diff --git a/lzma-report.xml b/lzma-report.xml
index 4294914..0dd9e88 100644
--- a/lzma-report.xml
+++ b/lzma-report.xml
@@ -1,340 +1,5176 @@
-<?xml version="1.0" encoding="utf-8"?><testsuites><testsuite name="pytest" errors="0" failures="113" skipped="0" tests="472" time="30.590" timestamp="2023-05-23T19:41:58.330131" hostname="FALCON-PUNCH"><testcase classname="test_lzma_detect" name="test_training_file" time="0.827"><properties><property name="score" value="3.3553375139606243" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[IFAW1.txt]" time="0.061"><properties><property name="score" value="0.635689245273463" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[cdc_covid.txt]" time="0.079"><properties><property name="score" value="0.033281207609958674" /></properties><failure message="AssertionError: cdc_covid.txt is a human-generated file, misclassified as AI-generated with confidence 0.03328121&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: cdc_covid.txt is a human-generated file, misclassified as AI-generated with confidence 0.03328121
+<?xml version="1.0" encoding="utf-8"?><testsuites><testsuite name="pytest" errors="0" failures="1725" skipped="0" tests="4225" time="133.806" timestamp="2023-05-31T08:21:32.092038" hostname="FALCON-PUNCH"><testcase classname="test_lzma_detect" name="test_training_file" time="0.369"><properties><property name="score" value="0.6548082558795084" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[IFAW1.txt]" time="0.054"><properties><property name="score" value="0.11089394659749915" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[cdc_covid.txt]" time="0.067"><properties><property name="score" value="0.05211779479496052" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[How_soon-Fans.txt]" time="0.054"><properties><property name="score" value="0.11874918746563057" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[hsus4.txt]" time="0.061"><properties><property name="score" value="0.08536936331899223" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[pg_robinson_crusoe.txt]" time="0.103"><properties><property name="score" value="0.08147970687180185" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[How_soon-Lebron-James.txt]" time="0.046"><properties><property name="score" value="0.11063212375655247" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Ohio_Steel.txt]" time="0.083"><properties><property name="score" value="0.08325107376730517" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[chapter-10.txt]" time="0.114"><properties><property name="score" value="0.08219067202887614" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Env_Prot_Agency-nov1.txt]" time="0.093"><properties><property name="score" value="0.02796117973983373" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[ch5.txt]" time="0.091"><properties><property name="score" value="0.05552852697494932" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[lw1.txt]" time="0.051"><properties><property name="score" value="0.1571472755753779" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Madame_White_Snake.txt]" time="0.060"><properties><property name="score" value="0.13476158795548432" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[110CYL070.txt]" time="0.011"><properties><property name="score" value="0.07113613799767937" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[audubon1.txt]" time="0.059"><properties><property name="score" value="0.09710520135096003" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[littleshelter2.txt]" time="0.012"><properties><property name="score" value="0.14711734575651447" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[guidedogs1.txt]" time="0.078"><properties><property name="score" value="0.10667204901102519" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[anth_essay_4.txt]" time="0.090"><properties><property name="score" value="0.03169431108608075" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[113CWL018.txt]" time="0.011"><properties><property name="score" value="0.08427863675998237" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[blog-monastery.txt]" time="0.056"><properties><property name="score" value="0.1064846949048715" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[115CVL037.txt]" time="0.020"><properties><property name="score" value="0.06652376679825633" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[115CVL036.txt]" time="0.025"><properties><property name="score" value="0.08434140943350556" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[A_defense_of_Michael_Moore.txt]" time="0.112"><properties><property name="score" value="0.08422627832093538" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[captured_moments.txt]" time="0.099"><properties><property name="score" value="0.17397547578235037" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[115CVL035.txt]" time="0.016"><properties><property name="score" value="0.08531489420113905" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[NWF1.txt]" time="0.053"><properties><property name="score" value="0.06262776429612216" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[WhereToHongKong.txt]" time="0.150"><properties><property name="score" value="0.16826744143453976" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[att2.txt]" time="0.019"><properties><property name="score" value="0.08782269284871481" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[c4_GOD.txt]" time="0.008"><properties><property name="score" value="0.06406097093498153" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[cdc_covid_asthma.txt]" time="0.075"><properties><property name="score" value="0.019490953523201154" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[fcic_final_report_conclusions.txt]" time="0.154"><properties><property name="score" value="0.061933984307962736" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[118CWL050.txt]" time="0.064"><properties><property name="score" value="0.08477265058365717" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[116CUL033.txt]" time="0.024"><properties><property name="score" value="0.06407752462125083" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[112C-L012.txt]" time="0.021"><properties><property name="score" value="0.07862490543384566" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[appalachian1.txt]" time="0.061"><properties><property name="score" value="0.1342189819398854" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[united1.txt]" time="0.016"><properties><property name="score" value="0.11840898199061421" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[117CWL009.txt]" time="0.008"><properties><property name="score" value="0.10786024947763706" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[marine1.txt]" time="0.061"><properties><property name="score" value="0.1421619229306082" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Ant_Robot.txt]" time="0.146"><properties><property name="score" value="0.038712171168113915" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[110CYL069.txt]" time="0.059"><properties><property name="score" value="0.09157264932137188" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[CUP2.txt]" time="0.114"><properties><property name="score" value="0.02158918625515785" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Tupelo-Honey-Cafe.txt]" time="0.051"><properties><property name="score" value="0.14950272283623164" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[119CWL041.txt]" time="0.013"><properties><property name="score" value="0.06691189466138114" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[lzma_readme.txt]" time="0.013"><properties><property name="score" value="0.006174266400283481" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Homosexuality.txt]" time="0.092"><properties><property name="score" value="0.06910862617883612" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[116CUL032.txt]" time="0.021"><properties><property name="score" value="0.054084195645810906" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[112C-L014.txt]" time="0.009"><properties><property name="score" value="0.10074219175651433" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Fermentation_HR5034.txt]" time="0.090"><properties><property name="score" value="0.09009918170890868" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Acephalous-Internet.txt]" time="0.054"><properties><property name="score" value="0.11843815576116173" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Effing-Idiot.txt]" time="0.102"><properties><property name="score" value="0.13795027744392074" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[114CUL059.txt]" time="0.018"><properties><property name="score" value="0.08418556274323144" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[sucker.txt]" time="0.026"><properties><property name="score" value="0.13225307660191118" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[c4_denver.txt]" time="0.019"><properties><property name="score" value="0.027867318350410786" /></properties><failure message="AssertionError: c4_denver.txt is a human-generated file, misclassified as AI-generated with confidence 0.02786732&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: c4_denver.txt is a human-generated file, misclassified as AI-generated with confidence 0.02786732
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_samples[How_soon-Fans.txt]" time="0.052"><properties><property name="score" value="0.7825709858086671" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[hsus4.txt]" time="0.070"><properties><property name="score" value="0.5917512337940967" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[pg_robinson_crusoe.txt]" time="0.138"><properties><property name="score" value="0.4227498199852101" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[How_soon-Lebron-James.txt]" time="0.057"><properties><property name="score" value="0.6631578470245175" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Ohio_Steel.txt]" time="0.178"><properties><property name="score" value="0.35614138590489014" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[chapter-10.txt]" time="0.247"><properties><property name="score" value="0.4436708019187562" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Env_Prot_Agency-nov1.txt]" time="0.185"><properties><property name="score" value="0.14783569336422925" /></properties><failure message="AssertionError: Env_Prot_Agency-nov1.txt is a human-generated file, misclassified as AI-generated with confidence 0.14783569&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: Env_Prot_Agency-nov1.txt is a human-generated file, misclassified as AI-generated with confidence 0.14783569
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_samples[blog-jet-lag.txt]" time="0.069"><properties><property name="score" value="0.09423934527113638" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[warner1.txt]" time="0.016"><properties><property name="score" value="0.1149024463850125" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Fermentation_Eminent-Domain.txt]" time="0.085"><properties><property name="score" value="0.1124238965222049" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[113CWL017.txt]" time="0.030"><properties><property name="score" value="0.08393069975080518" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[112C-L015.txt]" time="0.019"><properties><property name="score" value="0.00042659973741405316" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[blog-new-year's-resolutions.txt]" time="0.056"><properties><property name="score" value="0.016524892424573782" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[audubon2.txt]" time="0.054"><properties><property name="score" value="0.09881384770856343" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[The_Black_Willow.txt]" time="0.092"><properties><property name="score" value="0.12319734686527567" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Postal_Rate_Comm-ReportToCongress2002WEB.txt]" time="0.109"><properties><property name="score" value="0.02825713068741011" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[alumnifund1.txt]" time="0.018"><properties><property name="score" value="0.08396294463212785" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[112C-L016.txt]" time="0.008"><properties><property name="score" value="0.0433946554299347" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[114CUL057.txt]" time="0.011"><properties><property name="score" value="0.1306026598876464" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[110CYL067.txt]" time="0.026"><properties><property name="score" value="0.08521283249453868" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[hs_keyescrow.txt]" time="0.116"><properties><property name="score" value="0.06453433656153976" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[lessig_blog-carbon.txt]" time="0.060"><properties><property name="score" value="0.13523775016061282" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[114CUL060.txt]" time="0.016"><properties><property name="score" value="0.07840050791237094" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[vampires.txt]" time="0.078"><properties><property name="score" value="0.15678384165723758" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[chZ.txt]" time="0.092"><properties><property name="score" value="0.11400473800201787" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[ts_theme_intro.txt]" time="0.018"><properties><property name="score" value="0.07602784513293659" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[wildelifewatch1.txt]" time="0.089"><properties><property name="score" value="0.05609905234426857" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[aspca1.txt]" time="0.081"><properties><property name="score" value="0.12089759121374427" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[reuters_santos.txt]" time="0.034"><properties><property name="score" value="0.0582643559727527" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[110CYL071.txt]" time="0.018"><properties><property name="score" value="0.10844073611703929" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Uprooted_Bike.txt]" time="0.089"><properties><property name="score" value="0.1445643838714224" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[110CYL072.txt]" time="0.011"><properties><property name="score" value="0.06356350344637329" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Black_and_white.txt]" time="0.079"><properties><property name="score" value="0.07265941960798543" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[110CYL200.txt]" time="0.021"><properties><property name="score" value="0.09819805005532284" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[A_Wasted_Day.txt]" time="0.099"><properties><property name="score" value="0.1354063995711546" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[HistoryGreek.txt]" time="0.079"><properties><property name="score" value="0.13126995408995593" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[HistoryLasVegas.txt]" time="0.076"><properties><property name="score" value="0.14139894352840401" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[cnn_santos.txt]" time="0.050"><properties><property name="score" value="0.06419010365205424" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[detroit.txt]" time="0.076"><properties><property name="score" value="0.17692913736982213" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[118CWL049.txt]" time="0.020"><properties><property name="score" value="0.07606126857617934" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[hotel-california.txt]" time="0.133"><properties><property name="score" value="0.13758161379891432" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[about_me.txt]" time="0.010"><properties><property name="score" value="0.07119590600059844" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[602CZL285.txt]" time="0.019"><properties><property name="score" value="0.08808310600099495" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[WhatToHongKong.txt]" time="0.095"><properties><property name="score" value="0.1531564812140163" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Bartok.txt]" time="0.098"><properties><property name="score" value="0.03544361517480687" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Uprooted_Farming-on-Sand.txt]" time="0.060"><properties><property name="score" value="0.14672563231544922" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[116CUL034.txt]" time="0.052"><properties><property name="score" value="0.04146858970515002" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[114CUL058.txt]" time="0.017"><properties><property name="score" value="0.07495638629568618" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[IntroDublin.txt]" time="0.132"><properties><property name="score" value="0.113305964477298" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[defenders5.txt]" time="0.059"><properties><property name="score" value="0.08412387071346375" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Acephalous-Cant-believe.txt]" time="0.056"><properties><property name="score" value="0.11983969411447293" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[118CWL048.txt]" time="0.016"><properties><property name="score" value="0.0762750397531764" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[110CYL068.txt]" time="0.022"><properties><property name="score" value="0.09284413841263839" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[HistoryJerusalem.txt]" time="0.077"><properties><property name="score" value="0.14233588125297514" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[rybczynski-ch3.txt]" time="0.160"><properties><property name="score" value="0.09620412869669968" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[117CWL008.txt]" time="0.012"><properties><property name="score" value="0.08824191270048187" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Italy.txt]" time="0.064"><properties><property name="score" value="0.11944579806405398" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[CUP1.txt]" time="0.135"><properties><property name="score" value="0.03170345259863005" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Seedbombing.txt]" time="0.057"><properties><property name="score" value="0.1798065755985412" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Nathans_Bylichka.txt]" time="0.160"><properties><property name="score" value="0.1205193694010681" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Anti-Terrorist.txt]" time="0.170"><properties><property name="score" value="0.10673657128521682" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Protocol_Regarding_Access.txt]" time="0.228"><properties><property name="score" value="0.02395754292554558" /></properties><failure message="AssertionError: Protocol_Regarding_Access.txt is a human-generated file, misclassified as AI-generated with confidence 0.02395754&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: Protocol_Regarding_Access.txt is a human-generated file, misclassified as AI-generated with confidence 0.02395754
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_samples[ch5.txt]" time="0.154"><properties><property name="score" value="0.16652797279382162" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[lw1.txt]" time="0.049"><properties><property name="score" value="1.0814009877044146" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Madame_White_Snake.txt]" time="0.071"><properties><property name="score" value="0.8253672152757296" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[110CYL070.txt]" time="0.024"><properties><property name="score" value="0.4116860762837238" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[audubon1.txt]" time="0.065"><properties><property name="score" value="0.5561183405162048" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[littleshelter2.txt]" time="0.027"><properties><property name="score" value="0.2901890068449653" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[guidedogs1.txt]" time="0.060"><properties><property name="score" value="0.7548479059201335" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[anth_essay_4.txt]" time="0.129"><properties><property name="score" value="0.1424035885040198" /></properties><failure message="AssertionError: anth_essay_4.txt is a human-generated file, misclassified as AI-generated with confidence 0.14240359&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: anth_essay_4.txt is a human-generated file, misclassified as AI-generated with confidence 0.14240359
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_samples[IntroHongKong.txt]" time="0.070"><properties><property name="score" value="0.11958212796383579" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[wwf12.txt]" time="0.058"><properties><property name="score" value="0.06316740379128231" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[112C-L013.txt]" time="0.030"><properties><property name="score" value="0.09802891564239688" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[blog-varsity-athletics.txt]" time="0.093"><properties><property name="score" value="0.06722370987147686" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[AMC2.txt]" time="0.067"><properties><property name="score" value="0.07927838348409967" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.00667_generated.txt]" time="0.064"><properties><property name="score" value="0.026891321665825897" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.13472_generated.txt]" time="0.056"><properties><property name="score" value="0.011910534255182518" /></properties><failure message="AssertionError: 2110.13472_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.01191053&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.13472_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.01191053
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.03612_generated.txt]" time="0.091"><properties><property name="score" value="0.030446793618360602" /></properties><failure message="AssertionError: 2111.03612_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.03044679&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.03612_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.03044679
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.00310_generated.txt]" time="0.178"><properties><property name="score" value="0.03681195524085528" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.06230_generated.txt]" time="0.124"><properties><property name="score" value="0.015089880775588928" /></properties><failure message="AssertionError: 2111.06230_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.01508988&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.06230_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.01508988
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.06012_generated.txt]" time="0.069"><properties><property name="score" value="0.021465570194710386" /></properties><failure message="AssertionError: 2111.06012_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.02146557&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.06012_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.02146557
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.01676_generated.txt]" time="0.059"><properties><property name="score" value="0.0025651203638342467" /></properties><failure message="AssertionError: 2111.01676_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.00256512&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.01676_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.00256512
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2112.00405_generated.txt]" time="0.071"><properties><property name="score" value="0.010069357411523946" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[gpt_patient_summary.txt]" time="0.011"><properties><property name="score" value="0.01874326138860405" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[gpt_exercise.txt]" time="0.017"><properties><property name="score" value="0.039828809911594494" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.05241_generated.txt]" time="0.074"><properties><property name="score" value="0.03090638228254463" /></properties><failure message="AssertionError: 2111.05241_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.03090638&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.05241_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.03090638
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.12645_generated.txt]" time="0.057"><properties><property name="score" value="0.03142955710471515" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.07793_generated.txt]" time="0.081"><properties><property name="score" value="0.010046143403590354" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[cleaned_2111.00086_generated.txt]" time="0.066"><properties><property name="score" value="0.08741006339376889" /></properties><failure message="AssertionError: cleaned_2111.00086_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.08741006&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: cleaned_2111.00086_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.08741006
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.10340_generated.txt]" time="0.064"><properties><property name="score" value="0.07112407492772269" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.11115_generated.txt]" time="0.068"><properties><property name="score" value="0.042508213496260217" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.15093_generated.txt]" time="0.082"><properties><property name="score" value="0.10851395003974944" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.01515_generated.txt]" time="0.096"><properties><property name="score" value="0.002967395531662996" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.12501_generated.txt]" time="0.085"><properties><property name="score" value="0.009650421146984741" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.02041_generated.txt]" time="0.076"><properties><property name="score" value="0.019507383696192173" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.01023_generated.txt]" time="0.055"><properties><property name="score" value="0.0008952528864166788" /></properties><failure message="AssertionError: 2111.01023_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.00089525&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.01023_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.00089525
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.03715_generated.txt]" time="0.061"><properties><property name="score" value="0.03569432091862371" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[cgpt_analysis_explosion.txt]" time="0.070"><properties><property name="score" value="0.033067937750220516" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.02760_generated.txt]" time="0.094"><properties><property name="score" value="0.07724852395817007" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.02574_generated.txt]" time="0.065"><properties><property name="score" value="0.0030459930392749118" /></properties><failure message="AssertionError: 2111.02574_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.00304599&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.02574_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.00304599
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.15023_generated.txt]" time="0.084"><properties><property name="score" value="0.012822201466126431" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.03320_generated.txt]" time="0.079"><properties><property name="score" value="0.042565142156073245" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[gpt_patient_narrative.txt]" time="0.013"><properties><property name="score" value="0.053289218843100006" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.15317_generated.txt]" time="0.078"><properties><property name="score" value="0.007708554549460628" /></properties><failure message="AssertionError: 2110.15317_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.00770855&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.15317_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.00770855
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.06181_generated.txt]" time="0.064"><properties><property name="score" value="0.00019990109038806508" /></properties><failure message="AssertionError: 2111.06181_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.0001999&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.06181_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.0001999
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.03294_generated.txt]" time="0.057"><properties><property name="score" value="0.05282989061126761" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.15436_generated.txt]" time="0.066"><properties><property name="score" value="0.03255200191016512" /></properties><failure message="AssertionError: 2111.15436_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.032552&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.15436_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.032552
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.00867_generated.txt]" time="0.061"><properties><property name="score" value="0.038240293600237424" /></properties><failure message="AssertionError: 2111.00867_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.03824029&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.00867_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.03824029
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.05754_generated.txt]" time="0.065"><properties><property name="score" value="0.04686269560166414" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[dv_t1_aliens.txt]" time="0.048"><properties><property name="score" value="0.03428597947034474" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.12341_generated.txt]" time="0.063"><properties><property name="score" value="0.02775856397505502" /></properties><failure message="AssertionError: 2110.12341_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.02775856&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.12341_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.02775856
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.00086_generated.txt]" time="0.061"><properties><property name="score" value="0.07228922930222612" /></properties><failure message="AssertionError: 2111.00086_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.07228923&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.00086_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.07228923
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.02326_generated.txt]" time="0.071"><properties><property name="score" value="0.015437910727373754" /></properties><failure message="AssertionError: 2111.02326_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.01543791&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.02326_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.01543791
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.00514_generated.txt]" time="0.060"><properties><property name="score" value="0.015169026732234445" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.14532_generated.txt]" time="0.085"><properties><property name="score" value="0.01323241806589977" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.07408_generated.txt]" time="0.093"><properties><property name="score" value="0.0007878579979360121" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.00035_generated.txt]" time="0.100"><properties><property name="score" value="0.00027829857475752045" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.02687_generated.txt]" time="0.062"><properties><property name="score" value="0.054830247862516036" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.05204_generated.txt]" time="0.066"><properties><property name="score" value="0.0570873124357691" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.10478_generated.txt]" time="0.071"><properties><property name="score" value="0.041400091639774185" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.02188_generated.txt]" time="0.091"><properties><property name="score" value="0.01247571391482482" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.10575_generated.txt]" time="0.068"><properties><property name="score" value="0.011499984139099086" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.07525_generated.txt]" time="0.066"><properties><property name="score" value="0.015054160477072677" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.10577_generated.txt]" time="0.061"><properties><property name="score" value="0.007729289738388852" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.15473_generated.txt]" time="0.076"><properties><property name="score" value="0.020010137280216698" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[bing_c_abstract.txt]" time="0.016"><properties><property name="score" value="0.04670221893028348" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.10778_generated.txt]" time="0.056"><properties><property name="score" value="0.05529787751369155" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.13900_generated.txt]" time="0.069"><properties><property name="score" value="0.026022600732613144" /></properties><failure message="AssertionError: 2110.13900_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.0260226&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.13900_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.0260226
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.04416_generated.txt]" time="0.061"><properties><property name="score" value="0.0019234142215360695" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.03913_generated.txt]" time="0.062"><properties><property name="score" value="0.021247123633177312" /></properties><failure message="AssertionError: 2111.03913_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.02124712&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.03913_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.02124712
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.01706_generated.txt]" time="0.065"><properties><property name="score" value="0.009518074550131542" /></properties><failure message="AssertionError: 2111.01706_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.00951807&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.01706_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.00951807
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.01243_generated.txt]" time="0.056"><properties><property name="score" value="0.007957128745908945" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.11879_generated.txt]" time="0.062"><properties><property name="score" value="0.016032516286446655" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.10329_generated.txt]" time="0.061"><properties><property name="score" value="0.006344588252617486" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.04507_generated.txt]" time="0.064"><properties><property name="score" value="0.019815092789257284" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.03837_generated.txt]" time="0.069"><properties><property name="score" value="0.040895831019308826" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.00572_generated.txt]" time="0.066"><properties><property name="score" value="0.053053050109075225" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.11984_generated.txt]" time="0.064"><properties><property name="score" value="0.008917514894332602" /></properties><failure message="AssertionError: 2110.11984_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.00891751&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.11984_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.00891751
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.06464_generated.txt]" time="0.068"><properties><property name="score" value="0.006824490402934447" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[cgpt_hs_essay2.txt]" time="0.076"><properties><property name="score" value="0.0038316300121026977" /></properties><failure message="AssertionError: cgpt_hs_essay2.txt is an LLM-generated file, misclassified as human-generated with confidence 0.00383163&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: cgpt_hs_essay2.txt is an LLM-generated file, misclassified as human-generated with confidence 0.00383163
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.02110_generated.txt]" time="0.099"><properties><property name="score" value="0.041055041729549366" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[cgpt_hs_essay.txt]" time="0.101"><properties><property name="score" value="0.028425787834487747" /></properties><failure message="AssertionError: cgpt_hs_essay.txt is an LLM-generated file, misclassified as human-generated with confidence 0.02842579&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: cgpt_hs_essay.txt is an LLM-generated file, misclassified as human-generated with confidence 0.02842579
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.15802_generated.txt]" time="0.066"><properties><property name="score" value="0.024158792164691308" /></properties><failure message="AssertionError: 2110.15802_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.02415879&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.15802_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.02415879
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.11589_generated.txt]" time="0.062"><properties><property name="score" value="0.0395345960004591" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.15724_generated.txt]" time="0.070"><properties><property name="score" value="0.03361888707912073" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.15130_generated.txt]" time="0.070"><properties><property name="score" value="0.0006804434561482535" /></properties><failure message="AssertionError: 2110.15130_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.00068044&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.15130_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.00068044
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.06644_generated.txt]" time="0.123"><properties><property name="score" value="5.017205282722547e-05" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[bard_news.txt]" time="0.021"><properties><property name="score" value="0.07609946479712426" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.01322_generated.txt]" time="0.066"><properties><property name="score" value="0.0645060125260033" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[about_me_short.txt]" time="0.009"><properties><property name="score" value="0.022435947049664895" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[bing_b_song.txt]" time="0.013"><properties><property name="score" value="0.06349965698745907" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.00526_generated.txt]" time="0.101"><properties><property name="score" value="0.009700218738143262" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.12010_generated.txt]" time="0.068"><properties><property name="score" value="0.014198854202833615" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[cgpt_planet_abs.txt]" time="0.018"><properties><property name="score" value="0.007323842943515962" /></properties><failure message="AssertionError: cgpt_planet_abs.txt is an LLM-generated file, misclassified as human-generated with confidence 0.00732384&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: cgpt_planet_abs.txt is an LLM-generated file, misclassified as human-generated with confidence 0.00732384
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.13317_generated.txt]" time="0.080"><properties><property name="score" value="0.034368076530993025" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[gpt_santos_long.txt]" time="0.018"><properties><property name="score" value="0.08687296539291924" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[gpt_santos_short.txt]" time="0.010"><properties><property name="score" value="0.036748653640045204" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.02643_generated.txt]" time="0.095"><properties><property name="score" value="0.04323629523956635" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[about_me.txt]" time="0.031"><properties><property name="score" value="0.03650856210351417" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.15534_generated.txt]" time="0.078"><properties><property name="score" value="0.023258227358238577" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.07611_generated.txt]" time="0.091"><properties><property name="score" value="0.008878663310623147" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.13658_generated.txt]" time="0.066"><properties><property name="score" value="0.02699764271647433" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.15725_generated.txt]" time="0.059"><properties><property name="score" value="0.21004687858009405" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.01231_generated.txt]" time="0.059"><properties><property name="score" value="0.014329250024448388" /></properties><failure message="AssertionError: 2111.01231_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.01432925&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.01231_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.01432925
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.10319_generated.txt]" time="0.071"><properties><property name="score" value="0.054226302370966074" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.01340_generated.txt]" time="0.075"><properties><property name="score" value="0.030427286412469588" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.07699_generated.txt]" time="0.095"><properties><property name="score" value="0.004579660864132976" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.15705_generated.txt]" time="0.069"><properties><property name="score" value="0.019546349564217875" /></properties><failure message="AssertionError: 2110.15705_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.01954635&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.15705_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.01954635
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.02844_generated.txt]" time="0.064"><properties><property name="score" value="0.17955517433113896" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.04130_generated.txt]" time="0.061"><properties><property name="score" value="0.0014743456769695162" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.02259_generated.txt]" time="0.060"><properties><property name="score" value="0.03981341051323346" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.00180_generated.txt]" time="0.074"><properties><property name="score" value="0.0020256477506176702" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.04574_generated.txt]" time="0.068"><properties><property name="score" value="0.03738718715538765" /></properties><failure message="AssertionError: 2111.04574_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.03738719&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.04574_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.03738719
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.00808_generated.txt]" time="0.060"><properties><property name="score" value="0.005626824373011194" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.12765_generated.txt]" time="0.059"><properties><property name="score" value="0.003780733067831982" /></properties><failure message="AssertionError: 2110.12765_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.00378073&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.12765_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.00378073
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.13229_generated.txt]" time="0.059"><properties><property name="score" value="0.013793634739519476" /></properties><failure message="AssertionError: 2110.13229_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.01379363&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.13229_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.01379363
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.00554_generated.txt]" time="0.066"><properties><property name="score" value="0.06571039340181636" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.11207_generated.txt]" time="0.055"><properties><property name="score" value="0.004092658226009885" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.00607_generated.txt]" time="0.089"><properties><property name="score" value="0.057708731553457515" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.07267_generated.txt]" time="0.062"><properties><property name="score" value="0.06451354146371271" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.11205_generated.txt]" time="0.072"><properties><property name="score" value="0.01566518163773588" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[bard_internet_abstract.txt]" time="0.028"><properties><property name="score" value="0.1185537261588465" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.03945_generated.txt]" time="0.057"><properties><property name="score" value="0.017836148934244095" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.12383_generated.txt]" time="0.070"><properties><property name="score" value="0.028922123366813872" /></properties><failure message="AssertionError: 2110.12383_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.02892212&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.12383_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.02892212
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.02362_generated.txt]" time="0.073"><properties><property name="score" value="0.020314097358401232" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.15799_generated.txt]" time="0.067"><properties><property name="score" value="0.00583365003010263" /></properties><failure message="AssertionError: 2110.15799_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.00583365&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.15799_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.00583365
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[bard_abstract.txt]" time="0.012"><properties><property name="score" value="0.06758585441330456" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i0]" time="0.015"><properties><property name="score" value="0.0611721834392176" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255000 (len: 134) is a human-generated sample, misclassified as AI-generated with confidence 0.06117218&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255000 (len: 134) is a human-generated sample, misclassified as AI-generated with confidence 0.06117218
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_samples[113CWL018.txt]" time="0.020"><properties><property name="score" value="0.48214498662640937" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[blog-monastery.txt]" time="0.065"><properties><property name="score" value="0.5621473919967257" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[115CVL037.txt]" time="0.050"><properties><property name="score" value="0.2286792303657729" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[115CVL036.txt]" time="0.022"><properties><property name="score" value="0.7237798340359591" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[A_defense_of_Michael_Moore.txt]" time="0.141"><properties><property name="score" value="0.46668147295538176" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[captured_moments.txt]" time="0.137"><properties><property name="score" value="1.118506134821687" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[115CVL035.txt]" time="0.025"><properties><property name="score" value="0.675217256088656" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[NWF1.txt]" time="0.055"><properties><property name="score" value="0.019733771081920093" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[WhereToHongKong.txt]" time="0.244"><properties><property name="score" value="1.0093731570497944" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[att2.txt]" time="0.048"><properties><property name="score" value="0.5815950096597818" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[c4_GOD.txt]" time="0.027"><properties><property name="score" value="0.14205987975671186" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[cdc_covid_asthma.txt]" time="0.052"><properties><property name="score" value="0.09042061163495874" /></properties><failure message="AssertionError: cdc_covid_asthma.txt is a human-generated file, misclassified as AI-generated with confidence 0.09042061&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: cdc_covid_asthma.txt is a human-generated file, misclassified as AI-generated with confidence 0.09042061
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i1]" time="0.057"><properties><property name="score" value="0.08311791222576946" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i2]" time="0.013"><properties><property name="score" value="0.013949229033954902" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255002 (len: 173) is a human-generated sample, misclassified as AI-generated with confidence 0.01394923&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255002 (len: 173) is a human-generated sample, misclassified as AI-generated with confidence 0.01394923
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_samples[fcic_final_report_conclusions.txt]" time="0.196"><properties><property name="score" value="0.0369182544039419" /></properties><failure message="AssertionError: fcic_final_report_conclusions.txt is a human-generated file, misclassified as AI-generated with confidence 0.03691825&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: fcic_final_report_conclusions.txt is a human-generated file, misclassified as AI-generated with confidence 0.03691825
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i3]" time="0.017"><properties><property name="score" value="0.06407692136350143" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i4]" time="0.009"><properties><property name="score" value="0.16029518242817176" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i5]" time="0.014"><properties><property name="score" value="0.03149819336629167" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i6]" time="0.029"><properties><property name="score" value="0.04168443104275104" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i7]" time="0.016"><properties><property name="score" value="0.07460609767287463" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i8]" time="0.014"><properties><property name="score" value="0.17599322009336227" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i9]" time="0.008"><properties><property name="score" value="0.026344502408498505" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i10]" time="0.009"><properties><property name="score" value="0.04164683091568144" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i11]" time="0.054"><properties><property name="score" value="0.07663350725359369" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i12]" time="0.020"><properties><property name="score" value="0.08812680438886775" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i13]" time="0.060"><properties><property name="score" value="0.1219662067797188" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i14]" time="0.013"><properties><property name="score" value="0.07266007308955325" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i15]" time="0.075"><properties><property name="score" value="0.057847104443006045" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i16]" time="0.010"><properties><property name="score" value="0.044154202872553894" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i17]" time="0.009"><properties><property name="score" value="0.051087958556172985" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i18]" time="0.025"><properties><property name="score" value="0.09955453343130824" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i19]" time="0.032"><properties><property name="score" value="0.08255719449053153" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i20]" time="0.072"><properties><property name="score" value="0.09748343364056232" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i21]" time="0.069"><properties><property name="score" value="0.12656239296936744" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i22]" time="0.060"><properties><property name="score" value="0.15041361276918522" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i23]" time="0.011"><properties><property name="score" value="0.04490700689807059" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i24]" time="0.012"><properties><property name="score" value="0.03413391169426738" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i25]" time="0.053"><properties><property name="score" value="0.08993349990268334" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i26]" time="0.062"><properties><property name="score" value="0.023325381983297283" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i27]" time="0.077"><properties><property name="score" value="0.0033891496099920145" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i28]" time="0.040"><properties><property name="score" value="0.005764324565243512" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i29]" time="0.033"><properties><property name="score" value="0.11346315103934951" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i30]" time="0.054"><properties><property name="score" value="0.11150943859325076" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i31]" time="0.009"><properties><property name="score" value="0.009966675066641617" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i32]" time="0.013"><properties><property name="score" value="0.05225512398264187" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i33]" time="0.075"><properties><property name="score" value="0.08663817277775213" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i34]" time="0.017"><properties><property name="score" value="0.13263487349268743" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i35]" time="0.077"><properties><property name="score" value="0.09564532983842453" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i36]" time="0.023"><properties><property name="score" value="0.061010541395972104" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i37]" time="0.017"><properties><property name="score" value="0.11003873114352206" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i38]" time="0.007"><properties><property name="score" value="0.029450345649839615" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i39]" time="0.007"><properties><property name="score" value="0.0055932115215318046" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i40]" time="0.061"><properties><property name="score" value="0.07370051288056245" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i41]" time="0.020"><properties><property name="score" value="0.026230819705383605" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i42]" time="0.014"><properties><property name="score" value="0.04426705323418545" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255042 (len: 531) is a human-generated sample, misclassified as AI-generated with confidence 0.04426705&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255042 (len: 531) is a human-generated sample, misclassified as AI-generated with confidence 0.04426705
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_samples[118CWL050.txt]" time="0.052"><properties><property name="score" value="0.40014314800861095" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[116CUL033.txt]" time="0.026"><properties><property name="score" value="0.016012841885348905" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[112C-L012.txt]" time="0.023"><properties><property name="score" value="0.6141202572082505" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[appalachian1.txt]" time="0.066"><properties><property name="score" value="0.918961659612201" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[united1.txt]" time="0.052"><properties><property name="score" value="0.5939726790035923" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[117CWL009.txt]" time="0.032"><properties><property name="score" value="0.5827274445339148" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[marine1.txt]" time="0.056"><properties><property name="score" value="0.648306704003021" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Ant_Robot.txt]" time="0.201"><properties><property name="score" value="0.07938750635653476" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[110CYL069.txt]" time="0.051"><properties><property name="score" value="0.5256665165530305" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[CUP2.txt]" time="0.164"><properties><property name="score" value="0.046969781020568126" /></properties><failure message="AssertionError: CUP2.txt is a human-generated file, misclassified as AI-generated with confidence 0.04696978&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: CUP2.txt is a human-generated file, misclassified as AI-generated with confidence 0.04696978
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i43]" time="0.013"><properties><property name="score" value="0.003860792711257721" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i44]" time="0.008"><properties><property name="score" value="0.023492238904887497" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i45]" time="0.059"><properties><property name="score" value="0.08820578812536994" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i46]" time="0.051"><properties><property name="score" value="0.14294269488443276" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i47]" time="0.066"><properties><property name="score" value="0.09320861274941623" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i48]" time="0.074"><properties><property name="score" value="0.0689674833257161" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i49]" time="0.032"><properties><property name="score" value="0.010137090996656481" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i50]" time="0.068"><properties><property name="score" value="0.11177717593437811" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i51]" time="0.094"><properties><property name="score" value="0.006571132171763673" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i52]" time="0.014"><properties><property name="score" value="0.04546938168650816" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i53]" time="0.008"><properties><property name="score" value="0.13928940758211605" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i54]" time="0.017"><properties><property name="score" value="0.13043186315843594" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i55]" time="0.055"><properties><property name="score" value="0.11440516114070076" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i56]" time="0.071"><properties><property name="score" value="0.10295236917241662" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i57]" time="0.056"><properties><property name="score" value="0.08634080940460609" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i58]" time="0.059"><properties><property name="score" value="0.10607641406082467" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i59]" time="0.011"><properties><property name="score" value="0.12008826858081467" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i60]" time="0.050"><properties><property name="score" value="0.14548713100418986" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i61]" time="0.073"><properties><property name="score" value="0.09386605552998817" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i62]" time="0.016"><properties><property name="score" value="0.057446393787113914" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i63]" time="0.007"><properties><property name="score" value="0.07011687784464482" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i64]" time="0.052"><properties><property name="score" value="0.15938836895565955" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i65]" time="0.009"><properties><property name="score" value="0.064345172186836" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i66]" time="0.076"><properties><property name="score" value="0.11308677792157094" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i67]" time="0.015"><properties><property name="score" value="0.10682577347230626" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i68]" time="0.009"><properties><property name="score" value="0.11514572564247372" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i69]" time="0.009"><properties><property name="score" value="0.07094395444617319" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i70]" time="0.070"><properties><property name="score" value="0.08462038565567569" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i71]" time="0.018"><properties><property name="score" value="0.05784938768266545" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i72]" time="0.014"><properties><property name="score" value="0.09213878698797388" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i73]" time="0.008"><properties><property name="score" value="0.024507709569121516" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255074 (len: 355) is a human-generated sample, misclassified as AI-generated with confidence 0.02450771&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255074 (len: 355) is a human-generated sample, misclassified as AI-generated with confidence 0.02450771
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Tupelo-Honey-Cafe.txt]" time="0.053"><properties><property name="score" value="0.9137336815517316" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[119CWL041.txt]" time="0.026"><properties><property name="score" value="0.05162571604599986" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[lzma_readme.txt]" time="0.042"><properties><property name="score" value="0.1286665341867499" /></properties><failure message="AssertionError: lzma_readme.txt is a human-generated file, misclassified as AI-generated with confidence 0.12866653&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: lzma_readme.txt is a human-generated file, misclassified as AI-generated with confidence 0.12866653
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i74]" time="0.016"><properties><property name="score" value="0.11724451313939824" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i75]" time="0.010"><properties><property name="score" value="0.09652390798107757" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i76]" time="0.008"><properties><property name="score" value="0.034289911570406506" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i77]" time="0.052"><properties><property name="score" value="0.08676686469780753" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i78]" time="0.010"><properties><property name="score" value="0.059051073932792075" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i79]" time="0.057"><properties><property name="score" value="0.06678314096779375" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i80]" time="0.065"><properties><property name="score" value="0.055832695513019004" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i81]" time="0.010"><properties><property name="score" value="0.07941197946654821" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i82]" time="0.059"><properties><property name="score" value="0.14169465311113258" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i83]" time="0.047"><properties><property name="score" value="0.10980867445248498" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i84]" time="0.010"><properties><property name="score" value="0.12409236481312202" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i85]" time="0.069"><properties><property name="score" value="0.0566117222773485" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i86]" time="0.028"><properties><property name="score" value="0.07091170654974777" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i87]" time="0.008"><properties><property name="score" value="0.04048010875302155" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i88]" time="0.059"><properties><property name="score" value="0.09314910635177405" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i89]" time="0.009"><properties><property name="score" value="0.03016634089959558" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i90]" time="0.011"><properties><property name="score" value="0.06192038023590252" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i91]" time="0.068"><properties><property name="score" value="0.05526529485849066" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i92]" time="0.025"><properties><property name="score" value="0.10395206774504995" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i93]" time="0.016"><properties><property name="score" value="0.07653653747073354" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i94]" time="0.074"><properties><property name="score" value="0.07082671016313125" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i95]" time="0.033"><properties><property name="score" value="0.05916530828815852" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255096 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.05916531&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255096 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.05916531
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Homosexuality.txt]" time="0.121"><properties><property name="score" value="0.22304125136977243" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[116CUL032.txt]" time="0.047"><properties><property name="score" value="0.14123168307063802" /></properties><failure message="AssertionError: 116CUL032.txt is a human-generated file, misclassified as AI-generated with confidence 0.14123168&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: 116CUL032.txt is a human-generated file, misclassified as AI-generated with confidence 0.14123168
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i96]" time="0.018"><properties><property name="score" value="0.03212757863707005" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i97]" time="0.031"><properties><property name="score" value="0.05982369227281794" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i98]" time="0.011"><properties><property name="score" value="0.03995880213328862" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i99]" time="0.013"><properties><property name="score" value="0.0009483706931625105" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255100 (len: 512) is a human-generated sample, misclassified as AI-generated with confidence 0.00094837&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255100 (len: 512) is a human-generated sample, misclassified as AI-generated with confidence 0.00094837
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_samples[112C-L014.txt]" time="0.026"><properties><property name="score" value="0.05587060631945251" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Fermentation_HR5034.txt]" time="0.089"><properties><property name="score" value="0.5810542027073206" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Acephalous-Internet.txt]" time="0.057"><properties><property name="score" value="0.7633012382572524" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Effing-Idiot.txt]" time="0.137"><properties><property name="score" value="0.8767068234365423" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[114CUL059.txt]" time="0.023"><properties><property name="score" value="0.705263075428425" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[sucker.txt]" time="0.047"><properties><property name="score" value="0.6919920118111933" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[c4_denver.txt]" time="0.016"><properties><property name="score" value="0.08777060748943466" /></properties><failure message="AssertionError: c4_denver.txt is a human-generated file, misclassified as AI-generated with confidence 0.08777061&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: c4_denver.txt is a human-generated file, misclassified as AI-generated with confidence 0.08777061
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i100]" time="0.057"><properties><property name="score" value="0.10833685228987845" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i101]" time="0.012"><properties><property name="score" value="0.051724406487890784" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i102]" time="0.068"><properties><property name="score" value="0.059855689269458855" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i103]" time="0.011"><properties><property name="score" value="0.06310533164369247" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i104]" time="0.079"><properties><property name="score" value="0.03882885674539835" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i105]" time="0.021"><properties><property name="score" value="0.09899800808640255" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i106]" time="0.007"><properties><property name="score" value="0.11546846804143529" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i107]" time="0.014"><properties><property name="score" value="0.10090962620985416" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i108]" time="0.014"><properties><property name="score" value="0.09746681124539372" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i109]" time="0.008"><properties><property name="score" value="0.040909101240810264" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i110]" time="0.019"><properties><property name="score" value="0.08496287017585907" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i111]" time="0.019"><properties><property name="score" value="0.06898142236513165" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i112]" time="0.016"><properties><property name="score" value="0.09957241174803345" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i113]" time="0.007"><properties><property name="score" value="0.11899964856257883" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i114]" time="0.054"><properties><property name="score" value="0.011240598551779089" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255115 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.0112406&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255115 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.0112406
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_samples[blog-jet-lag.txt]" time="0.063"><properties><property name="score" value="0.29858053939392964" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[warner1.txt]" time="0.049"><properties><property name="score" value="0.7028671431591101" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Fermentation_Eminent-Domain.txt]" time="0.088"><properties><property name="score" value="0.5152209612162958" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[113CWL017.txt]" time="0.024"><properties><property name="score" value="0.1428027595141712" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[112C-L015.txt]" time="0.032"><properties><property name="score" value="0.37299215818922815" /></properties><failure message="AssertionError: 112C-L015.txt is a human-generated file, misclassified as AI-generated with confidence 0.37299216&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: 112C-L015.txt is a human-generated file, misclassified as AI-generated with confidence 0.37299216
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i115]" time="0.054"><properties><property name="score" value="0.05815740037418521" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i116]" time="0.055"><properties><property name="score" value="0.005576821707041851" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i117]" time="0.017"><properties><property name="score" value="0.10008928680535432" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i118]" time="0.055"><properties><property name="score" value="0.07550009610763285" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i119]" time="0.009"><properties><property name="score" value="0.05857008739899583" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i120]" time="0.016"><properties><property name="score" value="0.08210003166963055" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i121]" time="0.020"><properties><property name="score" value="0.10658662349390748" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i122]" time="0.048"><properties><property name="score" value="0.10124759175398045" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i123]" time="0.015"><properties><property name="score" value="0.14931050680073166" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i124]" time="0.011"><properties><property name="score" value="0.1539129060390998" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i125]" time="0.011"><properties><property name="score" value="0.12145803828710888" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i126]" time="0.021"><properties><property name="score" value="0.06087895746663695" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i127]" time="0.060"><properties><property name="score" value="0.07525904428686422" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i128]" time="0.056"><properties><property name="score" value="0.1155637329835077" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i129]" time="0.093"><properties><property name="score" value="0.05712630890263247" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i130]" time="0.066"><properties><property name="score" value="0.0463455903256647" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i131]" time="0.012"><properties><property name="score" value="0.03415516305748101" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i132]" time="0.058"><properties><property name="score" value="0.0735931971371119" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i133]" time="0.009"><properties><property name="score" value="0.07088499799987602" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i134]" time="0.020"><properties><property name="score" value="0.135500924929439" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i135]" time="0.010"><properties><property name="score" value="0.004791780050023009" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255136 (len: 80) is a human-generated sample, misclassified as AI-generated with confidence 0.00479178&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255136 (len: 80) is a human-generated sample, misclassified as AI-generated with confidence 0.00479178
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_samples[blog-new-year's-resolutions.txt]" time="0.066"><properties><property name="score" value="0.03853273265985724" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[audubon2.txt]" time="0.051"><properties><property name="score" value="0.7083282975687694" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[The_Black_Willow.txt]" time="0.121"><properties><property name="score" value="0.7991212392105086" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Postal_Rate_Comm-ReportToCongress2002WEB.txt]" time="0.157"><properties><property name="score" value="0.1366068786052456" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[alumnifund1.txt]" time="0.024"><properties><property name="score" value="0.7985963762248971" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[112C-L016.txt]" time="0.024"><properties><property name="score" value="0.1306064031072196" /></properties><failure message="AssertionError: 112C-L016.txt is a human-generated file, misclassified as AI-generated with confidence 0.1306064&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: 112C-L016.txt is a human-generated file, misclassified as AI-generated with confidence 0.1306064
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i136]" time="0.006"><properties><property name="score" value="0.08212302252030845" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i137]" time="0.009"><properties><property name="score" value="0.02009342972492112" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i138]" time="0.051"><properties><property name="score" value="0.04412444486731455" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i139]" time="0.009"><properties><property name="score" value="0.05756026681555393" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i140]" time="0.016"><properties><property name="score" value="0.01873933467352374" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i141]" time="0.053"><properties><property name="score" value="0.0892461830630699" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i142]" time="0.009"><properties><property name="score" value="0.12696564743750471" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i143]" time="0.058"><properties><property name="score" value="0.0338282509550876" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i144]" time="0.056"><properties><property name="score" value="0.1285160210611866" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i145]" time="0.017"><properties><property name="score" value="0.003615619415864302" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255146 (len: 377) is a human-generated sample, misclassified as AI-generated with confidence 0.00361562&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255146 (len: 377) is a human-generated sample, misclassified as AI-generated with confidence 0.00361562
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_samples[114CUL057.txt]" time="0.023"><properties><property name="score" value="0.19735403379250727" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[110CYL067.txt]" time="0.058"><properties><property name="score" value="0.35433689389482065" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[hs_keyescrow.txt]" time="0.104"><properties><property name="score" value="0.04309813698838427" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[lessig_blog-carbon.txt]" time="0.073"><properties><property name="score" value="0.7840444525208996" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[114CUL060.txt]" time="0.046"><properties><property name="score" value="0.564368960345211" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[vampires.txt]" time="0.060"><properties><property name="score" value="0.8823583649036904" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[chZ.txt]" time="0.061"><properties><property name="score" value="0.756033381880831" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[ts_theme_intro.txt]" time="0.025"><properties><property name="score" value="0.33405336688516196" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[wildelifewatch1.txt]" time="0.051"><properties><property name="score" value="0.10619794045219066" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[aspca1.txt]" time="0.055"><properties><property name="score" value="0.8701427000135653" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[reuters_santos.txt]" time="0.037"><properties><property name="score" value="0.040078306845686695" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[110CYL071.txt]" time="0.072"><properties><property name="score" value="0.44932901188415864" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Uprooted_Bike.txt]" time="0.084"><properties><property name="score" value="0.9297163581784957" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[110CYL072.txt]" time="0.013"><properties><property name="score" value="0.4105112031408398" /></properties><failure message="AssertionError: 110CYL072.txt is a human-generated file, misclassified as AI-generated with confidence 0.4105112&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: 110CYL072.txt is a human-generated file, misclassified as AI-generated with confidence 0.4105112
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i146]" time="0.082"><properties><property name="score" value="0.14051902308899028" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i147]" time="0.016"><properties><property name="score" value="0.059634818945056414" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i148]" time="0.017"><properties><property name="score" value="0.03472368199624146" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i149]" time="0.009"><properties><property name="score" value="0.033584620119964104" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i150]" time="0.008"><properties><property name="score" value="0.15369585697069255" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i151]" time="0.008"><properties><property name="score" value="0.03263693695154113" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i152]" time="0.012"><properties><property name="score" value="0.05787776328958216" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i153]" time="0.056"><properties><property name="score" value="0.1113696831204824" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i154]" time="0.056"><properties><property name="score" value="0.1040258888900088" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i155]" time="0.024"><properties><property name="score" value="0.02466792249293592" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i156]" time="0.039"><properties><property name="score" value="0.14611579476935427" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i157]" time="0.021"><properties><property name="score" value="0.020867756355019695" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i158]" time="0.059"><properties><property name="score" value="0.08180943908761427" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i159]" time="0.055"><properties><property name="score" value="0.02127618607021197" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i160]" time="0.025"><properties><property name="score" value="0.05260291079658215" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i161]" time="0.060"><properties><property name="score" value="0.0882656155502401" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i162]" time="0.073"><properties><property name="score" value="0.0448303414963331" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i163]" time="0.080"><properties><property name="score" value="0.11996314445246561" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i164]" time="0.036"><properties><property name="score" value="0.09868359089866174" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i165]" time="0.016"><properties><property name="score" value="0.032255398981811956" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i166]" time="0.020"><properties><property name="score" value="0.03364685997794825" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i167]" time="0.079"><properties><property name="score" value="0.10096977799594688" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i168]" time="0.082"><properties><property name="score" value="0.0470241291225912" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i169]" time="0.067"><properties><property name="score" value="0.1269760694884129" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i170]" time="0.066"><properties><property name="score" value="0.08235860787180177" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i171]" time="0.035"><properties><property name="score" value="0.0225922772555746" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i172]" time="0.017"><properties><property name="score" value="0.10176255933404084" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i173]" time="0.008"><properties><property name="score" value="0.06276846123823288" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i174]" time="0.051"><properties><property name="score" value="0.09739603267605096" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i175]" time="0.022"><properties><property name="score" value="0.10181288422940295" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i176]" time="0.006"><properties><property name="score" value="0.07773762350292701" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i177]" time="0.018"><properties><property name="score" value="0.10859253283200294" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i178]" time="0.017"><properties><property name="score" value="0.1450057495725282" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i179]" time="0.012"><properties><property name="score" value="0.15042566319670714" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i180]" time="0.061"><properties><property name="score" value="0.026844223870647412" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i181]" time="0.018"><properties><property name="score" value="0.05286801207482984" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i182]" time="0.058"><properties><property name="score" value="0.029052707643209538" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i183]" time="0.013"><properties><property name="score" value="0.07911645520753408" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i184]" time="0.009"><properties><property name="score" value="0.04500759533188625" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i185]" time="0.007"><properties><property name="score" value="0.0697315590214409" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i186]" time="0.015"><properties><property name="score" value="0.09398523201562183" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i187]" time="0.060"><properties><property name="score" value="0.056100186570181385" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255188 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.05610019&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255188 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.05610019
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Black_and_white.txt]" time="0.107"><properties><property name="score" value="0.23398112833891543" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[110CYL200.txt]" time="0.048"><properties><property name="score" value="0.5005388917362907" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[A_Wasted_Day.txt]" time="0.141"><properties><property name="score" value="0.9050069975729563" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[HistoryGreek.txt]" time="0.096"><properties><property name="score" value="0.8127580208246963" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[HistoryLasVegas.txt]" time="0.097"><properties><property name="score" value="0.8208391370003141" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[cnn_santos.txt]" time="0.052"><properties><property name="score" value="0.16685729310950137" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[detroit.txt]" time="0.103"><properties><property name="score" value="1.0786613220067809" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[118CWL049.txt]" time="0.046"><properties><property name="score" value="0.5654421247669831" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[hotel-california.txt]" time="0.153"><properties><property name="score" value="0.9104542752929639" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[about_me.txt]" time="0.012"><properties><property name="score" value="0.6744447024438771" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[602CZL285.txt]" time="0.023"><properties><property name="score" value="0.8651959086094785" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[WhatToHongKong.txt]" time="0.119"><properties><property name="score" value="0.9268188121898087" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Bartok.txt]" time="0.145"><properties><property name="score" value="0.1606593069525775" /></properties><failure message="AssertionError: Bartok.txt is a human-generated file, misclassified as AI-generated with confidence 0.16065931&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: Bartok.txt is a human-generated file, misclassified as AI-generated with confidence 0.16065931
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i188]" time="0.012"><properties><property name="score" value="0.1288339933386451" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i189]" time="0.051"><properties><property name="score" value="0.09577164922581884" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i190]" time="0.060"><properties><property name="score" value="0.04162759987326195" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i191]" time="0.011"><properties><property name="score" value="0.05238013002996089" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i192]" time="0.073"><properties><property name="score" value="0.06849418068783618" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i193]" time="0.070"><properties><property name="score" value="0.1102474376780178" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i194]" time="0.012"><properties><property name="score" value="0.07028079090569683" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i195]" time="0.011"><properties><property name="score" value="0.03397078422286426" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i196]" time="0.085"><properties><property name="score" value="0.06753859068643334" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i197]" time="0.019"><properties><property name="score" value="0.004715356895953397" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255198 (len: 500) is a human-generated sample, misclassified as AI-generated with confidence 0.00471536&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255198 (len: 500) is a human-generated sample, misclassified as AI-generated with confidence 0.00471536
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Uprooted_Farming-on-Sand.txt]" time="0.072"><properties><property name="score" value="0.9753671396823927" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[116CUL034.txt]" time="0.063"><properties><property name="score" value="0.04325603526836863" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[114CUL058.txt]" time="0.047"><properties><property name="score" value="0.5798466695012272" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[IntroDublin.txt]" time="0.061"><properties><property name="score" value="0.6041883976987366" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[defenders5.txt]" time="0.061"><properties><property name="score" value="0.5564494052139006" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Acephalous-Cant-believe.txt]" time="0.053"><properties><property name="score" value="0.6753657357551901" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[118CWL048.txt]" time="0.024"><properties><property name="score" value="0.12156192517884246" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[110CYL068.txt]" time="0.046"><properties><property name="score" value="0.6576298096870881" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[HistoryJerusalem.txt]" time="0.100"><properties><property name="score" value="0.8726776002020569" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[rybczynski-ch3.txt]" time="0.221"><properties><property name="score" value="0.5055652280614354" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[117CWL008.txt]" time="0.026"><properties><property name="score" value="0.5901984203896538" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Italy.txt]" time="0.061"><properties><property name="score" value="0.7110388050850792" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[CUP1.txt]" time="0.180"><properties><property name="score" value="0.12956900634661744" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Seedbombing.txt]" time="0.065"><properties><property name="score" value="1.0623688683712202" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Nathans_Bylichka.txt]" time="0.223"><properties><property name="score" value="0.8208399629359924" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Anti-Terrorist.txt]" time="0.066"><properties><property name="score" value="0.5715223782604972" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[Protocol_Regarding_Access.txt]" time="0.089"><properties><property name="score" value="0.10767971369684311" /></properties><failure message="AssertionError: Protocol_Regarding_Access.txt is a human-generated file, misclassified as AI-generated with confidence 0.10767971&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: Protocol_Regarding_Access.txt is a human-generated file, misclassified as AI-generated with confidence 0.10767971
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i198]" time="0.008"><properties><property name="score" value="0.04064809722122975" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i199]" time="0.061"><properties><property name="score" value="0.13971771733896504" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i200]" time="0.064"><properties><property name="score" value="0.12745927270159543" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i201]" time="0.020"><properties><property name="score" value="0.08386487546443355" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i202]" time="0.010"><properties><property name="score" value="0.06651755952430394" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i203]" time="0.029"><properties><property name="score" value="0.05954210669308312" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i204]" time="0.096"><properties><property name="score" value="0.039585277874222516" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i205]" time="0.101"><properties><property name="score" value="0.04865940279876629" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i206]" time="0.018"><properties><property name="score" value="0.09781460387668917" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i207]" time="0.098"><properties><property name="score" value="0.08979734591422188" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i208]" time="0.017"><properties><property name="score" value="0.04988516237735552" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i209]" time="0.016"><properties><property name="score" value="0.08077648732991152" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i210]" time="0.081"><properties><property name="score" value="0.018889924573255085" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i211]" time="0.018"><properties><property name="score" value="0.07703148257730552" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i212]" time="0.057"><properties><property name="score" value="0.08481841308258019" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i213]" time="0.011"><properties><property name="score" value="0.20120273157265278" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i214]" time="0.022"><properties><property name="score" value="0.10051275663233511" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i215]" time="0.051"><properties><property name="score" value="0.12453852076390998" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i216]" time="0.057"><properties><property name="score" value="0.07177386960148491" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i217]" time="0.072"><properties><property name="score" value="0.04506894425508982" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i218]" time="0.008"><properties><property name="score" value="0.02435518465301878" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i219]" time="0.006"><properties><property name="score" value="0.03824250539313234" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i220]" time="0.052"><properties><property name="score" value="0.14566660296990946" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i221]" time="0.009"><properties><property name="score" value="0.13439219917724676" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i222]" time="0.020"><properties><property name="score" value="0.1001477885911361" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i223]" time="0.008"><properties><property name="score" value="0.027658404666175418" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i224]" time="0.007"><properties><property name="score" value="0.10202878525111259" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i225]" time="0.054"><properties><property name="score" value="0.03337801174952685" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i226]" time="0.070"><properties><property name="score" value="0.1549492985274402" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i227]" time="0.032"><properties><property name="score" value="0.0646814634179644" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i228]" time="0.017"><properties><property name="score" value="0.027245295243460976" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i229]" time="0.013"><properties><property name="score" value="0.056930230239155" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i230]" time="0.013"><properties><property name="score" value="0.10066114429882522" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i231]" time="0.008"><properties><property name="score" value="0.030404341042289396" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i232]" time="0.008"><properties><property name="score" value="0.04237361722646549" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i233]" time="0.052"><properties><property name="score" value="0.08311632734970313" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i234]" time="0.056"><properties><property name="score" value="0.08391380646067235" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i235]" time="0.032"><properties><property name="score" value="0.0037522688665653577" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255237 (len: 749) is a human-generated sample, misclassified as AI-generated with confidence 0.00375227&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255237 (len: 749) is a human-generated sample, misclassified as AI-generated with confidence 0.00375227
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_samples[IntroHongKong.txt]" time="0.059"><properties><property name="score" value="0.732062324348981" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[wwf12.txt]" time="0.051"><properties><property name="score" value="0.06997558561247508" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[112C-L013.txt]" time="0.048"><properties><property name="score" value="0.5299177125091167" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[blog-varsity-athletics.txt]" time="0.061"><properties><property name="score" value="0.2913633480380688" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_samples[AMC2.txt]" time="0.051"><properties><property name="score" value="0.3719790511270554" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.00667_generated.txt]" time="0.065"><properties><property name="score" value="0.10116729391624572" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.13472_generated.txt]" time="0.062"><properties><property name="score" value="0.2665926582787563" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.03612_generated.txt]" time="0.084"><properties><property name="score" value="0.15897472379785652" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.00310_generated.txt]" time="0.066"><properties><property name="score" value="0.11124021841307785" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.06230_generated.txt]" time="0.083"><properties><property name="score" value="0.1442628929747188" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.06012_generated.txt]" time="0.077"><properties><property name="score" value="0.19809045784008328" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.01676_generated.txt]" time="0.062"><properties><property name="score" value="0.16449980077240126" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2112.00405_generated.txt]" time="0.063"><properties><property name="score" value="0.06640072521651204" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[gpt_patient_summary.txt]" time="0.012"><properties><property name="score" value="0.09009812040822185" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[gpt_exercise.txt]" time="0.026"><properties><property name="score" value="0.07891762449707684" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.05241_generated.txt]" time="0.062"><properties><property name="score" value="0.3151642745244693" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.12645_generated.txt]" time="0.058"><properties><property name="score" value="0.07220123799119478" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.07793_generated.txt]" time="0.100"><properties><property name="score" value="0.15686817876517725" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[cleaned_2111.00086_generated.txt]" time="0.074"><properties><property name="score" value="0.03880176914827088" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.10340_generated.txt]" time="0.073"><properties><property name="score" value="0.36538219748233347" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.11115_generated.txt]" time="0.059"><properties><property name="score" value="0.15189934846302858" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.15093_generated.txt]" time="0.065"><properties><property name="score" value="0.4582503442161698" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.01515_generated.txt]" time="0.101"><properties><property name="score" value="0.11939919635256535" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.12501_generated.txt]" time="0.067"><properties><property name="score" value="0.13689235826694346" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.02041_generated.txt]" time="0.094"><properties><property name="score" value="0.127001303665735" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.01023_generated.txt]" time="0.059"><properties><property name="score" value="0.07072099944743782" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.03715_generated.txt]" time="0.061"><properties><property name="score" value="0.15688165315287067" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.02760_generated.txt]" time="0.075"><properties><property name="score" value="0.4015923266996735" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.02574_generated.txt]" time="0.067"><properties><property name="score" value="0.15858013574474028" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.15023_generated.txt]" time="0.075"><properties><property name="score" value="0.13709927929628218" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.03320_generated.txt]" time="0.093"><properties><property name="score" value="0.21328961249449108" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[gpt_patient_narrative.txt]" time="0.027"><properties><property name="score" value="0.1591158574437146" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.15317_generated.txt]" time="0.063"><properties><property name="score" value="0.10494158793200592" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.06181_generated.txt]" time="0.075"><properties><property name="score" value="0.16492196967035325" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.03294_generated.txt]" time="0.062"><properties><property name="score" value="0.17150211806013108" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.15436_generated.txt]" time="0.084"><properties><property name="score" value="0.13040705443089864" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.00867_generated.txt]" time="0.068"><properties><property name="score" value="0.22260256499689185" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.05754_generated.txt]" time="0.072"><properties><property name="score" value="0.15329802213651794" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.12341_generated.txt]" time="0.063"><properties><property name="score" value="0.1588212777041594" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.00086_generated.txt]" time="0.067"><properties><property name="score" value="0.003705929850383871" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.02326_generated.txt]" time="0.079"><properties><property name="score" value="0.14772961768726545" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.00514_generated.txt]" time="0.060"><properties><property name="score" value="0.12141026318387056" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.14532_generated.txt]" time="0.073"><properties><property name="score" value="0.09938179851852623" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.07408_generated.txt]" time="0.063"><properties><property name="score" value="0.13731633068714277" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.00035_generated.txt]" time="0.083"><properties><property name="score" value="0.041933833554365994" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.02687_generated.txt]" time="0.061"><properties><property name="score" value="0.15418700697044413" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.05204_generated.txt]" time="0.070"><properties><property name="score" value="0.35810066097922977" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.10478_generated.txt]" time="0.055"><properties><property name="score" value="0.1673216805001987" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.02188_generated.txt]" time="0.088"><properties><property name="score" value="0.09992876488865154" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.10575_generated.txt]" time="0.067"><properties><property name="score" value="0.08380110001508789" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.07525_generated.txt]" time="0.062"><properties><property name="score" value="0.15343341603512345" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.10577_generated.txt]" time="0.060"><properties><property name="score" value="0.12709488306434738" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.15473_generated.txt]" time="0.059"><properties><property name="score" value="0.16005604022156783" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[bing_c_abstract.txt]" time="0.014"><properties><property name="score" value="0.23256193657605895" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.10778_generated.txt]" time="0.067"><properties><property name="score" value="0.19520470910313772" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.13900_generated.txt]" time="0.078"><properties><property name="score" value="0.02929535676057249" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.04416_generated.txt]" time="0.076"><properties><property name="score" value="0.14943594041116337" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.03913_generated.txt]" time="0.067"><properties><property name="score" value="0.2306938885144813" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.01706_generated.txt]" time="0.062"><properties><property name="score" value="0.22862803633219178" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.01243_generated.txt]" time="0.061"><properties><property name="score" value="0.12084361587882952" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.11879_generated.txt]" time="0.061"><properties><property name="score" value="0.0693840760386731" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.10329_generated.txt]" time="0.063"><properties><property name="score" value="0.154645482157196" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.04507_generated.txt]" time="0.077"><properties><property name="score" value="0.10437393577895761" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.03837_generated.txt]" time="0.077"><properties><property name="score" value="0.10426137788375756" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.00572_generated.txt]" time="0.063"><properties><property name="score" value="0.19038013511210197" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.11984_generated.txt]" time="0.062"><properties><property name="score" value="0.13284629488287866" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.06464_generated.txt]" time="0.070"><properties><property name="score" value="0.0743763852392304" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[cgpt_hs_essay2.txt]" time="0.059"><properties><property name="score" value="0.18899611393781435" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.02110_generated.txt]" time="0.083"><properties><property name="score" value="0.1250552307252176" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[cgpt_hs_essay.txt]" time="0.066"><properties><property name="score" value="0.2799628128786896" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.15802_generated.txt]" time="0.066"><properties><property name="score" value="0.20406962127838307" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.11589_generated.txt]" time="0.066"><properties><property name="score" value="0.08181074866302009" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.15724_generated.txt]" time="0.078"><properties><property name="score" value="0.06159829696565181" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.15130_generated.txt]" time="0.082"><properties><property name="score" value="0.025505120430824313" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.06644_generated.txt]" time="0.069"><properties><property name="score" value="0.10508123673550612" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[bard_news.txt]" time="0.027"><properties><property name="score" value="0.28555448914890924" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.01322_generated.txt]" time="0.070"><properties><property name="score" value="0.2563493558179128" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[about_me_short.txt]" time="0.016"><properties><property name="score" value="0.1433578162552962" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[bing_b_song.txt]" time="0.014"><properties><property name="score" value="0.2680560847190261" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.00526_generated.txt]" time="0.071"><properties><property name="score" value="0.21344810503626938" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.12010_generated.txt]" time="0.064"><properties><property name="score" value="0.08032055844605275" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.13317_generated.txt]" time="0.097"><properties><property name="score" value="0.15963170447596634" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[gpt_santos_long.txt]" time="0.046"><properties><property name="score" value="0.3349774826205132" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[gpt_santos_short.txt]" time="0.013"><properties><property name="score" value="0.26121964724495905" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.02643_generated.txt]" time="0.073"><properties><property name="score" value="0.1767332221932668" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[about_me.txt]" time="0.025"><properties><property name="score" value="0.12567781288247426" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.15534_generated.txt]" time="0.061"><properties><property name="score" value="0.05310623272177425" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.07611_generated.txt]" time="0.064"><properties><property name="score" value="0.07233264592779176" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.13658_generated.txt]" time="0.085"><properties><property name="score" value="0.1258906474883026" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.15725_generated.txt]" time="0.078"><properties><property name="score" value="0.7940571075586372" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.01231_generated.txt]" time="0.075"><properties><property name="score" value="0.13297706936420284" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.10319_generated.txt]" time="0.063"><properties><property name="score" value="0.21043336410369226" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.01340_generated.txt]" time="0.066"><properties><property name="score" value="0.12535789221614774" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.07699_generated.txt]" time="0.082"><properties><property name="score" value="0.04234328464014966" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.15705_generated.txt]" time="0.082"><properties><property name="score" value="0.18089691542569594" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.02844_generated.txt]" time="0.068"><properties><property name="score" value="0.9422335633988629" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.04130_generated.txt]" time="0.083"><properties><property name="score" value="0.13583067832779183" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.02259_generated.txt]" time="0.061"><properties><property name="score" value="0.30345696544274514" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.00180_generated.txt]" time="0.076"><properties><property name="score" value="0.12469225055529289" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.04574_generated.txt]" time="0.065"><properties><property name="score" value="0.09212211099027653" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.00808_generated.txt]" time="0.103"><properties><property name="score" value="0.09651974565396311" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.12765_generated.txt]" time="0.064"><properties><property name="score" value="0.13818280766048074" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.13229_generated.txt]" time="0.061"><properties><property name="score" value="0.1146552693758925" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.00554_generated.txt]" time="0.066"><properties><property name="score" value="0.17594252303347602" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.11207_generated.txt]" time="0.059"><properties><property name="score" value="0.08138217485473276" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.00607_generated.txt]" time="0.063"><properties><property name="score" value="0.17440839419042525" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.07267_generated.txt]" time="0.062"><properties><property name="score" value="0.2960343712419187" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.11205_generated.txt]" time="0.079"><properties><property name="score" value="0.11921510087875564" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[bard_internet_abstract.txt]" time="0.046"><properties><property name="score" value="0.39998917604031997" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.03945_generated.txt]" time="0.069"><properties><property name="score" value="0.13333471676045539" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.12383_generated.txt]" time="0.061"><properties><property name="score" value="0.2230628415747967" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2111.02362_generated.txt]" time="0.082"><properties><property name="score" value="0.18268483314272513" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[2110.15799_generated.txt]" time="0.073"><properties><property name="score" value="0.18169743896726911" /></properties></testcase><testcase classname="test_lzma_detect" name="test_llm_sample[bard_abstract.txt]" time="0.015"><properties><property name="score" value="0.5360835751112014" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i0]" time="0.053"><properties><property name="score" value="0.29829211768210573" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i1]" time="0.014"><properties><property name="score" value="0.0936455626249777" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255002 (len: 173) is a human-generated sample, misclassified as AI-generated with confidence 0.09364556&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255002 (len: 173) is a human-generated sample, misclassified as AI-generated with confidence 0.09364556
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i236]" time="0.011"><properties><property name="score" value="0.03721534755705136" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i237]" time="0.055"><properties><property name="score" value="0.10983186181966327" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i238]" time="0.050"><properties><property name="score" value="0.06972508575702205" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i239]" time="0.066"><properties><property name="score" value="0.1450614221525617" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255241 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.14506142&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255241 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.14506142
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i2]" time="0.048"><properties><property name="score" value="0.1509521097426784" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255003 (len: 490) is a human-generated sample, misclassified as AI-generated with confidence 0.15095211&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255003 (len: 490) is a human-generated sample, misclassified as AI-generated with confidence 0.15095211
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i240]" time="0.067"><properties><property name="score" value="0.04262947211048206" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i241]" time="0.066"><properties><property name="score" value="0.0585628562336774" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i242]" time="0.091"><properties><property name="score" value="0.1023653912456787" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i243]" time="0.065"><properties><property name="score" value="0.10606499329171051" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i244]" time="0.009"><properties><property name="score" value="0.01785567315766845" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i245]" time="0.054"><properties><property name="score" value="0.07001057346556139" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i246]" time="0.022"><properties><property name="score" value="0.04251139261653267" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i247]" time="0.051"><properties><property name="score" value="0.03821122635339738" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255249 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.03821123&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255249 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.03821123
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i3]" time="0.018"><properties><property name="score" value="1.3074058868729477" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i4]" time="0.010"><properties><property name="score" value="0.29581312612397825" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255005 (len: 198) is a human-generated sample, misclassified as AI-generated with confidence 0.29581313&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255005 (len: 198) is a human-generated sample, misclassified as AI-generated with confidence 0.29581313
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i248]" time="0.075"><properties><property name="score" value="0.04476121150137153" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i249]" time="0.061"><properties><property name="score" value="0.1403030090284212" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i250]" time="0.032"><properties><property name="score" value="0.06403532628123912" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255252 (len: 705) is a human-generated sample, misclassified as AI-generated with confidence 0.06403533&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255252 (len: 705) is a human-generated sample, misclassified as AI-generated with confidence 0.06403533
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i5]" time="0.023"><properties><property name="score" value="0.1844374786750258" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i6]" time="0.011"><properties><property name="score" value="0.6041080419381117" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i7]" time="0.015"><properties><property name="score" value="1.360350591812609" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i8]" time="0.054"><properties><property name="score" value="0.11011824360084022" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i9]" time="0.021"><properties><property name="score" value="0.679269389823628" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i10]" time="0.051"><properties><property name="score" value="0.9066848060801935" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i11]" time="0.068"><properties><property name="score" value="0.4876731040583254" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i12]" time="0.034"><properties><property name="score" value="0.21519501308250866" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i13]" time="0.050"><properties><property name="score" value="0.5543394813119865" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i14]" time="0.065"><properties><property name="score" value="0.7103646659874138" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i15]" time="0.067"><properties><property name="score" value="0.7264590899276335" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i16]" time="0.058"><properties><property name="score" value="0.9173715488847712" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i17]" time="0.025"><properties><property name="score" value="0.39863590927285975" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i18]" time="0.061"><properties><property name="score" value="0.664964225090034" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i19]" time="0.071"><properties><property name="score" value="0.1159900071643627" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255026 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.11599001&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255026 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.11599001
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i251]" time="0.017"><properties><property name="score" value="0.057028891051088805" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i252]" time="0.009"><properties><property name="score" value="0.01675958596172311" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i253]" time="0.064"><properties><property name="score" value="0.08024868201492527" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i254]" time="0.064"><properties><property name="score" value="0.10331669848106254" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i255]" time="0.011"><properties><property name="score" value="0.020809458376880308" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i256]" time="0.016"><properties><property name="score" value="0.22192419476519398" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i257]" time="0.073"><properties><property name="score" value="0.09670419540596355" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i258]" time="0.084"><properties><property name="score" value="0.018097681051861523" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i259]" time="0.083"><properties><property name="score" value="0.05159190350078752" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i260]" time="0.061"><properties><property name="score" value="0.07022437751142244" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i261]" time="0.018"><properties><property name="score" value="0.023270601597063123" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i262]" time="0.053"><properties><property name="score" value="0.06123653894052012" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i263]" time="0.052"><properties><property name="score" value="0.010047907773608629" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255265 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.01004791&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255265 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.01004791
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i20]" time="0.058"><properties><property name="score" value="0.19671090629426458" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255027 (len: 694) is a human-generated sample, misclassified as AI-generated with confidence 0.19671091&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255027 (len: 694) is a human-generated sample, misclassified as AI-generated with confidence 0.19671091
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i264]" time="0.010"><properties><property name="score" value="0.05291566462207742" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i265]" time="0.057"><properties><property name="score" value="0.12172759328373003" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i266]" time="0.053"><properties><property name="score" value="0.07872413627033391" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i267]" time="0.063"><properties><property name="score" value="0.10708986929139863" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i268]" time="0.024"><properties><property name="score" value="0.07310053247487339" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i269]" time="0.021"><properties><property name="score" value="0.0473644670090176" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i270]" time="0.069"><properties><property name="score" value="0.05164215524143325" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i271]" time="0.058"><properties><property name="score" value="0.07232812397501653" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i272]" time="0.021"><properties><property name="score" value="0.13036715573446572" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i273]" time="0.014"><properties><property name="score" value="0.07463429886733763" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i274]" time="0.016"><properties><property name="score" value="0.15260563136737515" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i275]" time="0.022"><properties><property name="score" value="0.1290565296482701" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i276]" time="0.021"><properties><property name="score" value="0.016326489624133664" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255278 (len: 369) is a human-generated sample, misclassified as AI-generated with confidence 0.01632649&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255278 (len: 369) is a human-generated sample, misclassified as AI-generated with confidence 0.01632649
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i21]" time="0.048"><properties><property name="score" value="0.2812184034674166" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255028 (len: 512) is a human-generated sample, misclassified as AI-generated with confidence 0.2812184&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255028 (len: 512) is a human-generated sample, misclassified as AI-generated with confidence 0.2812184
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i277]" time="0.070"><properties><property name="score" value="0.044562340709108354" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i278]" time="0.024"><properties><property name="score" value="0.045468347872815784" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i279]" time="0.016"><properties><property name="score" value="0.060761875543069443" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i280]" time="0.072"><properties><property name="score" value="0.13247707711355955" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i281]" time="0.071"><properties><property name="score" value="0.13349769749980006" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i282]" time="0.078"><properties><property name="score" value="0.1209421672795068" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i283]" time="0.017"><properties><property name="score" value="0.1070197919561311" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i284]" time="0.083"><properties><property name="score" value="0.08206083020478047" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i285]" time="0.033"><properties><property name="score" value="0.10521375251233742" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i286]" time="0.013"><properties><property name="score" value="0.13110804159749967" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i287]" time="0.078"><properties><property name="score" value="0.1344733534245389" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i288]" time="0.076"><properties><property name="score" value="0.1250200108281232" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i289]" time="0.067"><properties><property name="score" value="0.10703504599974448" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i290]" time="0.015"><properties><property name="score" value="0.15919150197471477" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i291]" time="0.091"><properties><property name="score" value="0.034998286353871116" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i292]" time="0.017"><properties><property name="score" value="0.038689935180891144" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i293]" time="0.087"><properties><property name="score" value="0.07785489105462047" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i294]" time="0.016"><properties><property name="score" value="0.09887723730396258" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i295]" time="0.029"><properties><property name="score" value="0.03380442717606425" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255297 (len: 460) is a human-generated sample, misclassified as AI-generated with confidence 0.03380443&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255297 (len: 460) is a human-generated sample, misclassified as AI-generated with confidence 0.03380443
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i22]" time="0.026"><properties><property name="score" value="0.8106792980107635" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i23]" time="0.048"><properties><property name="score" value="0.8184550436154033" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i24]" time="0.069"><properties><property name="score" value="0.7067828849700031" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i25]" time="0.018"><properties><property name="score" value="0.1722479302765645" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i26]" time="0.055"><properties><property name="score" value="0.2976267626889961" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i27]" time="0.036"><properties><property name="score" value="0.22570020516188283" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i28]" time="0.021"><properties><property name="score" value="0.8410607272330561" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i29]" time="0.014"><properties><property name="score" value="0.26986376011489677" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255039 (len: 203) is a human-generated sample, misclassified as AI-generated with confidence 0.26986376&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255039 (len: 203) is a human-generated sample, misclassified as AI-generated with confidence 0.26986376
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i296]" time="0.020"><properties><property name="score" value="0.12825131351074104" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i297]" time="0.077"><properties><property name="score" value="0.021001766210502728" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i298]" time="0.088"><properties><property name="score" value="0.10274767183301034" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i299]" time="0.038"><properties><property name="score" value="0.09939809099584507" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i300]" time="0.071"><properties><property name="score" value="0.062421931906279135" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i301]" time="0.087"><properties><property name="score" value="0.05235775613841629" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i302]" time="0.014"><properties><property name="score" value="0.03475330820619238" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i303]" time="0.014"><properties><property name="score" value="0.18434223908029845" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i304]" time="0.015"><properties><property name="score" value="0.1477390917221877" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i305]" time="0.084"><properties><property name="score" value="0.15302981952604838" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i306]" time="0.030"><properties><property name="score" value="0.09791664192636959" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i307]" time="0.082"><properties><property name="score" value="0.06515776963044054" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i308]" time="0.086"><properties><property name="score" value="0.1203459389735214" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i309]" time="0.033"><properties><property name="score" value="0.07705016849572013" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255312 (len: 641) is a human-generated sample, misclassified as AI-generated with confidence 0.07705017&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255312 (len: 641) is a human-generated sample, misclassified as AI-generated with confidence 0.07705017
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i30]" time="0.059"><properties><property name="score" value="0.4446386743358621" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i31]" time="0.066"><properties><property name="score" value="0.280114990351558" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i32]" time="0.063"><properties><property name="score" value="0.1852895644209318" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255042 (len: 531) is a human-generated sample, misclassified as AI-generated with confidence 0.18528956&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255042 (len: 531) is a human-generated sample, misclassified as AI-generated with confidence 0.18528956
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i310]" time="0.037"><properties><property name="score" value="0.0821225696715686" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i311]" time="0.022"><properties><property name="score" value="0.11174732388828379" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i312]" time="0.033"><properties><property name="score" value="0.08320262495460806" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i313]" time="0.018"><properties><property name="score" value="0.1298195119338874" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i314]" time="0.019"><properties><property name="score" value="0.0355363049225621" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i315]" time="0.100"><properties><property name="score" value="0.10007220393226403" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i316]" time="0.097"><properties><property name="score" value="0.09878781529918129" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i317]" time="0.020"><properties><property name="score" value="0.05916780367937746" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i318]" time="0.036"><properties><property name="score" value="0.19864020040470287" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i319]" time="0.079"><properties><property name="score" value="0.15630072042842666" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i320]" time="0.090"><properties><property name="score" value="0.03964450321874048" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i321]" time="0.087"><properties><property name="score" value="0.11228407656255764" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i322]" time="0.021"><properties><property name="score" value="0.10897743548957672" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i323]" time="0.020"><properties><property name="score" value="0.03460139922910055" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i324]" time="0.095"><properties><property name="score" value="0.05293166939943489" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i325]" time="0.096"><properties><property name="score" value="0.10810693865957306" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i326]" time="0.017"><properties><property name="score" value="0.056050291439518896" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i327]" time="0.090"><properties><property name="score" value="0.07139164865821539" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i328]" time="0.089"><properties><property name="score" value="0.06645432363124061" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i329]" time="0.036"><properties><property name="score" value="0.07853967509262685" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i330]" time="0.017"><properties><property name="score" value="0.03007922848971556" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i331]" time="0.017"><properties><property name="score" value="0.05065067213986807" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i332]" time="0.035"><properties><property name="score" value="0.12076510611819988" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i333]" time="0.021"><properties><property name="score" value="0.12820366914127468" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i334]" time="0.016"><properties><property name="score" value="0.0938599388789707" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i335]" time="0.022"><properties><property name="score" value="0.10563923831061195" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i336]" time="0.096"><properties><property name="score" value="0.02961542763970453" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255339 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.02961543&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255339 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.02961543
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i33]" time="0.020"><properties><property name="score" value="0.15591184213412101" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255044 (len: 243) is a human-generated sample, misclassified as AI-generated with confidence 0.15591184&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255044 (len: 243) is a human-generated sample, misclassified as AI-generated with confidence 0.15591184
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i337]" time="0.101"><properties><property name="score" value="0.07268058200620808" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i338]" time="0.034"><properties><property name="score" value="0.07648300874437616" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i339]" time="0.035"><properties><property name="score" value="0.06345730303442643" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i340]" time="0.019"><properties><property name="score" value="0.052046785704140675" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i341]" time="0.034"><properties><property name="score" value="0.1494716524480996" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i342]" time="0.024"><properties><property name="score" value="0.001486500579650496" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i343]" time="0.043"><properties><property name="score" value="0.08513587730642826" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255346 (len: 593) is a human-generated sample, misclassified as AI-generated with confidence 0.08513588&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255346 (len: 593) is a human-generated sample, misclassified as AI-generated with confidence 0.08513588
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i34]" time="0.060"><properties><property name="score" value="0.4192314662291974" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i35]" time="0.053"><properties><property name="score" value="0.9073179671391096" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i36]" time="0.063"><properties><property name="score" value="0.6237778115469106" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i37]" time="0.053"><properties><property name="score" value="0.2933308982308924" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i38]" time="0.026"><properties><property name="score" value="0.23153852198753988" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255049 (len: 424) is a human-generated sample, misclassified as AI-generated with confidence 0.23153852&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255049 (len: 424) is a human-generated sample, misclassified as AI-generated with confidence 0.23153852
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i344]" time="0.036"><properties><property name="score" value="0.1469408301014431" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i345]" time="0.102"><properties><property name="score" value="0.04882564724565437" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i346]" time="0.019"><properties><property name="score" value="0.06679729831061687" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i347]" time="0.035"><properties><property name="score" value="0.06993100367140911" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i348]" time="0.018"><properties><property name="score" value="0.010821817329270678" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i349]" time="0.038"><properties><property name="score" value="0.08091373685485062" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i350]" time="0.100"><properties><property name="score" value="0.03041829428972203" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i351]" time="0.021"><properties><property name="score" value="0.03088271574231971" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i352]" time="0.090"><properties><property name="score" value="0.06616468386995396" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255355 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.06616468&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255355 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.06616468
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i39]" time="0.055"><properties><property name="score" value="0.735018297399237" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i40]" time="0.059"><properties><property name="score" value="0.20433647763521412" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255051 (len: 891) is a human-generated sample, misclassified as AI-generated with confidence 0.20433648&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255051 (len: 891) is a human-generated sample, misclassified as AI-generated with confidence 0.20433648
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i353]" time="0.038"><properties><property name="score" value="0.010632149150735145" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255356 (len: 326) is a human-generated sample, misclassified as AI-generated with confidence 0.01063215&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255356 (len: 326) is a human-generated sample, misclassified as AI-generated with confidence 0.01063215
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i41]" time="0.022"><properties><property name="score" value="0.7164398991476861" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i42]" time="0.028"><properties><property name="score" value="1.0341092884764147" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i43]" time="0.067"><properties><property name="score" value="0.6706066389121301" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i44]" time="0.058"><properties><property name="score" value="0.38576260653840067" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i45]" time="0.063"><properties><property name="score" value="0.6630139850711037" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i46]" time="0.055"><properties><property name="score" value="0.46867711921499966" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i47]" time="0.037"><properties><property name="score" value="0.5384769899553221" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i48]" time="0.075"><properties><property name="score" value="0.9649441472234233" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i49]" time="0.071"><properties><property name="score" value="0.6294255426586681" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i0]" time="0.016"><properties><property name="score" value="0.5747089671988004" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255001 (text: What is this?  This is a neat little plugin to giv) is an LLM-generated sample, misclassified as human-generated with confidence 0.57470897&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255001 (text: What is this?  This is a neat little plugin to giv) is an LLM-generated sample, misclassified as human-generated with confidence 0.57470897
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i1]" time="0.059"><properties><property name="score" value="0.7394625297881263" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255002 (text: People carry a memorial sign to honor Robert Paxto) is an LLM-generated sample, misclassified as human-generated with confidence 0.73946253&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255002 (text: People carry a memorial sign to honor Robert Paxto) is an LLM-generated sample, misclassified as human-generated with confidence 0.73946253
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i2]" time="0.025"><properties><property name="score" value="0.12802778034834428" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255003 (text: Google is launching a new scholarship campaign for) is an LLM-generated sample, misclassified as human-generated with confidence 0.12802778&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255003 (text: Google is launching a new scholarship campaign for) is an LLM-generated sample, misclassified as human-generated with confidence 0.12802778
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i3]" time="0.058"><properties><property name="score" value="0.6873756304713874" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255004 (text: Authored by: Kristin Scavuzzo, Managing Director o) is an LLM-generated sample, misclassified as human-generated with confidence 0.68737563&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255004 (text: Authored by: Kristin Scavuzzo, Managing Director o) is an LLM-generated sample, misclassified as human-generated with confidence 0.68737563
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i4]" time="0.048"><properties><property name="score" value="1.4644040332755237" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255005 (text: &gt;at some point the MPP team stop standardizing pee) is an LLM-generated sample, misclassified as human-generated with confidence 1.46440403&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255005 (text: &gt;at some point the MPP team stop standardizing pee) is an LLM-generated sample, misclassified as human-generated with confidence 1.46440403
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i5]" time="0.038"><properties><property name="score" value="0.31445391909268317" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255006 (text: ISIS is breeding from among the refugees forced ou) is an LLM-generated sample, misclassified as human-generated with confidence 0.31445392&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255006 (text: ISIS is breeding from among the refugees forced ou) is an LLM-generated sample, misclassified as human-generated with confidence 0.31445392
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i6]" time="0.059"><properties><property name="score" value="0.23245211573891256" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255007 (text: More than a third of population as of recently, Ma) is an LLM-generated sample, misclassified as human-generated with confidence 0.23245212&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255007 (text: More than a third of population as of recently, Ma) is an LLM-generated sample, misclassified as human-generated with confidence 0.23245212
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i7]" time="0.049"><properties><property name="score" value="0.7450008571483849" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255008 (text: © Provided by Texas State Historical Association' ) is an LLM-generated sample, misclassified as human-generated with confidence 0.74500086&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255008 (text: © Provided by Texas State Historical Association' ) is an LLM-generated sample, misclassified as human-generated with confidence 0.74500086
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i8]" time="0.058"><properties><property name="score" value="0.4550383866430363" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255009 (text: Mongolia has banned the sale and consumption of me) is an LLM-generated sample, misclassified as human-generated with confidence 0.45503839&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255009 (text: Mongolia has banned the sale and consumption of me) is an LLM-generated sample, misclassified as human-generated with confidence 0.45503839
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i9]" time="0.053"><properties><property name="score" value="0.8296108281736029" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255010 (text: Bots and AI are everywhere these days, and it's no) is an LLM-generated sample, misclassified as human-generated with confidence 0.82961083&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255010 (text: Bots and AI are everywhere these days, and it's no) is an LLM-generated sample, misclassified as human-generated with confidence 0.82961083
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i10]" time="0.041"><properties><property name="score" value="0.6199065176327179" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i11]" time="0.063"><properties><property name="score" value="0.7859802287694673" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255012 (text: Dr. Karl Widerquist says his study &quot;opens up an in) is an LLM-generated sample, misclassified as human-generated with confidence 0.78598023&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255012 (text: Dr. Karl Widerquist says his study "opens up an in) is an LLM-generated sample, misclassified as human-generated with confidence 0.78598023
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i12]" time="0.080"><properties><property name="score" value="1.0894579392178305" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255013 (text: 2. Every nature, forest, even in the front pasture) is an LLM-generated sample, misclassified as human-generated with confidence 1.08945794&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255013 (text: 2. Every nature, forest, even in the front pasture) is an LLM-generated sample, misclassified as human-generated with confidence 1.08945794
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i13]" time="0.020"><properties><property name="score" value="0.6665580718309694" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255014 (text: Fact: The Justice 4 Steph showing can be good tele) is an LLM-generated sample, misclassified as human-generated with confidence 0.66655807&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255014 (text: Fact: The Justice 4 Steph showing can be good tele) is an LLM-generated sample, misclassified as human-generated with confidence 0.66655807
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i14]" time="0.048"><properties><property name="score" value="0.28114683893502396" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255015 (text: CALGARY — Tory MP Mac Harb resigned from the Conse) is an LLM-generated sample, misclassified as human-generated with confidence 0.28114684&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255015 (text: CALGARY — Tory MP Mac Harb resigned from the Conse) is an LLM-generated sample, misclassified as human-generated with confidence 0.28114684
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i15]" time="0.059"><properties><property name="score" value="0.7333037239752871" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255016 (text: SERVO, Jupiter - Since the discovery of the swirli) is an LLM-generated sample, misclassified as human-generated with confidence 0.73330372&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255016 (text: SERVO, Jupiter - Since the discovery of the swirli) is an LLM-generated sample, misclassified as human-generated with confidence 0.73330372
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i16]" time="0.068"><properties><property name="score" value="1.280155437691319" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255018 (text: Cipei-Rochel  Installerete Generasseanunterzeichnu) is an LLM-generated sample, misclassified as human-generated with confidence 1.28015544&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255018 (text: Cipei-Rochel  Installerete Generasseanunterzeichnu) is an LLM-generated sample, misclassified as human-generated with confidence 1.28015544
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i17]" time="0.022"><properties><property name="score" value="0.3606658836891874" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i18]" time="0.057"><properties><property name="score" value="0.7705397839151584" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255020 (text: Albert Einstein, one of the world's most famous sc) is an LLM-generated sample, misclassified as human-generated with confidence 0.77053978&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255020 (text: Albert Einstein, one of the world's most famous sc) is an LLM-generated sample, misclassified as human-generated with confidence 0.77053978
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i19]" time="0.060"><properties><property name="score" value="0.7983287107058312" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255021 (text: Barney: They weren't really that opposed to me if ) is an LLM-generated sample, misclassified as human-generated with confidence 0.79832871&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255021 (text: Barney: They weren't really that opposed to me if ) is an LLM-generated sample, misclassified as human-generated with confidence 0.79832871
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i20]" time="0.054"><properties><property name="score" value="0.3150988634771229" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i21]" time="0.055"><properties><property name="score" value="0.9838118628334724" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255023 (text: My Flamin' Hot Chocolate Rd2o   ❮❮ Newer Download ) is an LLM-generated sample, misclassified as human-generated with confidence 0.98381186&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255023 (text: My Flamin' Hot Chocolate Rd2o   ❮❮ Newer Download ) is an LLM-generated sample, misclassified as human-generated with confidence 0.98381186
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i22]" time="0.057"><properties><property name="score" value="0.6945611225753434" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255024 (text: I wake up with a fever. This isn't good. No, no, n) is an LLM-generated sample, misclassified as human-generated with confidence 0.69456112&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255024 (text: I wake up with a fever. This isn't good. No, no, n) is an LLM-generated sample, misclassified as human-generated with confidence 0.69456112
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i23]" time="0.031"><properties><property name="score" value="0.5227485311408986" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255025 (text: Convicted cop killer should spend more time in jai) is an LLM-generated sample, misclassified as human-generated with confidence 0.52274853&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255025 (text: Convicted cop killer should spend more time in jai) is an LLM-generated sample, misclassified as human-generated with confidence 0.52274853
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i24]" time="0.070"><properties><property name="score" value="0.8864005997061553" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255026 (text: Issue Width: 500px  Posted on: 01/25/2013  Twitter) is an LLM-generated sample, misclassified as human-generated with confidence 0.8864006&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255026 (text: Issue Width: 500px  Posted on: 01/25/2013  Twitter) is an LLM-generated sample, misclassified as human-generated with confidence 0.8864006
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i25]" time="0.024"><properties><property name="score" value="0.4473284661710708" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i26]" time="0.084"><properties><property name="score" value="0.5863753489827706" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255028 (text: I know this is a book about Holocaust survivors, b) is an LLM-generated sample, misclassified as human-generated with confidence 0.58637535&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255028 (text: I know this is a book about Holocaust survivors, b) is an LLM-generated sample, misclassified as human-generated with confidence 0.58637535
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i27]" time="0.020"><properties><property name="score" value="0.6863817359076263" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255029 (text: We are pleased to announce that our previously ann) is an LLM-generated sample, misclassified as human-generated with confidence 0.68638174&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255029 (text: We are pleased to announce that our previously ann) is an LLM-generated sample, misclassified as human-generated with confidence 0.68638174
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i28]" time="0.026"><properties><property name="score" value="1.0161670731977657" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255030 (text: Stitch Fuse Box by Kink Novaphor Supporting Kinky ) is an LLM-generated sample, misclassified as human-generated with confidence 1.01616707&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255030 (text: Stitch Fuse Box by Kink Novaphor Supporting Kinky ) is an LLM-generated sample, misclassified as human-generated with confidence 1.01616707
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i29]" time="0.058"><properties><property name="score" value="0.8179028633525653" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255031 (text: A small group of Soviet &quot;Oh-On's&quot; fled North Vietn) is an LLM-generated sample, misclassified as human-generated with confidence 0.81790286&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255031 (text: A small group of Soviet "Oh-On's" fled North Vietn) is an LLM-generated sample, misclassified as human-generated with confidence 0.81790286
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i30]" time="0.051"><properties><property name="score" value="0.2557140884569635" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255033 (text: PALLBRIDGE, Mass. -- The New Hampshire Department ) is an LLM-generated sample, misclassified as human-generated with confidence 0.25571409&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255033 (text: PALLBRIDGE, Mass. -- The New Hampshire Department ) is an LLM-generated sample, misclassified as human-generated with confidence 0.25571409
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i31]" time="0.049"><properties><property name="score" value="0.5014843638383448" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255035 (text: Speaker Paul Ryan Paul Davis RyanHouse passes reso) is an LLM-generated sample, misclassified as human-generated with confidence 0.50148436&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255035 (text: Speaker Paul Ryan Paul Davis RyanHouse passes reso) is an LLM-generated sample, misclassified as human-generated with confidence 0.50148436
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i32]" time="0.053"><properties><property name="score" value="1.0960667473809833" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255036 (text: To revisit an idea that a writer once put forth in) is an LLM-generated sample, misclassified as human-generated with confidence 1.09606675&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255036 (text: To revisit an idea that a writer once put forth in) is an LLM-generated sample, misclassified as human-generated with confidence 1.09606675
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i33]" time="0.102"><properties><property name="score" value="0.68772860237859" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255037 (text: Jimmy Pastore worked his way up from the lightweig) is an LLM-generated sample, misclassified as human-generated with confidence 0.6877286&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255037 (text: Jimmy Pastore worked his way up from the lightweig) is an LLM-generated sample, misclassified as human-generated with confidence 0.6877286
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i34]" time="0.014"><properties><property name="score" value="0.36032573820427016" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i35]" time="0.037"><properties><property name="score" value="0.07067396119847524" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i36]" time="0.075"><properties><property name="score" value="0.806656333057909" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255040 (text: http://tvtropes.org/pmwiki/pmwiki.php/Main/Dunes  ) is an LLM-generated sample, misclassified as human-generated with confidence 0.80665633&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255040 (text: http://tvtropes.org/pmwiki/pmwiki.php/Main/Dunes  ) is an LLM-generated sample, misclassified as human-generated with confidence 0.80665633
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i37]" time="0.034"><properties><property name="score" value="1.1667494841276493" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255041 (text: You have a choice when your four year old gets to ) is an LLM-generated sample, misclassified as human-generated with confidence 1.16674948&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255041 (text: You have a choice when your four year old gets to ) is an LLM-generated sample, misclassified as human-generated with confidence 1.16674948
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i38]" time="0.070"><properties><property name="score" value="0.766933002494341" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255042 (text: So why the disregard for danger? Contrary to popul) is an LLM-generated sample, misclassified as human-generated with confidence 0.766933&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255042 (text: So why the disregard for danger? Contrary to popul) is an LLM-generated sample, misclassified as human-generated with confidence 0.766933
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i39]" time="0.016"><properties><property name="score" value="0.7635362301405824" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255043 (text: review by Carter118 on 6/26/2017  Recommend  This ) is an LLM-generated sample, misclassified as human-generated with confidence 0.76353623&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255043 (text: review by Carter118 on 6/26/2017  Recommend  This ) is an LLM-generated sample, misclassified as human-generated with confidence 0.76353623
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i40]" time="0.073"><properties><property name="score" value="0.4634385266315544" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255044 (text: AryanTheFascist writes...  I am sorry to hear it M) is an LLM-generated sample, misclassified as human-generated with confidence 0.46343853&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255044 (text: AryanTheFascist writes...  I am sorry to hear it M) is an LLM-generated sample, misclassified as human-generated with confidence 0.46343853
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i41]" time="0.071"><properties><property name="score" value="0.25626607990368866" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255045 (text: Fear that mirrors a changing political dynamic in ) is an LLM-generated sample, misclassified as human-generated with confidence 0.25626608&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255045 (text: Fear that mirrors a changing political dynamic in ) is an LLM-generated sample, misclassified as human-generated with confidence 0.25626608
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i42]" time="0.064"><properties><property name="score" value="0.9316689103661494" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255046 (text: By Vigilant Citizen Staff   Just a restaurant that) is an LLM-generated sample, misclassified as human-generated with confidence 0.93166891&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255046 (text: By Vigilant Citizen Staff   Just a restaurant that) is an LLM-generated sample, misclassified as human-generated with confidence 0.93166891
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i43]" time="0.030"><properties><property name="score" value="0.7301406539053923" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255047 (text: 16 #geekyvimcovers submitted after watching this v) is an LLM-generated sample, misclassified as human-generated with confidence 0.73014065&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255047 (text: 16 #geekyvimcovers submitted after watching this v) is an LLM-generated sample, misclassified as human-generated with confidence 0.73014065
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i44]" time="0.059"><properties><property name="score" value="0.9748955768255965" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255048 (text: Watch a video recording from the Oct. 29 La Plata ) is an LLM-generated sample, misclassified as human-generated with confidence 0.97489558&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255048 (text: Watch a video recording from the Oct. 29 La Plata ) is an LLM-generated sample, misclassified as human-generated with confidence 0.97489558
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i45]" time="0.027"><properties><property name="score" value="0.22555194417506197" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255049 (text: Watch Rick Santorum's full interview with Glenn Be) is an LLM-generated sample, misclassified as human-generated with confidence 0.22555194&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255049 (text: Watch Rick Santorum's full interview with Glenn Be) is an LLM-generated sample, misclassified as human-generated with confidence 0.22555194
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i46]" time="0.064"><properties><property name="score" value="0.9053171770180279" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255050 (text: Passengers and regulators round up illegal drugs d) is an LLM-generated sample, misclassified as human-generated with confidence 0.90531718&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255050 (text: Passengers and regulators round up illegal drugs d) is an LLM-generated sample, misclassified as human-generated with confidence 0.90531718
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i47]" time="0.015"><properties><property name="score" value="0.6485415219582014" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255051 (text: Former Texas governor Rick Perry will sit down wit) is an LLM-generated sample, misclassified as human-generated with confidence 0.64854152&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255051 (text: Former Texas governor Rick Perry will sit down wit) is an LLM-generated sample, misclassified as human-generated with confidence 0.64854152
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i48]" time="0.057"><properties><property name="score" value="0.9221870811583655" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255052 (text: Beloved writer and poet J.R.R. Tolkien once sugges) is an LLM-generated sample, misclassified as human-generated with confidence 0.92218708&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255052 (text: Beloved writer and poet J.R.R. Tolkien once sugges) is an LLM-generated sample, misclassified as human-generated with confidence 0.92218708
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i49]" time="0.064"><properties><property name="score" value="0.7624104846322255" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255053 (text: 2000 Year Oro Jackson  The tomb was previously ide) is an LLM-generated sample, misclassified as human-generated with confidence 0.76241048&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255053 (text: 2000 Year Oro Jackson  The tomb was previously ide) is an LLM-generated sample, misclassified as human-generated with confidence 0.76241048
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Glacier Ridge Christian School\n\nGlacier Ridge Christian School is a private Christian school in Johnstown, Ohio. It was founded in the fall of 1999 by Gary and Tammy Smith.\n\nThe school started with just 13 students in grades 5-8. By the end of the first year, the school had grown to 65 students in those same grades. As the school has continued to grow, so has the number of teachers.\n\nBy the year 2009, Glacier Ridge had expanded to offer a preschool through 12th grade education, and the school had grown to almost 200 students. In the fall of 2013, the enrollment was over 300 students.\n\nThe school started with the first graduating class of 2002. They are now on their 11th graduating class, and the school's graduates continue to receive acceptance to universities across the United States. The school offers several Advanced Placement classes.\n\nGlacier Ridge has had its students participate in various competitions and competitions. Most notable are the sports teams, especially their basketball teams, which have qualified for state competition multiple times. The school has also had various music groups participate in state and regional competitions, and in the 2010-2011 school year, the school sent students to participate in an International Children's Games competition in Romania.\n\nIn the 2010-2011 school year, a group of the school's teachers were sent to Haiti to help establish a new orphanage, &quot;Eli's Children's House&quot;, which was dedicated to Eli Pierre, who had died in a motorcycle accident. They were instrumental in bringing the orphanage into the community by helping establish new programs for the children, including English as a Second Language, Computer Lab, and Physical Education classes. The teachers assisted in the education of the orphans as well as working with the staff on the property.\n\nThe school has also taken various trips over the years, including trips to Virginia and New York City. The school also hosted an educational trip to Washington D.C. for one of their graduating classes.\n\nThe school has participated in various community events and service projects. One of the most notable is their annual &quot;Twilight Prom&quot; that is held each spring. Each year, the school hosts a prom for local high school students who have special needs. The students' dates are typically their own siblings.\n\nOn March 21, 2011, a fire destroyed the school's auditorium, causing over $1 million in damages. After the school received an outpouring of support, the school was able to begin construction on a new auditorium in time for the next school year.\n\nIn 2015, Glacier Ridge Christian School became part of the Christian School Association of Northeast Ohio.\n\nThe School is affiliated with the Association of Christian Schools International, the Ohio Christian Education Association, and the Christian School Association of Northeast Ohio.\n\n]" time="0.058"><properties><property name="score" value="0.05387341804247159" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[2003 NCAA Division I-AA football season\n\nThe 2003 NCAA Division I-AA football season, part of college football in the United States organized by the National Collegiate Athletic Association at the Division I Football Championship Subdivision level, began on September 6, 2003, and concluded with the National Championship Game of the NCAA Division I Football Championship on December 21, 2003, at Finley Stadium in Chattanooga, Tennessee. The Georgia Southern Eagles won their first I-AA championship, defeating the Furman Paladins by a final score of 27\u221210. \n\n&lt;nowiki&gt;*&lt;/nowiki&gt; &quot;Denotes host institution&quot;\n\n&lt;nowiki&gt;*&lt;/nowiki&gt; &quot;Denotes host institution&quot;\n\n&lt;nowiki&gt;*&lt;/nowiki&gt; &quot;Denotes host institution&quot;\n]" time="0.026"><properties><property name="score" value="0.13355222586508564" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Adiadochokinesia\n\nAdiadochokinesia, or more fully dysdiadochokinesia, is a neurologic symptom, described as the inability to perform rapid alternating movements, as seen in ataxia.\n\nExamples of this are \n\nA classic test is &quot;the finger to nose test&quot;. The examiner repeatedly taps the patient's finger with the examiner's finger and asks the patient to reciprocate in the same rhythm.\n\nThis should be distinguished from dysdiadochokinesia, a progressive condition involving more or less the same symptoms.\n]" time="0.018"><properties><property name="score" value="0.1260299905372908" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Akyem Kotoku\n\nAkyem Kotoku is one of the five traditional districts of the Ashanti Empire. The people are collectively known as Akyem Kotokus.\n\nAkyem Kotoku lies in the Eastern region of Ashanti and is bounded by the East Mamprusi to the East, Nkwantanan to the North, Ashanti to the West and Ahanta West to the South. It is part of Ashanti Region. The area is mainly mountainous with some rich valleys and hills. Akyem Kotoku is known as the &quot;Area of Plenty&quot;.\n\nThe people are mainly farmers and fishermen. Its capital is Akyem Kotoku. It has two paramountcies, Buokrom and Kumasi. Buokrom, the newest paramountcy, was established in 1975. The Akyem Kotoku Traditional Council has nine chiefs.\n\nThe major towns are Abokobi, Domenase, Akropong, and Nsuaem. Other settlements include Aduanful, Boankraful, Denkyemerem, Anhwiaso, Gyankroma, Kwabenya, Nananom, Obofram, Okwensu, Ofoase, Nkoranza, Nsutam, Aduaso, Adom, Ofram, Odomankoma, Asafo, Anyigbe, Obofram, Kwabenya, Denkyemerem, Odumase, Nsutam, Aduaso, Nsuaem, Nsuaso, Tafo-Nsuaem, Manhyia, Boaso, Adom, Ofram, Anhwiaso, Adoase, Anyigbe, Boankraful, Okwensu, Ofoase, Nkoranza, Akropong, Gyankroma, Domenase, Egyir, Afram, Denkyerem, Nananom, Akuapem South, Asafo, Aboadze and Koma.\n\nThe Akyem Kotoku Traditional Council has nine paramountcies.\n\nThe Akyem Kotoku people have two coats of arms, one to use at home and the other for international use.\n\nThe arms used at home is a gold shield with a black anchor in the center. On the shield is a white dove with a golden olive branch in its beak. The dove is supported by two black wild dogs facing each other.\n\nThe shield is surmounted by a gold mural crown with five points and black stripes.\n\nThe shield is supported by two black lions standing on a white scroll. The scroll bears the motto: &quot;MUSTER FAITHFULNESS&quot; (Cape Coast Latin: &quot;Fidelitas Summa Est&quot;).\n\nThe arms used outside the country is a gold shield with a black ram's head in the center. On the shield is a green laurel wreath.\n\nThe shield is surmounted by a gold mural crown with five points and black stripes.\n\nThe shield is supported by two black and white wild dogs standing on a white scroll. The scroll bears the motto: &quot;MUSTER FAITHFULNESS&quot; (Cape Coast Latin: &quot;Fidelitas Summa Est&quot;).\n\nThe two dogs represent vigilance. The two lions in the home arms represent fortitude and strength.\n\n]" time="0.060"><properties><property name="score" value="0.2005504021114398" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.2005504&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.2005504
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Aulus Atilius Calatinus\n\nAulus Atilius Calatinus was consul of the Roman Republic in 221 BC with Gaius Claudius Nero. The primary action of the consulship was the siege and capture of the Volscian city of Suessa Pometia.\n\nIn the same year he was sent to Tarentum with a fleet of 10,000 infantry and 800 cavalry to aid the Tarentines in the Second War with the Bruttians. Calatinus was captured in a naval action off the coast near Tarentum, and the Bruttian admiral Lucerius was rewarded with his command for this victory. The Roman troops under his command, however, escaped to Thurii.\n\nThe earlier Aulus Atilius Calatinus was elected consul in 222 BC, and killed in a battle with the Samnites.\n\n]" time="0.019"><properties><property name="score" value="0.5281359200828839" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.52813592&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.52813592
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Frederic Myers\n\nFrederic William Henry Myers (15 January 1843 \u2013 18 January 1901) was a classical scholar, poet, playwright, novelist, and the founder of the Society for Psychical Research, which he led until his death.\n\nFrederic Myers was born at Kew, Surrey, the eldest son of Frederic Myers, a surgeon. He was educated at King's College School, and at St Paul's School in London. After a year at Trinity College, Cambridge (1861\u20132), he was a schoolmaster at Rugby from 1862 to 1871. In 1872 he became classical lecturer at University College, London. He became a member of the Council of the College (now part of UCL), and was appointed professor of Latin there in 1880. In 1876 Myers joined the Metaphysical Society, along with Alfred Tennyson, J. M. Barrie and others. His friendship with Barrie lasted for the remainder of his life, and Myers even accompanied Barrie and his wife on their honeymoon. Myers and Tennyson became close friends and met frequently in later years, as seen in their correspondence.\n\nMyers wrote prose and verse for periodicals, including a parody of Gerard Manley Hopkins, and some detective stories, and contributed to books of poetry and belles lettres, such as &quot;Whistler and His Art&quot; and the &quot;Yellow Book&quot;. His poetry was praised by Henry James, who]" time="0.033"><properties><property name="score" value="0.5557702860330693" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.55577029&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.55577029
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[His father was a professor and his mother was an economist. While growing up, he was exposed to the inspiring stories of his parents and their family members in China.\n\n\n\n\n\nKoh attended the University of Michigan where he received a bachelor of arts degree in economics, then went to the University of California, Berkeley where he earned a master of arts and doctor of philosophy degrees in economics.\n\n\n\n\n\nHe returned to Singapore in 1998 and joined the Singapore Civil Service.\n\n\n\n\n\nAs Deputy Secretary of the Ministry of Finance (2001-2007), he was involved in the design and implementation of the Singapore Savings Bond and the Central Provident Fund (CPF) investments.\n\n\n\n\n\nIn 2008, Koh was appointed as the Managing Director of the Monetary Authority of Singapore (MAS), where he oversees the development and implementation of macroeconomic policies for Singapore.\n\n\n\n\n\nHe is also the Chairman of the Government of Singapore Investment Corporation (GIC) which is responsible for the management of Singapore\u2019s foreign reserves and investment portfolio.]" time="0.017"><properties><property name="score" value="0.2400253671426822" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Little Tikes LTX Battery Operated Tricycle\n\nHe\u2019s been dreaming about his first bike for a while, and now he can hop on and ride around with friends and siblings alike on this bright red Little Tikes LTX Battery Operated Tricycle. It features an adjustable seat, a footrest and a chain guard so he\u2019ll stay safe while riding. Battery-operated, it requires 2 AA batteries (not included). This is a toy for children 3 years and older.]" time="0.017"><properties><property name="score" value="0.24318361167285918" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Trixie Textured - My cat licked my thumb.\n\n\n\nJoin Date: Nov 2006 Posts: 4,592\n\nI have not been to China, though I have been to Japan and love it.\n\n\n\nI have friends in China who have also been to Japan, and they both love it, and neither of them think of it as a &quot;poor&quot; country. In fact, a lot of people who have been there say that China is much nicer.\n\n\n\nChina is not a poor country. They have tons of money and resources. Their problem is that they do not spend enough on infrastructure and don't value education as highly as they could. So their people do not make as much money as they could. And they don't have nice things. It is much nicer than India, but not as good as Japan. I would like to visit it, because I do not think I would be disappointed, but I don't expect it to be better than Japan.\n\n\n\nThey have a different mindset than us, and do things differently. Their government is not a democracy. It is a communist government. I think that is not a bad thing, and has it's benefits. They have much fewer guns and gun crime than we do. They are, for the most part, more honest than we are. Their infrastructure, while not as nice as ours, is very good. They spend more on infrastructure than we do. They have much more public transportation, and better public transportation. They are working on their environmental problems, and they have made more progress than we have. They have fewer social problems, and less crime than we do. They have a very good education system, especially compared to India.\n\n\n\nThey spend more money on schools and education, and invest more in their future, than we do.\n\n\n\nChina is a country that is far more well developed than many people in the US would think.\n\n\n\nThe fact that they have fewer guns, and better education and fewer social problems is probably why they have fewer school shootings than we do. I think we have way too many guns in this country, and that is a big part of the problem.\n\n\n\nJust my thoughts.]" time="0.038"><properties><property name="score" value="0.3292728125198763" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[&quot;Say to them: 'As surely as I live, declares the Sovereign Lord, I take no pleasure in the death of the wicked, but rather that they turn from their ways and live. Turn! Turn from your evil ways! Why will you die, people of Israel?' declares the Sovereign Lord.&quot;\n\nEzekiel 33:11-11\n\nGod wants every sinner to be saved. He does not want people to perish but be saved. God hates sin, and sin destroys a person. But He loves people. He created them to have fellowship with Him and one day live in His Kingdom, and He does not want that for one person to be lost.\n\n\n\n\n\nGod's command to Ezekiel was that He was telling him to go and tell the people of Israel that they need to turn to God, because they were going to die if they did not. We know that Ezekiel told them the same thing, because God's word is truth. God's Word does not fail.\n\n\n\n\n\nWhat God is saying is that people are going to die if they do not turn to Him. People who reject God will die. The Bible says in Romans 1:32, &quot;Though they know God's righteous decree that those who do such things deserve death, they not only continue to do these very things but also approve of those who practice them.&quot; God says that those people who do these things deserve death, and yet people continue to do them and approve of them.\n\n\n\n\n\nGod will not tolerate the sinful acts of humanity, but He wants every person to be saved. There is only one way to be saved. The Bible says in John 14:6, &quot;Jesus answered, 'I am the way and the truth and the life. No one comes to the Father except through me.'\u201d\n\n\n\n\n\nNo one can be saved except through Jesus Christ, but when people put Jesus as their savior, they are saved. It's that simple.\n\n\n\n\n\nThere are many people who call themselves Christians who say that God wants everyone to be saved, and they quote Ezekiel 33:11 and John 3:16, but they don't realize that God's word is truth. God says in Revelation 21:8, &quot;But the cowardly, the unbelieving, the vile, the murderers, the sexually immoral, those who practice magic arts, the idolaters and all liars\u2014their place will be in the fiery lake of burning sulfur. This is the second death.&quot;\n\n\n\n\n\nThere is only one way to be saved, and it is not through anyone or anything but Jesus Christ. Jesus is the only way to heaven. God says that all the people who are going to perish are going to go to a place called the second death. That's hell.\n\n\n\n\n\nWhen people quote Ezekiel 33:11 and John 3:16, they don't realize that God's word is truth. They don't realize that when they say that God wants everyone to be saved, they are not telling the truth. When they say that God wants everyone to be saved, they are putting God in a box, because God's word is truth. God says that those who do not accept Jesus as their Savior are going to perish. God's word is truth, and people who do not put their faith in Jesus Christ are going to die.\n\n\n\n\n\nThe Bible says in John 14:6, &quot;Jesus answered, 'I am the way and the truth and the life. No one comes to the Father except through me.'\u201d\n\n\n\n\n\nJesus is the only way to the Father, and no one can be saved except through Him. He died on the cross for our sins, and He rose from the dead on the third day, and He sits at the right hand of God, interceding for us. It's that simple.\n\n\n\n\n\nLet's take a look at Ezekiel 33:11-11, and read it carefully.\n\n\n\n\n\n&quot;Say to them: 'As surely as I live, declares the Sovereign Lord, I take no pleasure in the death of the wicked, but rather that they turn from their ways and live. Turn! Turn from your evil ways! Why will you die, people of Israel?' declares the Sovereign Lord.&quot;\n\n\n\n\n\nGod says that He takes no pleasure in the death of the wicked. God does not like it when people die. But God hates sin, and sin leads to death. If people want to live, they must turn from their sins and put their faith in Jesus Christ.\n\n\n\n\n\nIf people don't put their faith in Jesus Christ, they are going to perish. God doesn't want that. He doesn't want people to perish, but to live.\n\n\n\n\n\nGod wants people to put their faith in Jesus Christ. If people put their faith in Jesus Christ, they are saved. If they don't put their faith in Jesus Christ, they are going to perish. That's the truth. That's the message of Ezekiel 33:11-11.\n\n\n\n\n\nPeople are going to die if they don't put their faith in Jesus Christ. People will perish. People will be lost. People will be doomed. People will go to hell.\n\n\n\n\n\nLet's read Ezekiel 33:11-12.\n\n\n\n\n\n&quot;For I take no pleasure in the death of anyone,&quot; declares the Sovereign Lord. &quot;Repent and live!&quot;\n\n\n\n\n\nGod takes no pleasure in the death of anyone. God does not want anyone to perish. He doesn't want anyone to die. He wants people to repent and live. God is saying that He is not going to stand back and watch as people perish. He's going to do something about it.\n\n\n\n\n\nGod says in Ezekiel 33:13-16, &quot;Turn! Turn from your evil ways! Why will you die, people of Israel? I have no pleasure in the death of the wicked, but rather that they turn from their ways and live. Do you think I take pleasure in the death of the wicked? declares the Sovereign Lord. Rather, am I not pleased when they turn from their ways and live? Don't presume that you can say to yourselves]" time="0.079"><properties><property name="score" value="0.08402352517560274" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Image caption Bailiffs removed protesters from Grangemouth plant in October 2013\n\nFive people have been charged in connection with an anti-fracking protest at an oil refinery in Grangemouth.\n\nThe protesters climbed on to the chimney at Ineos petrochemicals plant on 9 October.\n\nPolice Scotland said two men, aged 21 and 24, had been charged in connection with the protest and would appear at Falkirk Sheriff Court on Monday.\n\nThree women, aged 40, 44 and 48, were charged with breach of the peace and will also appear at court.\n\nThe 40-year-old was also charged with breach of the peace.\n\nThe other three were arrested in connection with a protest outside a hotel in Glasgow on 10 October.]" time="0.028"><properties><property name="score" value="0.2945623072513892" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Working with an example - rotating through an array\n\nFor a number of years, I've been fascinated by the idea of writing code that can rotate through an array. Let's say I have the following array of numbers:\n\nvar myNumbers = [ 1, 2, 3, 4, 5 ];\n\nThe following snippet of code would display the same numbers in reverse order:\n\nfor ( var i = myNumbers.length - 1; i &gt;= 0; i-- ) { console.log( myNumbers[i] ); }\n\nWhat's going on here?\n\nWe know that the index of an array can be used as a &quot;pointer&quot; to the location in memory that contains a particular item in an array. If the array myNumbers has five items, then myNumbers[0] , myNumbers[1] , myNumbers[2] , myNumbers[3] and myNumbers[4] will point to the values 1 , 2 , 3 , 4 and 5 respectively.\n\nIn the previous example, the code starts with the length of the array (5) and subtracts one from it. As we know that index 0 in an array contains the first item in the array, the previous code will execute the following steps:\n\nTake the length of the array (5) Subtract 1 (so we now have 4) Start at the index that contains the first item in the array (myNumbers[4] = 3) and run down the array until we reach the last item in the array (myNumbers[3] = 5)\n\nSo the above code will end up displaying the items in the array in reverse order, starting at index 4 and working backwards to index 0.\n\nAt this point, we have a good idea of how we can access the values in an array in a specific order, but how do we go the other way?\n\nIt would be great if we could simply use the same code as above, but add a &quot;--&quot; operator before the last number. That way, we could add a simple parameter to the code to control the range of numbers it will use.\n\nUnfortunately, we can't do that. While JavaScript will happily work with -- , it will also happily take -- in code and convert it to -1 , or the last number in an array. That's not going to do us any good, as we want the code to be flexible enough that we can work with a number that we specify as the range of numbers.\n\nSo let's look at a simple function that can give us the results we need.\n\nfunction rotate(numbers, direction, number) { numbers.reverse(); numbers.push(number); numbers.unshift(number); };\n\nThe rotate function above has three parameters, and will rotate the values in the numbers array.\n\nnumbers is the array we want to use.\n\ndirection is the parameter that allows us to choose whether we want the values to be rotated in the direction we specify. In this case, we've chosen direction to be either &quot;right&quot; or &quot;left&quot;\n\nnumber is the number we're using as the upper bound of the range that we want to use.\n\nLet's take a look at a simple example:\n\nvar numbers = [ 1, 2, 3, 4, 5 ]; rotate( numbers, &quot;right&quot; ); console.log( numbers ); //output: [ 1, 2, 3, 4, 5 ]\n\nAs we can see, the numbers array is simply rotated right by the number we specified. That's fine, but let's add in some code that we can use to control which numbers are displayed:\n\nvar numbers = [ 1, 2, 3, 4, 5 ]; rotate( numbers, &quot;right&quot;, 1 ); console.log( numbers ); //output: [ 1, 4, 3, 5, 2 ]\n\nWe can see that we've rotated the numbers so that the first number in the array is now number 1 , and the next number in the array is 4 .\n\nWe've also specified that we want to rotate in a right-handed manner, by passing &quot;right&quot; as the second parameter. This means that we have access to the following rotation values:\n\nrotate( numbers, &quot;left&quot; ); rotate( numbers, &quot;right&quot; ); rotate( numbers, &quot;right&quot; ); rotate( numbers, &quot;right&quot; ); rotate( numbers, &quot;left&quot; );\n\nIf we try to access the rotate function with &quot;left&quot; as the parameter, we'll see the following output:\n\nvar numbers = [ 1, 2, 3, 4, 5 ]; rotate( numbers, &quot;left&quot; ); console.log( numbers ); //output: [ 4, 3, 5, 2, 1 ]\n\nNot quite what we wanted, right?\n\nWhat we want is to be able to take an array, specify the number of numbers we want to display, and be able to choose the direction of rotation.\n\nWe also want to ensure that the values we are displaying are always within the array's bounds, and we're using them correctly.\n\nThe good news is that it's not too difficult to do this. We simply need to know what we want the code to look like, and write the code to do what we want it to do.\n\nLet's take a look at some simple code to display the numbers in an array in a given direction:\n\nfunction rotateArray( numbers, direction, startAt ) { numbers.reverse(); numbers.push( startAt ); numbers.unshift( startAt ); };\n\nrotateArray function\n\nThis is the code we'll use to display the numbers in the numbers array in the direction we specify. We'll start with a simple parameter:\n\nstartAt is the number we're displaying. It can either be the first number in the array, or any number within the array.\n\nLet's take a look at the code again:\n\nfunction rotateArray( numbers, direction, startAt ) { numbers.reverse(); numbers.push( startAt ); numbers.unshift( startAt ); };\n\nUsing this simple function, we can take any number of numbers, and display them in the direction we want. Let's look at a quick example:\n\nvar numbers = [ 1, 2, 3, 4, 5 ]; rotateArray( numbers, &quot;left&quot;, 3 ); console.log( numbers ); //output: [ 4, 3, 5, 2, 1 ]\n\nThis time, we've used &quot;left&quot; as the direction, and we've also used the parameter &quot;3&quot; as the startAt value. The result is that we now have the numbers displayed in the array starting at the number 3 .\n\nIf we use a number that isn't within the bounds of the array, we'll get the following result:\n\nvar numbers = [ 1, 2, 3, 4, 5 ]; rotateArray( numbers, &quot;right&quot;, 3 ); console.log( numbers ); //output: [ 5, 2, 4, 1, 3 ]\n\nAs you can see, the direction we've specified is &quot;right&quot;, and the value we've specified for startAt is 3 . Unfortunately, this value isn't within the bounds of the array, so the values in the array have been displayed starting from the last number in the array.\n\nLet's take a look at how we can use this code to work with a number that's outside the bounds of the array:\n\nvar numbers = [ 1, 2, 3, 4, 5 ]; rotateArray( numbers, &quot;right&quot;, 6 ); console.log( numbers ); //output: [ 5, 2, 4, 1, 3 ]\n\nThis time, we've specified &quot;right&quot; as the direction, and a value of 6 as the startAt parameter. The result is that we now have the values displayed in the array starting at the number 6 .\n\nThis time, the result we want is displayed in the array, but there's a problem. We don't want the number 6 to be the start of the array. Instead, we want it to be the last number in the array.\n\nThe good news is that we can easily handle this by modifying the code slightly:\n\nfunction rotateArray( numbers, direction, startAt ) { numbers.reverse(); numbers.push( startAt ); numbers.unshift( startAt ); };\n\nUsing the simple rotation code above, we can now take any array, specify the number we want to display, and choose the direction we want to use. This means we can write some simple code that can work with any number, regardless of its position in the array.\n\nThe next step is to write some code that can ensure the number we specify as the startAt is in the range that we expect.\n\nLet's take a look at how we can do that:\n\nfunction rotateArray( numbers, direction, startAt ) { if( startAt &gt;= numbers.length ) { throw new RangeError(&quot;Start at is outside of the]" time="0.076"><properties><property name="score" value="0.18576626194144463" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Contents\n\n1. The Meaning of Love\n\n2. Our Attitude Toward Others\n\n3. The Two Great Commandments\n\n4. Developing Love\n\n5. How to Help Others\n\n6. The Golden Rule\n\n7. The Commandments in the New Testament\n\n8. Tolerance\n\n\nCHAPTER 1\n\nThe Meaning of Love\n\nBefore we can love others, we must know the meaning of love. It is important to love God and obey Him and to have a right relationship with Him. But to have a right relationship with Him we must love Him. That means that we must have a good relationship with all His laws.\n\nOf course we can't keep all the commandments all the time. Yet we must have a good relationship with them. Otherwise we will not know what God is like.\n\nLove is based on God's law.\n\nRomans 13:10. The commandments, You shall not commit adultery, You shall not steal, You shall not bear false witness, You shall not covet; and any other commandment, are summed up in this word, You shall love your neighbor as yourself.\n\nGalatians 5:14. For all the law is fulfilled in one word, even in this: You shall love your neighbor as yourself.\n\nLove is based on God's law. We should study God's commandments and know the laws He has given us. Then we will know how to have a good relationship with Him.\n\nOne reason we are to love God and obey Him is that it is important to have a good relationship with God. We need to know how much God loves us. Without knowing this we cannot have a right relationship with God. We need to have a good relationship with Him so we can be happy and serve Him better.\n\nLet us understand what love is.\n\n1 John 4:8,9. He who does not love does not know God, for God is love. In this the love of God was manifested toward us, that God has sent His only begotten Son into the world, that we might live through Him.\n\nIn these verses John explained that love is the opposite of lawlessness. Love is the opposite of murder, hatred, malice, evil, adultery, stealing, lying, and a lot of other bad things.\n\nRomans 13:8-10. Owe no man anything, but to love one another: for he that loves another has fulfilled the law. For this, You shall not commit adultery, You shall not kill, You shall not steal, You shall not covet, and if there be any other commandment, it is briefly comprehended in this saying, namely, You shall love your neighbor as yourself. Love does no harm to a neighbor; therefore love is the fulfillment of the law.\n\nWe learn from this that the opposite of love is lawlessness. The opposite of love is murder, hatred, envy, malice, evil, and all the other things mentioned in these verses.\n\nVerse 10 says, &quot;Love does no harm to a neighbor.&quot; In other words, love is a good thing. It is a good thing to have a good relationship with God.\n\nRomans 13:8 says, &quot;Love does no harm to a neighbor.&quot; Love helps and blesses others.\n\nLove is a good thing. The reason love is a good thing is because God is love.\n\n1 John 4:16. And we have known and believed the love that God has for us. God is love; and he who abides in love abides in God, and God in him.\n\nGod is love.\n\n1 John 4:7-11. Beloved, let us love one another: for love is of God; and every one that loves is born of God, and knows God. He that loves not knows not God; for God is love. In this was manifested the love of God toward us, because that God sent his only begotten Son into the world, that we might live through him. Herein is love, not that we loved God, but that he loved us, and sent his Son to be the propitiation for our sins. Beloved, if God so loved us, we ought also to love one another. No man has seen God at any time. The only begotten Son, which is in the bosom of the Father, he has declared him.\n\nVerse 8 says, &quot;God is love.&quot; God's name is Love.\n\nGod's name is Love.\n\nJohn 4:8. \u2026 the well was deep; and I said, &quot;How shall I get down?&quot;\n\nAnd he said, &quot;Come, and I will show you.&quot;\n\n9. And he said, &quot;Draw you [a] little water out of the well, and \u2026 drink.&quot;\n\n10. And I said, &quot;Sir, give me [a] little to drink.&quot;\n\n11. And he said, &quot;Drink, my [b] son, and \u2026 your eyes shall be opened.&quot;\n\n12. And I drank, and it was as if I had had new vision given to me; for I could see [c] through those walls of water.\n\n13. And he said to me, &quot;Go [a] down, for the water is [b] yet a little way below the earth. \u2026\n\n15. \u2026 This water is [c] that which I said \u2026 will make you come up hither.&quot;\n\n16. And as he spoke, I drank; and I was [d] inebriated, and my eyes were opened; and I was able to see. \u2026\n\n19. \u2026 &quot;Come,&quot; he said, &quot;follow me \u2026 and I will give you \u2026 rest.&quot;\n\nGod is the God of rest.\n\n1 Thessalonians 4:9. \u2026 God has not appointed us to wrath, but to obtain salvation by our Lord Jesus Christ.\n\nGod is our salvation. God is our rest. God is our peace.\n\nGod is our rest.\n\n1 Peter 2:6. \u2026 you may be \u2026 a kind of first-fruits of his creatures.\n\nGod is our first fruits.\n\n2 Timothy 2:13. If we are faithless, he remains faithful; for he cannot deny himself.\n\nGod is faithful.\n\nGod is the God of rest.\n\nGod is our salvation.\n\nGod is our peace.\n\nGod is our rest.\n\nHebrews 4:10-13. For he who has entered His rest has himself also ceased from his own works, as God did from His. Let us therefore be diligent to enter that rest \u2026\n\n12. \u2026 For the word of God is living and active, and sharper than any two-edged sword, and piercing even to the division of soul and spirit, and of joints and marrow, and is a discerner of the thoughts and intents of the heart.\n\n13. \u2026 neither is there any creature that is not manifest in his sight; but all things are naked and opened to the eyes of him with whom we have to do.\n\nGod is a discerner of the thoughts and intents of the heart.\n\nGod sees everything.\n\nGod sees everything.\n\n1 John 4:8,9. He who does not love does not know God, for God is love. In this the love of God was manifested toward us, that God has sent His only begotten Son into the world, that we might live through Him.\n\nGod is love.\n\n1 John 3:16,18. Hereby perceive we the love of God, because he laid down his life for us: and we ought to lay down our lives for the brethren.\n\n19. But whoso has this world's goods, and sees his brother in need, and shuts up his heart from him, how does the love of God abide in him?\n\nGod is love.\n\nGod is our rest.\n\nGod is our salvation.\n\nGod is our peace.\n\nGod is a discerner of the thoughts and intents of the heart.\n\nGod is love.\n\nGod is our rest.\n\nGod is our salvation.\n\nGod is our peace.\n\nGod is a discerner of the thoughts and intents of the heart.\n\nThe Bible says that God is love.\n\nWe can study love.\n\n1 John 4:7-11. Beloved, let us love one another: for love is of God; and every one that loves is born of God, and knows God. He that loves not knows not God; for God is love. In this was manifested the love of God toward us, because that God sent his only begotten Son into the world, that we might live through him. Herein is love, not that we loved God, but that he loved us, and sent his Son to be the propitiation for our sins. Beloved, if God so loved us, we ought also to love one another. No man has seen God at any time. The only begotten Son, which is in the bosom of the Father, he has declared him.\n\nVerse 11 says, &quot;Herein is love, not that we loved God, but that he loved us.&quot; The opposite of love is lawlessness. But the opposite of love is not lawlessness. The opposite of love is law.\n\nLoving God and obeying Him is loving Him.\n\nThe Bible says that God is love.\n\nLoving God and obeying Him is loving Him.\n\nLoving God and obeying Him is loving God.\n\nJohn 14:21. He that has my commandments,]" time="0.078"><properties><property name="score" value="0.06954537065634314" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06954537&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06954537
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Interviews\n\nInterview with Ariel Efraim (Soilwork)\n\nInterview with Steven Rosen (August Burns Red)\n\nInterview with Brent Hinds (Mastodon)\n\nInterview with Paul Kuhr (I See Stars)\n\nInterview with Christian Antidormi (Hatebreed)\n\nInterview with Jack Syperek (Structures)\n\nInterview with Ben Gordon (The Treatment)\n\nInterview with Doc Coyle (God Forbid)\n\nInterview with Chris Barnes (Six Feet Under)\n\nInterview with Mike D (The Sword)\n\nInterview with Russ Rankin (The Slackers)\n\nInterview with Claudio Sanchez (Coheed and Cambria)\n\nInterview with Udo Dirkschneider (Accept)\n\nInterview with Ben Weinman (Dillinger Escape Plan)\n\nInterview with Matt Pike (High On Fire)\n\nInterview with Joe Duplantier (Gojira)\n\nInterview with Johannes Eckerstr\xf6m (Evergrey)\n\nInterview with Pete Loose (Reel Big Fish)\n\nInterview with Pete Wentz (Fall Out Boy)\n\nInterview with Daniel Svensson (Raubtier)\n\nInterview with Chino Moreno (Deftones)\n\nInterview with Pierre Bouvier (Simple Plan)\n\nInterview with Jon Egan (The Ghost Of A Thousand)\n\nInterview with Mike D (Lack Of Afro)\n\nInterview with Michael Crafter (Sevendust)\n\nInterview with Charlie Benante (Anthrax)\n\nInterview with Rob Caggiano (Volbeat)\n\nInterview with Ronnie Atkins (Pretty Maids)\n\nInterview with Gee Anzalone (Exhorder)\n\nInterview with Micky Shirley (Five Finger Death Punch)\n\nInterview with Oderus Urungus (GWAR)\n\nInterview with Ryan Downey (Texas Hippie Coalition)\n\nInterview with Chris Chasse (Unearth)\n\nInterview with Lee Altus (Suffocation)\n\nInterview with Roy Mayorga (Soulfly)\n\nInterview with Gus G (Firewind)\n\nInterview with Paul Allender (In Flames)\n\nInterview with Francesco Artusato (All Shall Perish)\n\nInterview with Rody Walker (Pyogenesis)\n\nInterview with Mike Dean (Slipknot)\n\nInterview with Angela Gossow (Arch Enemy)\n\nInterview with Trevor Strnad (The Black Dahlia Murder)\n\nInterview with Mike Peters (The Alarm)\n\nInterview with Jonas Ekdahl (Freak Kitchen)\n\nInterview with Dave Ellefson (Megadeth)\n\nInterview with Corey Taylor (Slipknot)\n\nInterview with Keith Buckley (Every Time I Die)\n\nInterview with Jack Frost (Kill Devil Hill)\n\nInterview with Josh Rand (Temple of Brutality)\n\nInterview with Michael Amott (Arch Enemy)\n\nInterview with Scott Gorham (Thin Lizzy)\n\nInterview with Matt Heafy (Trivium)\n\nInterview with Joe Kresge (Unearth)\n\nInterview with Randy Blythe (Lamb of God)\n\nInterview with Benji Webbe (Dimmu Borgir)\n\nInterview with Max Cavalera (Soulfly)\n\nInterview with John Bush (Armored Saint)\n\nInterview with Davey Muise (Exhorder)\n\nInterview with Stephen Carpenter (Deftones)\n\nInterview with Dani Filth (Cradle Of Filth)\n\nInterview with Leif Cuzner (Darkane)\n\nInterview with George \u201cCorspegrinder\u201d Fisher (Cannibal Corpse)\n\nInterview with Steve Asheim (Hatebreed)\n\nInterview with Bobby Blitz Ellsworth (Overkill)\n\nInterview with Adam Dutkiewicz (Killswitch Engage)\n\nInterview with Mick Murphy (Exhumed)\n\nInterview with Doug Pinnick (King\u2019s X)\n\nInterview with Alexi Laiho (Children Of Bodom)\n\nInterview with M. Shadows (Avenged Sevenfold)\n\nInterview with Alex Holzwarth (Rhapsody Of Fire)\n\nInterview with Carlos Cruz (Soulfly)\n\nInterview with Glenn Hughes (Deep Purple/Black Country Communion)\n\nInterview with Oliver Palotai (Kamelot)\n\nInterview with Meegs Rascon (The Mars Volta)\n\nInterview with Joakim Brod\xe9n (Sabaton)\n\nInterview with Tim \u201cRipper\u201d Owens (Iced Earth/Yngwie Malmsteen)\n\nInterview with Matt Heafy (Trivium)\n\nInterview with Warrel Dane (Nevermore)\n\nInterview with Zach Myers (Shinedown)\n\nInterview with Michael Poulsen (Volbeat)\n\nInterview with Michael Amott (Arch Enemy)\n\nInterview with Mat Sinner (Primal Fear)\n\nInterview with Bj\xf6rn \u201cSpeed\u201d Strid (Soilwork)\n\nInterview with Matt Heafy (Trivium)\n\nInterview with Rob Flynn (Machine Head)\n\nInterview with Serj Tankian (System Of A Down)\n\nInterview with Ville Valo (HIM)\n\nInterview with Jimmy DeGrasso (Megadeth)\n\nInterview with Joey Jordison (Slipknot)\n\nInterview with Chris Jericho (Fozzy/Fozzy Osbourne)\n\nInterview with Jon Donais (Shadows Fall)\n\nInterview with Jason Hook (Five Finger Death Punch)\n\nInterview with Steve Mazur (Grave Digger)\n\nInterview with Bj\xf6rn \u201cSpeed\u201d Strid (Soilwork)\n\nInterview with Sonny Sandoval (P.O.D.)\n\nInterview with Serj Tankian (System Of A Down)\n\nInterview with Fredrik Thordendal (Meshuggah)\n\nInterview with Wes Borland (Limp Bizkit)\n\nInterview with Alexi Laiho (Children Of Bodom)\n\nInterview with Blaze Bayley (Iron Maiden)\n\nInterview with Mark Hunter (Chimaira)\n\nInterview with Max Cavalera (Soulfly/Sepultura)\n\nInterview with Dave Lombardo (Slayer)\n\nInterview with Bruce Dickinson (Iron Maiden)\n\nInterview with Morgan Lander (Fear Factory)\n\nInterview with Oliver Palotai (Kamelot)\n\nInterview with James Hetfield (Metallica)\n\nInterview with Al Jourgensen (Ministry)\n\nInterview with Corey Taylor (Stone Sour/Slipknot)\n\nInterview with Alex Skolnick (Testament)\n\nInterview with Charlie Benante (Anthrax)\n\nInterview with Chris Reifert (Autopsy)\n\nInterview with Lajon Witherspoon (Sevendust)\n\nInterview with Jason McMaster (Grave Robber)\n\nInterview with Jon Howard (Light The Torch)\n\nInterview with Neil Fallon (Clutch)\n\nInterview with Dino Cazares (Fear Factory)\n\nInterview with Oderus Urungus (GWAR)\n\nInterview with Rex Brown (Pantera)\n\nInterview with Dave Lombardo (Slayer)\n\nInterview with Serj Tankian (System Of A Down)\n\nInterview with John Moyer (Disturbed)\n\nInterview with Mike Portnoy (Adrenaline Mob)\n\nInterview with King Diamond\n\nInterview with Trevor Strnad (The Black Dahlia Murder)\n\nInterview with Mikael \xc5kerfeldt (Opeth)\n\nInterview with Scott Ian (Anthrax)\n\nInterview with Paul Gray (Slipknot)\n\nInterview with Casey Chaos (Amen)\n\nInterview with Opeth\u2019s Mikael \xc5kerfeldt\n\nInterview with Mark Morton (Lamb Of God)\n\nInterview with Shawn Drover (Megadeth)\n\nInterview with Chino Moreno (Deftones)\n\nInterview with Serj Tankian (System Of A Down)\n\nInterview with Bruce Corbitt (Corrosion Of Conformity)\n\nInterview with Ray Alder (Fates Warning)\n\nInterview with Rob \u201cBlasko\u201d Nicholson (Ozzy Osbourne)\n\nInterview with Tom Araya (Slayer)\n\nInterview with Max Cavalera (Soulfly)\n\nInterview with Ryan Martinie (Cynic)\n\nInterview with Max Cavalera (Soulfly)\n\nInterview with Jason Newsted (Voivod)\n\nInterview with Mick Murphy (Exhumed)\n\nInterview with Lars Frederiksen (Rancid)\n\nInterview with Mike Muir (Suicidal Tendencies)\n\nInterview with Dave Mustaine (Megadeth)\n\nInterview with Neil Fallon (Clutch)\n\nInterview with Bruce Dickinson (Iron Maiden)\n\nInterview with Maynard James Keenan (Tool)\n\nInterview with Mike \u201cRat\u201d Ratner (S.O.D.)\n\nInterview with Max Cavalera (Soulfly)\n\nInterview with Tom Araya (Slayer)\n\nInterview with Chuck Billy (Testament)\n\nInterview with Tom Araya (Slayer)\n\nInterview with Phil Anselmo (Down)\n\nInterview with King Diamond\n\nInterview with Corey Taylor (Stone Sour/Slipknot)\n\nInterview with Dave Mustaine (Megadeth)\n\nInterview with Lars Ulrich (Metallica)\n\nInterview with Jerry Cantrell]" time="0.081"><properties><property name="score" value="0.6886236319633857" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.68862363&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.68862363
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Apple\u2019s ResearchKit is off to a good start, but there\u2019s plenty of work to be done.\n\n\n\nResearchKit was launched in March as a way for Apple\u2019s developers to easily tap into medical data from the iPhone to help researchers track and collect information about a variety of health conditions. Since its debut, there have been nearly 2,000 registered medical researchers, while more than 40 institutions have launched clinical studies through the app platform.\n\n\n\nBut there are several hurdles for Apple and its partners. On Wednesday, Dr. Harvey Fineberg, president of the Institute of Medicine, will join at a panel discussion at Mount Sinai Hospital in New York to examine what works and what\u2019s missing in the fast-evolving field of mobile health.\n\n\n\nThe session, titled \u201cWhat Are the Future Directions for ResearchKit?\u201d comes after a first wave of successful projects on ResearchKit, including apps that help people manage diabetes, heart disease and asthma.\n\n\n\nThere are also some examples of the problems Apple is facing as it tries to get more partners to adopt the tool. Earlier this month, a major player in the diabetes community said it would not be taking part in the app\u2019s data collection. The Juvenile Diabetes Research Foundation said in a statement that its members were not interested in the iPhone platform and \u201cwould prefer to conduct research studies using our own patient data that is gathered on an Android platform.\u201d\n\n\n\nThat\u2019s likely an expensive prospect for the organization, given the popularity of iPhones in the United States. As of last year, Android devices accounted for just over half of all U.S. smartphone subscribers, while iOS devices made up around 41 percent. That\u2019s according to the Pew Research Center, which estimated last year that there were 121 million iPhone users in the United States.\n\n\n\nThe success of Apple\u2019s platform will hinge in part on how it handles data privacy and security concerns. Apple had to clear some hurdles in the early days of ResearchKit because many people were concerned about privacy when they heard it would rely on data from Apple\u2019s servers. Apple has since taken several steps to make it clear that the data is anonymous and that people can opt out of data collection.\n\n\n\nThe app program could also be a major part of the iPhone\u2019s future if it takes off. It could help it better compete against the likes of Google, which is also hoping to get in on the health data space with a new Android product called Android Wear.\n\n\n\nThe first public version of ResearchKit will be unveiled on Wednesday, the panel will also include representatives from Mount Sinai and Pfizer.\n\n\n\nThat event will be at 11:30 a.m. EDT in New York City, and it will be livestreamed on Fortune\u2019s website.]" time="0.071"><properties><property name="score" value="0.08212340796613186" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[\n\nSend this thread via email by filling out the fields below\n\n\n\nThread: Lost 10lbs in the last 3 weeks Send To:\n\nUse semi-colon to separate multiple addresses Your Name: Your Email: Add a personal message: Send this thread via email by filling out the fields below\n\n\n\n]" time="0.026"><properties><property name="score" value="0.08522751945544416" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08522752&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08522752
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[I don't think it is your attitude. I think it is just the way you are.\n\nIt isn't a criticism, just a statement of fact. It's like saying I don't think you are stupid.\n\nJust that you are naive, sweet, guileless, open. You have so much going for you, you don't have to be cynical.\n\nBut it isn't fair to the rest of us, not fair to expect us to feel your innocence and not to respond to it.\n\nYou need to toughen up, otherwise you will just be hurt. But don't ever be cynical.]" time="0.023"><properties><property name="score" value="0.42586882921175073" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Get the biggest daily news stories by email Subscribe Thank you for subscribing We have more newsletters Show me See our privacy notice Could not subscribe, try again later Invalid Email\n\nAs the Queen's cousin, Lady Amelia Windsor is more used to socialising with the royal family than you or I.\n\nBut even for her, it seems the sight of seeing Kate Middleton and Prince William in attendance at Pippa Middleton's wedding - as well as the newlyweds' own big day - was enough to make her feel &quot;uncomfortable&quot;.\n\nSpeaking at the wedding of one of her friend's brothers in Kent, the 28-year-old spoke about how it made her feel to have them there, reports the Daily Mail .\n\nShe said: &quot;Well I think it is a beautiful wedding, obviously, because it's Pippa and James.\n\n&quot;I've never been to a wedding before, so I don't know really what it should be like. But I do find it quite uncomfortable seeing them.&quot;\n\nThe Royal family came together to celebrate the wedding of Kate's sister to hedge fund manager James Matthews in the Berkshire village of Englefield.\n\n(Image: Getty Images Europe)\n\n(Image: PA)\n\n(Image: PA)\n\nAnd while Pippa and James' wedding was a private affair with just 200 guests, they were still surrounded by some royals as well as other notable names.\n\nLady Amelia's father is the Queen's first cousin, so it was no surprise to see the Queen herself attend the wedding, alongside Prince Harry and Duchess Meghan.\n\nThe Princesses Beatrice and Eugenie were also there, as well as Kate's sister, Pippa's brother James Middleton, David and Victoria Beckham and Sir Elton John.\n\nLady Amelia revealed that she even had a few words with Harry and Meghan, with the former asking her how she felt about her cousin's big day.\n\nShe said: &quot;I had a little conversation with Harry and Meghan, and they said how lovely the wedding was. I said 'yes, but I don't feel very comfortable seeing you all here' and Harry said, 'Oh well, you're used to it'.\n\n(Image: PA)\n\n(Image: PA)\n\n&quot;I said, 'Yes, but it is still very strange to see you here'. I suppose you can't blame me, I am a bit naive. But the whole thing was really lovely.&quot;\n\nShe added: &quot;I mean, James is my cousin, I grew up with him and I love him dearly, so it was really lovely to see him. It was a lovely wedding, they are such a lovely couple. I am so pleased they are so happy.&quot;\n\n(Image: Getty Images Europe)\n\n(Image: Getty Images Europe)\n\nAnd as for what she thought of her father's outfit - which caused much comment among fans - she admitted she wasn't sure.\n\nShe said: &quot;Well, he always wears an odd choice of clothes, but that is what he likes. It wasn't bad, I mean he didn't look like a big prawn.&quot;\n\nShe added: &quot;It was lovely to see him happy.&quot;]" time="0.084"><properties><property name="score" value="0.4539904578798974" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.45399046&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.45399046
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Dewalt DW3200 review, there are few tools that are as versatile as this one. You may have a thousand bucks of cordless tools, but there\u2019s nothing that will do as much as this one will do. As you might have noticed, the construction of this unit is incredibly sturdy, so that you don\u2019t have to worry about it breaking when you use it. Also, it is a bit heavy, and it will take a while to get used to using it, especially if you have not used any other heavy drills.\n\nPower\n\nWe also tested the power of this Dewalt DW3200 and we must say that it is very impressive. It delivers 3,300 RPM, which is more than enough to bore holes and drill in all kinds of materials. It can easily penetrate into wood, brick, or concrete and it can bore holes up to two inches. The trigger lock mechanism also works well, which makes this drill safer to use.\n\nCordless or corded?\n\nIt is important to remember that this drill can work either as a corded or a cordless unit, and this is very important for you to know. This Dewalt DW3200 uses the 20 volt lithium-ion batteries and they can last longer than you might expect. They are also very light and durable, so you will be able to use them for a long time. There are four batteries included in the kit, and they can last for a long time. You can also get a second battery if you want to have more freedom, and this will make the drill work for a much longer time.\n\nExtra Features\n\nYou may also like the extra features included in this Dewalt DW3200. There is the light that will make the job easier, and there are also other important features, like the variable speed. This is also a very durable drill, so it will last for a very long time. Also, the build quality is fantastic, so you don\u2019t have to worry about it breaking easily.\n\nOverall\n\nIf you have a lot of different drilling needs, this is the drill for you. This drill is perfect for the casual user, but it can also be used by professionals. It is also very durable, so you can use it for a very long time, and you will get the most for your money. You should also remember that this drill is very powerful, and it can work for a long time, without breaking down.\n\nIf you want to see what more we have, you can visit Amazon.com and check all of the available drills. You can also check our article on the best reciprocating saw, to see what the best product in this category is. Also, don\u2019t forget to like and share this article, if you liked it.]" time="0.083"><properties><property name="score" value="0.1659203428388175" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.16592034&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.16592034
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Source: Food\n\nAs I\u2019m sure you\u2019re well aware by now, it\u2019s officially National Grilled Cheese Sandwich Month. With that in mind, this week\u2019s Food Week: Hors D\u2019Oeuvre features a variety of hors d\u2019oeuvres you can make in your own home. After all, a good grilled cheese is just like a fine wine: it gets better with age.\n\nContinue reading]" time="0.041"><properties><property name="score" value="0.43056044707984187" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.43056045&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.43056045
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Jack and Justin are two of the hottest and most popular guys on the block. They make the perfect couple and know exactly what each other likes. But when they meet Tyler, things get complicated. This sizzling trio enjoys their time together and both Tyler and Jack crave more. But Justin is not so easily convinced. How will this all work out?\n\nWarning: This story contains three scorching hot guys and lots of hot and hard action.\n\n\n\n\xa92015 Scarlet Rose (P)2017 Scarlet Rose]" time="0.032"><properties><property name="score" value="0.4643192365798021" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[LAS VEGAS, NV - JANUARY 07: Recording artist Demi Lovato performs onstage during the 2017 DIRECTV NOW Super Saturday Night Concert at Park Theater at Park MGM on January 7, 2017 in Las Vegas, Nevada. (Photo by Kevin Winter/Getty Images for DIRECTV) (Photo: Kevin Winter, 2017 Getty Images)\n\nLAS VEGAS - Former Disney star Demi Lovato will release a new single from her sixth studio album on June 15th.\n\nThat's according to a video of Lovato, 26, posted on the YouTube channel of Las Vegas radio station KVEG.\n\nThe new album is the follow-up to 2015's &quot;Confident,&quot; which features the top 10 hits &quot;Cool for the Summer&quot; and &quot;Confident.&quot;\n\nMore: Everything you need to know about this year's 2017 Route 91 Harvest festival\n\nMore: Demi Lovato set for national tour with DJ Khaled\n\nMore: Demi Lovato re-engages with ex-boyfriend Wilmer Valderrama, so could there be more?\n\nThe pop singer also revealed she is launching a partnership with American Airlines to celebrate Hispanic Heritage Month, and she plans to go on a national tour with DJ Khaled.\n\nDemi Lovato to Release New Single, Announce Album Title on June 15th https://t.co/OBkKbXWySZ \u2014 Pop Crave (@PopCrave) June 1, 2017\n\nRead or Share this story: https://usat.ly/2rVY4mA]" time="0.054"><properties><property name="score" value="0.8216711598806603" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.82167116&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.82167116
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[RACISM IN BRITAIN HAS BECOME SO entrenched that it is becoming a bigger problem than it was at the time of the Birmingham pub bombings, when the threat of violent nationalism in the late 1970s prompted the first official inquiry into racial attacks, according to a report published today by the Equality and Human Rights Commission (EHRC).\n\nThe report, We\u2019re still struggling: the legacy of the Stephen Lawrence inquiry, says that despite years of measures to increase the representation of ethnic minorities in public life and protect them from discrimination, they are still \u201con the receiving end of racism\u201d, with black people twice as likely to suffer police stop and search tactics as whites.\n\nFigures from the Crown Prosecution Service (CPS) show that more than 40 per cent of cases involving charges of racially aggravated harassment are dropped, with charges reduced to more minor ones, or not pursued. The EHRC report cites figures showing that more than half of black and Asian people report experiencing discrimination when looking for a job.\n\nIn October, the Telegraph reported how]" time="0.058"><properties><property name="score" value="0.35982707010781545" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.35982707&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.35982707
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[# This Source Code Form is subject to the terms of the Mozilla Public\n\n# License, v. 2.0. If a copy of the MPL was not distributed with this\n\n# file, You can obtain one at http://mozilla.org/MPL/2.0/.\n\n# The following comment must be at the start of a new line.\n\n#\n\n#\n\n# The contents of this file are subject to the Mozilla Public\n\n# License Version 1.1 (the &quot;License&quot;); you may not use this file\n\n# except in compliance with the License. You may obtain a copy of\n\n# the License at http://www.mozilla.org/MPL/\n\n# Software distributed under the License is distributed on an &quot;AS\n\n# IS&quot; basis, WITHOUT WARRANTY OF ANY KIND, either express or implied.\n\n# See the License for the specific language governing rights and\n\n# limitations under the License.\n\n# The Original Code is Mozilla Communicator client code, released\n\n# March 31, 1998.\n\n# The Initial Developer of the Original Code is\n\n# Netscape Communications Corporation.\n\n# Portions created by the Initial Developer are Copyright (C) 1998\n\n# the Initial Developer. All Rights Reserved.\n\n# Contributor(s):\n\n#\n\n# Alternatively, the contents of this file may be used under the terms of\n\n# either of the GNU General Public License Version 2 or later (the &quot;GPL&quot;),\n\n# or the GNU Lesser General Public License Version 2.1 or later (the &quot;LGPL&quot;),\n\n# in which case the provisions of the GPL or the LGPL are applicable instead\n\n# of those above. If you wish to allow use of your version of this file only\n\n# under the terms of either the GPL or the LGPL, and not to allow others to\n\n# use your version of this file under the terms of the MPL, indicate your\n\n# decision by deleting the provisions above and replace them with the notice\n\n# and other provisions required by the GPL or the LGPL. If you do not delete\n\n# the provisions above, a recipient may use your version of this file under\n\n# the terms of any one of the MPL, the GPL or the LGPL.\n\n#\n\n# ***** END LICENSE BLOCK *****\n\nMOCHITEST_MANIFESTS += [\n\n' xpcshell.ini ' ,\n\n' mochitest-glue.ini ' ,\n\n' mochitest-plain.ini ' ,\n\n]\n\nDIST_DIRS += [\n\n' external/ ' ,\n\n' ipc/ ' ,\n\n' modules/ ' ,\n\n]\n\nEXTRA_COMPONENTS += [\n\n' mozLoop.js ' ,\n\n' webextensions.js ' ,\n\n]\n\ndeps = [\n\n'../base ' ,\n\n'../browser/base ' ,\n\n'../browser/components/migration ' ,\n\n'../browser/components/places ' ,\n\n'../browser/themes/shared ' ,\n\n'../content/base ' ,\n\n'../content/html ' ,\n\n'../content/xul ' ,\n\n'../embedding/browser ' ,\n\n'../embedding/components/browser ' ,\n\n'../embedding/components/toolkit ' ,\n\n'../extensions/browser/api ' ,\n\n'../extensions/common/moz.build ' ,\n\n'../gfx/2d ' ,\n\n'../gfx/layers ' ,\n\n'../intl ' ,\n\n'../netwerk ' ,\n\n'../printing ' ,\n\n'../services/sync ' ,\n\n'../toolkit/components/global ' ,\n\n'../toolkit/components/printing ' ,\n\n'../toolkit/components/satchel ' ,\n\n'../toolkit/crashreporter ' ,\n\n'../toolkit/library ' ,\n\n'../toolkit/mozapps/extensions ' ,\n\n'../toolkit/mozapps/extensions/extensions-types.js ' ,\n\n'../xpcom/base ' ,\n\n'../xpcom/build ' ,\n\n'../xpcom/build/shared-nspr ' ,\n\n'../xpcom/build/shared-nspr/src ' ,\n\n'../xpcom/ds ' ,\n\n'../xpcom/ds/io ' ,\n\n'../xpcom/ds/xpcom ' ,\n\n'../xpcom/glue ' ,\n\n'../xpcom/io ' ,\n\n'../xpcom/threads ' ,\n\n'../xpcom/threads/src ' ,\n\n'../xpcom/threads/xpcom ' ,\n\n' ../xpfe/appshell/nsAppShell.cpp ' ,\n\n' ../xpfe/appshell/nsAppShell.h ' ,\n\n' ../xpfe/appshell/nsAppsShellService.cpp ' ,\n\n' ../xpfe/appshell/nsAppsShellService.h ' ,\n\n' ../xpfe/appshell/nsNativeAppSupportWin.cpp ' ,\n\n' ../xpfe/appshell/nsNativeAppSupportWin.h ' ,\n\n' ../xpfe/appshell/nsNativeAppSupportXPConnect.cpp ' ,\n\n' ../xpfe/appshell/nsNativeAppSupportXPConnect.h ' ,\n\n' ../xpfe/appshell/nsWindowRoot.cpp ' ,\n\n' ../xpfe/appshell/nsWindowRoot.h ' ,\n\n' ../xpfe/appshell/nsXULAppAPI.h ' ,]" time="0.099"><properties><property name="score" value="0.44271170105005436" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The world of professional wrestling is always changing, but the superstars themselves can sometimes get lost in the shuffle. Sometimes, the changes within the industry are so drastic that a superstar's popularity can get left in the past. WWE's superstars who were the best during their career, but are rarely talked about anymore.\n\nWWE is still the leader of the wrestling world. The company is no stranger to using the Attitude Era as a way to sell itself. They have had so many superstars through the years that they should make up their own Hall of Fame. Some of those superstars are the ones who are mentioned the most.\n\nThe casual fan knows about Hulk Hogan, Stone Cold, and The Rock. That is what makes this so fun. Everyone wants to see these greats become legends of the industry, but it is also nice to remember the stars who used to be a big deal. Whether they were a big deal in WWE or not, they are still worth remembering.\n\nContinue scrolling to keep reading Click the button below to start this article in quick view Share Tweet Email Copy Link Copied\n\n10 Sin Cara\n\nvia trbimg.com\n\nSin Cara has become a joke of a superstar. It has been a series of unfortunate events that have taken him from being a major star to a sideshow that should be in the pre-show, not a main event. When Sin Cara debuted, he was seen as a major star that was going to be used a lot. It took some time, but it finally happened.\n\nThe problem was that he never delivered.\n\nHe was so unimpressive that he was booted from WWE in 2016. He is making an appearance at WrestleMania 33. For his sake, it would be best if he kept quiet and didn't interact with the crowd.\n\n9 Kelly Kelly\n\nvia tumblr.com\n\nKelly Kelly was a fan favorite during her time in WWE. She made her debut as a popular model who was wrestling on ECW. From there, she rose to the top as a valuable part of the roster. Kelly became a popular figure who had her own dedicated fans. She had the right amount of talent to be a decent worker, but it was her good looks that gave her the extra push.\n\nAfter leaving WWE in 2012, Kelly kept wrestling. She is back on the roster with TNA, but she has not appeared on a television episode of Impact Wrestling since December of 2015.\n\n8 Alberto Del Rio\n\nvia wrestlingrumors.net\n\nIt is a sad truth that Alberto Del Rio has fallen from his former position in the industry. He was a very big part of WWE. He was not only the United States Champion, but also a World Champion. Alberto had a unique presence as a superstar that could do things no one else could do. He also had the microphone skills to back up his moves.\n\nHis tenure in WWE was marred with incidents that made it clear he was not going to be there much longer. His time in Lucha Underground was more memorable, but it was a step down from what he used to be.\n\n7 The Great Khali\n\nvia cagesideseats.com\n\nThe Great Khali is not really a superstar that anyone really misses. His presence was laughable at times. The thing about Khali is that he had a lot of potential. He was very big and very strong, but he never seemed to have any clue about how to use those traits to his advantage.\n\nKhali is now an actor and a former wrestler. He is part of the reality show Khali Kisses, which involves him doing a lot of kisses on people. It is a far cry from his WWE days.\n\n6 Kelly Kelly's Boyfriend\n\nvia youtube.com\n\nWrestlers who are lucky enough to find a wife have a much easier time in the wrestling industry. The same could not be said for Mike Knox. He was only a superstar for a very short time, and he was a part of one of the most bizarre moments in wrestling history when he got the Attitude Adjustment from Kane. Knox is not talked about much these days, but there was a time when people were curious about him.\n\nHe married the popular WWE superstar Kelly Kelly. The couple got married in 2014. It was a marriage that was not meant to last, as the couple divorced in 2016.\n\n5 Santino Marella\n\nvia ringsidenews.com\n\nSantino Marella will forever be a part of WWE. The WWE Hall of Fame is home to some great wrestlers and some very bad wrestlers. Santino is among those who have made an impression for one reason or another. His career with the company was not as successful as he had hoped, but he made a huge impact. He was the guy who was hilarious because he was so bad.\n\nSantino's last match was in 2014. He has done some backstage work, but the star has pretty much disappeared from WWE.\n\n4 Tyson Kidd\n\nvia heavy.com\n\nTyson Kidd was a true performer in WWE. He had a unique look, and he could really do it all. The man was skilled on the mic, in the ring, and he had a good personality. Everything came together for Tyson, but that would all change during an unfortunate injury. Tyson was a mid-carder who could have risen to a top level, but that was not meant to be.\n\nKidd is now recovering from his neck injury. He is still hoping to get into the ring again, but it is looking unlikely.\n\n3 Christian\n\nvia bodybuilding.com\n\nChristian is not a name that is being mentioned much these days. It is a shame, as he is one of the best performers of his time. Christian has made a career of being known for his impressive in-ring work and his ability to entertain fans. Christian was a popular superstar during his time in WWE and he did some good things for TNA.\n\nChristian has not been in the ring since December of 2015. He is trying to get back into the ring, but there is a possibility that his neck injury may have ended his career.\n\n2 Dolph Ziggler\n\nvia wrestlingnewssource.com\n\nDolph Ziggler had the chance to become a very big star in WWE. Unfortunately, things have not been going well for Ziggler. He was once a main eventer, but he has fallen to the bottom of the card. He is one of the many guys who will be pushed as the underdog who is going to have the moment of their life.\n\nZiggler has not been the focus of WWE. It is likely that he will be entering a new chapter of his career in the near future. The most likely place would be in Ring of Honor, but WWE has shown a lot of interest in Ziggler.\n\n1 Wade Barrett\n\nvia wrestlingrumors.net\n\nIt is going to be very hard for the casual fan to remember Wade Barrett. He had a short run in WWE, but he did his part as a mid-carder. Barrett is a former Intercontinental Champion. He did well in WWE as a member of The Nexus, but his star power never quite took off.\n\nBarrett has been out of WWE since 2014. He has had some ups and downs in the independent circuit, but his career has largely gone downhill. It is possible that he could get another chance with WWE. That would probably be the only way that fans would get to see him again.\n\nBarrett has been trying to make a name for himself in the business world. He launched a website last year. It is not likely that he will become a success, but it is nice to see him try to make something of his life.]" time="0.097"><properties><property name="score" value="0.5388144646339252" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.53881446&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.53881446
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[University: The primary mission of the University is to provide education. For that purpose, the University has established different faculties and departments, of which the oldest and most important is the faculty of theology. The latter is a privileged entity, by virtue of its academic reputation, both in the local and the global sphere. The mission of the University is to educate the professors and the students in the science of theology, which means, it also needs to equip them with the specific professional skills, which make them able to exercise the profession of the ministry.]" time="0.026"><properties><property name="score" value="0.06968902871701266" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The brother-in-law of the Palestinian-American teenager who was filmed being brutally beaten by Israeli police at a West Bank checkpoint says the incident \u201chas put a wedge\u201d between the Palestinians and Israelis.\n\nYoussef Almogy said he has had more than 50 phone calls from news media outlets following the brutal arrest of 16-year-old Tariq Abu Khdeir, the Associated Press reported.\n\n\u201cWhen I saw it, I cried,\u201d Almogy said. \u201cI was so shocked, but after that, I started thinking about my brother, who was jailed for 11 years, and then released for nothing. And I said, \u2018If they can do that to him, then imagine what they can do to us.\u2019\u201d\n\n\u201cAnd this has put a wedge between Palestinians and Israelis,\u201d he added.\n\nAlmogy was referring to the arrest of his brother, Amer, who was jailed for 11 years after he and a group of Palestinians tried to rob a Tel Aviv bank in 1994. Almogy said Amer was never convicted of a crime and that he believes the former prisoner is being held because of his brother\u2019s connection to the teen, the AP said.\n\n&quot;We're used to seeing this sort of thing,&quot; he said.\n\nAbu Khdeir, who was born in Florida but living with his family in East Jerusalem, was arrested Thursday after he was identified as being involved in a clash between]" time="0.050"><properties><property name="score" value="0.3857518696008255" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.38575187&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.38575187
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[\n\nToday we\u2019re excited to announce that Bethesda Game Studios has joined Oculus as a developer for the Oculus Rift. The partnership will see the creation of two exclusive, made-for-VR games from Bethesda Game Studios; the first is currently being designed for virtual reality and the second is being built for traditional monitors.\n\n\n\n\u201cAt Oculus, we\u2019re dedicated to helping developers and partners to grow, and the support and dedication we\u2019ve received from Bethesda Game Studios has been outstanding,\u201d said Brendan Iribe, CEO of Oculus. \u201cJohn Carmack, founder of id Software and now CTO at Oculus, believes in our vision and has created a custom version of Doom 3 BFG Edition for Oculus. We look forward to continuing to work with Bethesda as we develop our first generation of virtual reality games and experiences.\u201d\n\n\n\n\u201cI have a lot of personal interest in making sure that the VR gaming space is successful because I really believe in the technology and its potential to change gaming,\u201d said Carmack. \u201cAs a developer, the Rift is something that I\u2019ve wanted to work with for a long time, but the existing technology wasn\u2019t quite there yet. I\u2019m happy to be able to help to make the Rift a success and contribute to the rebirth of VR gaming.\u201d\n\n\n\nAt Bethesda Game Studios, developers are working on a number of unannounced projects including new entries in major franchises. Those who have seen this year\u2019s E3 show floor, Gamescom or Quakecon presentations have seen teases of what the studio is up to and a few of the titles we have in development are also being designed for virtual reality. The two games announced today are entirely new IP and we can\u2019t wait to show them to you when they\u2019re ready for primetime.\n\n\n\nFor those of you unfamiliar with Bethesda Game Studios, the acclaimed developer is responsible for creating many of the industry\u2019s most popular games including the \u201cElder Scrolls\u201d series, the single-player add-on to the original \u201cDoom\u201d and the \u201cFallout\u201d series. In fact, \u201cFallout 3,\u201d \u201cFallout: New Vegas\u201d and \u201cThe Elder Scrolls V: Skyrim\u201d were each awarded game of the year at the 2008, 2010 and 2011 BAFTA awards, respectively.\n\n\n\n\u201cWe have made some long-term technology investments with VR in mind,\u201d said Todd Howard, Game Director and Executive Producer at Bethesda Game Studios. \u201cHardware is finally catching up to the visions developers have had for virtual reality. Now when you put on a VR headset, you\u2019re transported to new and amazing worlds.\u201d\n\n\n\nThis year, Bethesda Game Studios showcased \u201cDoom 3 BFG Edition\u201d for Oculus. \u201cDoom 3 BFG Edition\u201d is a complete overhaul of \u201cDoom 3\u201d that brings modern technology and a classic, storied franchise into the world of virtual reality. This is the first title designed for virtual reality.\n\n\n\n\u201cJohn Carmack\u2019s work in the early 90s helped usher in the era of modern 3D gaming, and his genius continues to inspire developers to innovate,\u201d said Brendan Iribe, CEO of Oculus. \u201cThe team at Bethesda Game Studios is made up of the most talented developers in the industry and we can\u2019t wait to bring their newest games, made for virtual reality, to life.\u201d\n\n\n\nOver the past 20 years, Bethesda Game Studios has earned its reputation as one of the industry\u2019s most respected and accomplished game development studios. Creators of the 2006 and 2008 Game of the Year, \u201cThe Elder Scrolls IV: Oblivion\u201d and 2010 Game of the Year, \u201cThe Elder Scrolls V: Skyrim,\u201d Bethesda Game Studios has earned a reputation for excellence and quality with each of its five releases.\n\n\n\nThis partnership is another step toward realizing the future of virtual reality. It\u2019s an exciting time to be a gamer with the release of these groundbreaking VR titles, the arrival of next generation consoles and the continued evolution of the PC.\n\n\n\nStay tuned for more news on Bethesda Game Studios\u2019 titles for Oculus.\n\n\n\n\u2013 The Oculus Team]" time="0.085"><properties><property name="score" value="0.1127998710733491" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.11279987&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.11279987
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[\n\nHi~ B een a long time no see. Thank you for the continued support on the blog. So, how have you all been doing?\n\nToday, I have a special post. I know most of you do not know but it is the 'World Animal Day'. The day is celebrated worldwide on October 4 to create awareness of animal rights and welfare. The day is a very special one for me because I am a vegetarian, animal lover and a dog-person.\n\n\n\n\n\nSo for the last one year, I have been trying to spread the awareness among the people. I have been trying to encourage people to stop using animals as fur coats, cosmetics and for food. I want people to know the feelings of these animals, the pains they suffer and what they go through while they are alive. Also, the environmental damage they cause to the planet. It is very important to spread this message because animals are treated as if they are some sort of toys that we can play with.\n\n\n\n\n\nToday, I would like to share with you some tips on how we can help the animals and make the planet a better place to live. I also have a small personal pledge to make. Also, if you have not signed the pledge yet, you can also make your own pledge to protect the animals and the planet.\n\n\n\n\n\n\n\n\n\nTips and pledges to help the animals\n\n(made by my friend!)\n\nOne way we can make a change is by making an effort to not use cosmetics that are tested on animals. We can choose to use all cruelty free products and make a conscious effort not to use anything that has been tested on animals.\n\nWays to reduce our carbon footprints and help protect the environment:\n\nAvoid eating meat, or even better, go vegetarian (dairy, eggs are not vegan but I still encourage you to stop using dairy products as well)\n\nUse eco-friendly products - Not just cosmetics but also in our everyday lives.\n\nGo cruelty free - Choose cruelty free products and wear clothes that have not been made out of leather, fur or wool.\n\nStop buying those animals that we do not need. Also, choose to buy second hand or recycled clothes. Avoid buying pets if you cannot provide the proper care for them.\n\nAdopt, don't buy! You can adopt a dog or cat from an animal shelter. Do not buy animals from pet stores or breeders. They are sold to you as &quot;breed-free&quot; animals but most of them have been bred in horrible conditions. Also, make sure you know the sex of the animal you are getting, because they usually are sold as 'un-neutered' or 'un-spayed' animals.\n\nMake a conscious effort to not use plastic bags. Use reusable bags or better still, try to avoid using bags at all.\n\nTry to use products that do not harm the environment and are biodegradable.\n\nWays to help the animals:\n\nSign the pledge - To show that you support the cause to protect the animals.\n\nShare the post - Let others know of the cruelty that the animals go through. Let them know how they can help the animals too.\n\nJoin or start a protest or campaign to help the animals.\n\nThe small pledge I am making is to help the animals. Every month, I will be donating some money to an animal shelter in my town to help feed the dogs and cats there.\n\nThese are just some of the ways we can help the animals and the planet. We can all make a change, and make this planet a better place to live.\n\n\n\n\n\nIf you want to know how to become a vegan, and which cosmetics are tested on animals, you can read this post: How to be a vegan?\n\nRemember, we are what we eat, so be what you want to be. Be a good person, who loves animals and helps to save the planet.\n\nWith lots of love,\n\nPriya]" time="0.098"><properties><property name="score" value="0.00786284861861719" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00786285&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00786285
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Top Customer Reviews\n\nWhere reviews refer to foods or cosmetic products, results may vary from person to person. Customer reviews are independent and do not represent the views of The Hut Group.\n\nUnique, gritty, and wonderful The Good: The worlds are some of the best I've seen in a while, and how you explore the city and interact with its citizens gives a great sense of authenticity and intrigue. As a huge comic fan, the references are abound. The story and narrative is one of the best I've ever experienced. The Bad: The game is very linear and the combat while satisfying can get old at times. Overall: This is one of the best games I've ever played, period. Everything about it is gritty, dark, atmospheric, and enticing. While you have a very linear path to follow in the main story, the side quests and interactions you can have with the citizens is one of the best I've seen. You are truly the villain in this game, so if you want to be good and helpful to everyone, this is not for you. But for the fans of dark, gritty stories, this is one of the best. The story and narrative are top notch. The villains are some of the most compelling I've ever seen in a video game, with even the main villains being so much more compelling than the good guys. I loved this game from start to finish, and can't wait to see where the series goes from here. The ending while great is somewhat confusing, so if you aren't]" time="0.049"><properties><property name="score" value="0.21473640003200067" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Mrs. Mahilani Konczey\n\n&quot;The tutors at HW have made a huge difference in my child's academic performance, as well as in her confidence. The tutors are friendly, professional and punctual. My child enjoys her sessions and she has learned a lot from them. HW has a good location and easy parking. I highly recommend them.&quot;]" time="0.028"><properties><property name="score" value="0.14251445744389857" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.14251446&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.14251446
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Flavor Pairs\n\n(Home)\n\nPickled Meat\n\nBacon, Flank Steak, Pork Butt, Bacon, Hams, Sausage, Canned Ham\n\nFor Cooking/Grilling\n\nPeppers, Onions, and other Vegetables\n\nPotatoes, Onion, Green Peppers, Jalapenos, Garlic, and other Vegetables\n\nPastas\n\nPasta, Macaroni, Shells, Ramen Noodles, Ziti, Gnocchi, Fusilli, and other Pastas\n\nSoups and Stews\n\nSoups, Stews, Chili, and other One-Pot Meals\n\nCorn, Peas, and other Grains\n\nCorn, Peas, Risotto, Polenta, Grits, Quinoa, Rice, and other Grains\n\nBeverages\n\nBeer, Alcohol, Soda, Juices, Tea, Coffee, Soda, and other Beverages]" time="0.029"><properties><property name="score" value="0.8464864388901261" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.84648644&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.84648644
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[And now, ladies and gentlemen, I give you the breathless high-speed ranting I did when I saw the (potentially) wackadoo conclusion to this storyline.\n\nOh, my God, you guys, this episode has a freaking atomic bomb on the side of it.\n\nOkay, cool. It\u2019s the mid-season finale. I know there\u2019s going to be some serious plotlines.\n\nAnd seriously, what was up with the weird science in this episode?\n\nNo. No, I did not just watch an episode with the headline: \u201cOuch, My Butt!\u201d\n\nThat plotline was weird as hell.\n\nBut I can understand why they did it. He\u2019s trying to be a \u201csophisticated ladies man\u201d and not the boy they\u2019ve been dealing with all these years.\n\nOh, now this is hilarious. But it\u2019s also dangerous, if you\u2019ve watched any superhero movie ever.\n\nIn the brief moment that I was allowed to breathe, I actually enjoyed this part.\n\nNot only is this scientifically improbable, but also\u2026 how does a security camera shoot a laser?\n\nI know this is only in fiction, but I do believe that they\u2019re committing suicide.\n\nThis is also part of the reason why I wanted to talk about this episode. I know it\u2019s not a good look. It\u2019s not exactly smart of him to do this.\n\nI don\u2019t even know what to say here.\n\nWHY?\n\nHe\u2019s not afraid of bugs, but he\u2019s scared of that snake?\n\nI mean, yeah, I get that he\u2019s doing it for a different reason, but\u2026\n\nIf it weren\u2019t for that, then I would not have laughed so hard at this.\n\nOkay, look, I was done with this season for a while. I thought that it was becoming a boring clich\xe9d plot that they were sticking to, so I was ready to say goodbye to it. But then I watched this, and I just felt like I was watching a new show. There was so much going on, and they were throwing so much at us at once, and it was just such a good use of time. I\u2019m really happy that they had this show, because it was seriously pretty funny and it was also action-packed. There was a lot going on, but it was very clear, and it was fun to watch. I just loved how much was going on. I loved it, and I really did enjoy it. It was a bit much, and they took a lot of crazy chances, but it was just so fun to watch. That was some crazy stuff, and it was nice to see something so good after so much of the regular garbage we were getting. Overall, I loved it.]" time="0.095"><properties><property name="score" value="0.46968671381229327" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.46968671&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.46968671
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Specification:\n\nProduct Name Bags Model SA01A Material Canvas + Zipper Size 29*15.5*10.5cm Color Red, Green, Black, Blue, Gray, Purple, Brown, Khaki, Yellow, Orange Weight 10g Quantity 1 Piece(s) Pattern Solid Inner Pocket Yes Closure Drawstring\n\nPackage include:\n\n1 X Canvas Bags\n\n\n\n\n\nDescription:\n\nDimensions of the storage bag, cool way to store the supplies for travel and daily life.\n\nLightweight, durable and sturdy canvas construction.\n\nA simple pouch, plenty of storage space to put many small items.\n\nThere is a handle for easy carrying and a good choice for daily life.\n\nUsed for carrying, practical, suitable for outdoor, camping, traveling, adventure, daily life, perfect for storing jewelry, makeup, keys, etc.\n\n\n\n\n\nSize: 29*15.5*10.5cm\n\n\n\nWeight: 10g\n\nMaterial: Canvas + Zipper\n\n\n\nColor: Red, Green, Black, Blue, Gray, Purple, Brown, Khaki, Yellow, Orange\n\nPattern: Solid\n\nInner Pocket: Yes\n\n\n\nClosure: Drawstring\n\n\n\n\n\n\n\nPackage include:\n\n1 X Canvas Bags]" time="0.026"><properties><property name="score" value="0.4372006411832996" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Author:\n\nFollowing a stint as a park ranger in Wyoming, Dan Beal relocated to Seattle and received a degree in science writing from the University of Washington.\n\nFollowing a stint as a park ranger in Wyoming, Dan Beal relocated to Seattle and received a degree in science writing from the University of Washington.]" time="0.024"><properties><property name="score" value="0.3393735529644837" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.33937355&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.33937355
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[(Adapted from the 2014 Urban Tree Canopy Conference)\n\nPotholes are problematic for all of us, and for our city\u2019s trees they represent a real danger. Roots can become entangled in the small cracks between the asphalt and the cement to create the impression that the tree is growing out of the road! And when the roots do become entwined, the pressure can be so great that the roots are destroyed. We all know the danger this creates for the trees.\n\nThe solution to this problem is obvious: pave over the whole road surface. Not so fast! In the City of Chicago, for example, there is a practice that preserves the utility of the tree\u2019s roots and of the road surface itself. The practice is called \u201cmilling\u201d. Milling involves cutting the asphalt out in a shape that will not disturb the roots and adding back in a new surface. While the tree is protected, the roots are not disturbed and the road surface is left intact. When a tree is scheduled for removal, this is often the method of choice. Milling also involves improving the root zone around the tree. The roots of many trees can\u2019t penetrate the compacted soil that is found beneath pavement, so it is necessary to mill the area to promote root growth.\n\nMilling is less expensive than paving over a road surface and can also be less expensive than removal of a tree. Milling should be done whenever possible and certainly before a tree is scheduled for removal. There is no need to sacrifice a mature, healthy tree.\n\nAdvertisements]" time="0.046"><properties><property name="score" value="0.2669239211351704" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[How do i search for files with a specific suffix using Unix?\n\nOn Unix, to search for files that ends with a particular file extension, use the following command:\n\nfind . -name '*.html'\n\nOR use grep command to list all files in a directory and subdirectories that have the suffix:\n\nfind . -type f | xargs grep -l &lt;suffix&gt;\n\nSample outputs:\n\n./Ajax/contact-us/index.html ./test/index.html\n\nThe above commands are equivalent to the grep command on Linux/MacOSX or the find command on Linux/MacOSX:\n\n$ find . -name &quot;*.html&quot; ./test/index.html ./Ajax/contact-us/index.html $ grep -l -Ri -f -E &quot;.*\\.html&quot; ./ test/index.html Ajax/contact-us/index.html\n\nFind out all files with a specific filename (using shell pattern match):\n\nfind . -name '*.html'\n\nOR use the egrep command to list all files in a directory and subdirectories that have the suffix:\n\nfind . -type f | xargs egrep -l &lt;suffix&gt;\n\nSample outputs:\n\n./Ajax/contact-us/index.html ./test/index.html\n\nOR\n\n$ find . -name &quot;*.html&quot; ./test/index.html ./Ajax/contact-us/index.html $ egrep -i &quot;.*\\.html&quot; . Ajax/contact-us/index.html test/index.html\n\nFind files in a directory and subdirectories:\n\nfind . -name &quot;*.html&quot;\n\nOR\n\nfind . -type f | xargs grep -l &lt;suffix&gt;\n\nSample outputs:\n\n./Ajax/contact-us/index.html ./test/index.html\n\nOR\n\n$ find . -name &quot;*.html&quot; ./test/index.html ./Ajax/contact-us/index.html $ egrep -i &quot;.*\\.html&quot; . Ajax/contact-us/index.html test/index.html\n\nFind files in a directory and all]" time="0.047"><properties><property name="score" value="0.3291496555819229" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Russia has expanded its ban on U.S. pork to include all American meat imports as the diplomatic row over alleged election meddling and the situation in Syria rumbles on.\n\nLast week, the Russian government added five new countries to the blacklist: South Korea, Montenegro, Iceland, Liechtenstein and Ukraine. While the list included major agricultural producers such as Brazil and Canada, U.S. meat producers were noticeably absent. However, on Wednesday, the government added the U.S. to the list of banned meat importers.\n\nThe ban was announced by the state consumer watchdog Rospotrebnadzor. A government official told Russian media that American meat imports were now being kept out of Russia on a \u201ctemporary basis\u201d.\n\nRussia is the fourth largest importer of U.S. meat, buying $51 million worth of poultry and $22 million of pork in 2016. According to the National Pork Producers Council, the U.S. shipped 27,600 metric tons of pork to Russia in 2016, which made up 0.3 percent of total U.S. pork exports.\n\nRead more: Polls Say Few Americans Think Trump is Putin's Puppet\n\nThe ban comes at a particularly difficult time for the U.S. meat industry, which is still struggling to recover from a disastrous salmonella outbreak that began in 2015. China, Vietnam and Mexico have all banned U.S. pork imports, while the EU and South Korea are still testing shipments.\n\nThe bans come as U.S. meat exporters have had difficulty convincing importers to trust American meat following the 2015 crisis. The outbreak, which resulted in more than 60 percent of pork being recalled, was linked to at least 184 illnesses in 20 states. While U.S. authorities traced the outbreak back to a farm in Missouri, they were unable to identify the exact source of contamination.\n\n\u201cThis has been a major setback for us and it will take a while for the industry to recover,\u201d said National Pork Producers Council spokesman Dave Warner.\n\nIt's not clear when Russia will lift its ban on U.S. meat, but Russia's state consumer protection watchdog has said it would allow Ukrainian meat imports.]" time="0.063"><properties><property name="score" value="0.6871628742597946" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.68716287&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.68716287
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Is there anything you would change about this book?\n\nIt's too long. I got bored. It's the wrong length for a book about the Crusades. I know you want to make it as comprehensive as possible, but let's face it, the Muslim perspective is the only one that matters. You can understand why people fought for this or that minor piece of territory, but you can't understand why they fought in the first place.\n\nHow would you have changed the story to make it more enjoyable?\n\nI wish they had done a better job of establishing the atmosphere of the time, instead of just going for sensationalism.\n\nHow could the performance have been better?\n\nThe narrator had a good voice for this.\n\nWas Crusade the right choice for you? Why or why not?\n\nNo, I'd never read about the Crusades before.]" time="0.022"><properties><property name="score" value="0.34538476559653475" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Most Helpful Positive Review\n\nI would purchase this product again\n\nI have used a baby wrap before for a previous baby and loved it, but this one is my favorite. The material feels great and is breathable, the design is cute and easy to use, and I love...Read complete review\n\nI have used a baby wrap before for a previous baby and loved it, but this one is my favorite. The material feels great and is breathable, the design is cute and easy to use, and I love how it fits under my shirt to show a little but isn't quite as obvious as a sling. I use this for my 1 month old and we have never had any problems with the carrier not staying tight enough. I would definitely recommend this product!]" time="0.022"><properties><property name="score" value="0.16637482131404147" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[A northwest Houston woman is under arrest, accused of child abuse.\n\nPolice say they were called to the woman's home on Almeda School Road around 10:45 a.m. Thursday. They say her husband had to take their child to the emergency room after the child suffered a burn.\n\nPolice say the child suffered burns to the back of his neck, shoulder and upper back area.\n\nPolice say the wife, 31-year-old Mercedes Victoria Gonzalez, was arrested.]" time="0.021"><properties><property name="score" value="0.3297026107240697" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[(NC)\u2014Despite the odds, there are some consumers who decide to head to the stores this weekend to start Christmas shopping. For some, the reason is because they still have some of their shopping list to complete. Others, however, have an entirely different reason.\n\nAccording to Statistics Canada, the average Canadian adult will spend $473.12 on gifts this holiday season. That means the average consumer will spend almost $35 more this year than last year.\n\n\u201cA lot of consumers have gone overboard,\u201d says Ted Michalos, founder of CMG Financial. \u201cIt\u2019s not necessarily the spending that\u2019s the problem, it\u2019s what they\u2019re spending the money on.\u201d\n\nThat\u2019s because many shoppers are buying gifts for themselves. That doesn\u2019t necessarily mean they\u2019re spending on the latest fashions, but rather toys and electronics. And the majority of these gifts, at 41.8 per cent, are going to be purchased online.\n\nHowever, the reason why so many people are spending online is because they\u2019re trying to avoid the crowds and waiting in lines.\n\n\u201cSome consumers just get too stressed and they go online to get their stuff,\u201d says Michalos. \u201cAnd if they do that, they\u2019re likely going to be paying more for the gifts.\u201d\n\nThat\u2019s because while there are fewer people at the store, they\u2019re paying more per person. For example, if someone is paying $50 for a gift in the store, they\u2019re likely to pay more than $50 online.\n\n\u201cThis is really the time where you\u2019re going to have to do your best to have everything done by the 15th,\u201d says Michalos. \u201cAnd it\u2019s definitely a good idea to sit down with your spouse or significant other and look at your budget. And when you have a child, the best thing to do is to write down what you\u2019re going to spend. That\u2019s a real good thing to do, because if you don\u2019t, you\u2019ll overspend.\u201d\n\nAnd that could come with consequences, because if you\u2019re overspending, you could be limiting your ability to spend on your children or other family members in the future.\n\n\u201cYou can go overboard and give away everything you have to your kids and that could be a problem,\u201d says Michalos. \u201cThis could be an opportunity to teach your kids about moderation, and also the difference between \u2018I want\u2019 and \u2018I need.\u2019\u201d\n\nBecause with Christmas, most people are shopping with the mindset of \u2018I want,\u2019 not \u2018I need.\u2019 And that can be a problem, because it will prevent you from saving or investing.\n\n\u201cIf you\u2019re going to be shopping on a credit card, you\u2019re going to be in big trouble,\u201d says Michalos. \u201cIf you\u2019re not able to pay off that debt by the end of January, you\u2019re going to be in big trouble.\u201d\n\nWant to know more about debt? Try these articles:\n\nNew Christmas ad showcases true meaning of the holidays\n\nTake advantage of these holiday deals\n\nHow to get your kids to be thankful for what they have\n\nWhat do you think? Join the conversation in the comments below or contact us using the \u201cHave Your Say\u201d form on this page.]" time="0.081"><properties><property name="score" value="0.013214120874886307" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Developing in a Virtual Machine on your computer or laptop\n\nHello , Here we are going to learn how to create virtual machine on your computer for testing purpose. Before we start , it is better to know what is virtual machine. A virtual machine is a simulated hardware that allows you to run other operating systems (or other versions of the same OS) at the same time as your current OS, inside a window on your current OS. It can also be thought of as a computer inside a computer. With a virtual machine, you can: Run another operating system (or another version of the same operating system) at the same time as your current operating system. Try out a new operating system without having to install it on your hard disk. Run applications for one operating system on another, without having to reboot your computer.\n\nStep 1: Create a New Virtual Machine\n\nVirtualBox provides a wizard that will assist you in creating a virtual machine. Launch the program and select \u201cNew\u201d.\n\nStep 2: Create a Virtual Machine\n\nThe first window you see is the New Virtual Machine Wizard. Select \u201cCreate a virtual machine\u201d.\n\nStep 3: Create a Virtual Machine\n\nIn the next window, enter the name for your virtual machine. This name will appear in the main VirtualBox window next to the icon for the virtual machine.\n\nStep 4: Create a Virtual Machine\n\nThe next window is where you select the operating system you want to install.\n\nStep 5: Create a Virtual Machine\n\nIn the next window, select the amount of memory you want to allocate to your virtual machine. The more memory you have, the more applications you can run simultaneously. However, keep in mind that the more memory you allocate, the slower your computer will be.\n\nStep 6: Create a Virtual Machine\n\nSelect \u201cCreate a virtual hard disk\ufffd]" time="0.047"><properties><property name="score" value="0.23405024230799265" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The internet is a funny place, where people can get away with saying and doing the weirdest stuff that would otherwise be totally unacceptable.\n\nBut it turns out not even the internet is a free pass to sexually harass people.\n\nThat's what one woman learned when she was sent a totally inappropriate, NSFW photo from a stranger, that happened to be someone she knew in real life.\n\nTwitter user @kendtee, who describes herself as &quot;an arm candy&quot;, tweeted about her encounter with the man she knew.\n\nAccording to @kendtee, the incident occurred after she agreed to take a picture with a guy. She then deleted the picture from her camera roll when it was over, and the man asked if she had. She said yes, and then he said this:\n\n&quot;It's a shame we can't see the photo anymore because it was pretty good&quot;\n\nShe didn't know how to respond. He then tweeted her back, asking her to &quot;leave [the photo] up for a little while&quot;. She declined.\n\nApparently, he then texted her back, asking her to send him a copy of the picture.\n\nShocked by this behaviour, she shared the exchange on Twitter, and said the worst part was how &quot;unbothered&quot; he seemed about it.\n\nEventually she shared a screenshot of the exchange, and people were not impressed.\n\n&quot;Oh wow. 'I'd do ya if I could' and then the photo request. You should probably just block him and be done with it. 'Doing ya' or not,&quot; one person said.\n\nAnother person had an idea on how to deal with it, tweeting, &quot;You know what? Be a bigger person, send it to him. For posterity.&quot;\n\nMany others were in complete agreement with this sentiment.\n\n&quot;I don't know this man and I probably never will. But I 100% understand why he's going to be blocked. A man who says that is definitely not a good person. All women should know their boundaries with men and this seems to me to be clear cut,&quot; one person said.\n\nIn the end, @kendtee went back to him and asked him to explain why he'd sent her that photo.\n\nAccording to @kendtee, he sent her this message in response:\n\n&quot;How do you think I know that you're gonna see this tweet and block me or delete it.\n\n&quot;I'm not trying to be a creep you just said that you'd do me so I figured we had a shot lol\n\n&quot;I like your style and want to get to know you better\n\n&quot;If you send the picture I promise not to be creepy and stop bothering you\n\n&quot;I'm being serious here and I don't think I deserve the reaction I got\n\n&quot;All I wanted to do was take a picture with a nice girl and see if she'd let me\n\n&quot;I don't want to be a bad person or make you feel uncomfortable.\n\n&quot;If you like I'll just go away\n\n&quot;You're gorgeous and I'm sorry I don't deserve to meet you but I thought I'd try my luck\n\n&quot;I won't message you again if you don't want me to\n\n&quot;I'll delete the tweet and block you\n\n&quot;I just wanted to try\n\n&quot;Please&quot;\n\nThe rest of Twitter wasn't impressed, calling him out for still not getting it, and for trying to shift the blame.\n\n&quot;Like you have to explain what 'likes your style' means, like you're trying to shift the blame. But it's not ok. It's not ok for you to message someone for a photo. It's not ok for you to say 'I'd do you' and it's not ok for you to try and justify your actions,&quot; one person said.\n\nAnother pointed out that he was trying to pressure her into sending him the photo, and wasn't taking responsibility for his actions.\n\n&quot;He's a creep, and still wants to force you to send him a picture of your body. He'll delete the tweet if you do, otherwise you can block him. This is on him, not you. You have no responsibility to 'meet' him,&quot; another person said.\n\n&quot;Ok let's stop there with the 'I would've taken a photo with you anyways' thing. That's not what this is about. This is about someone sending unsolicited nudes,&quot; another person pointed out.\n\nIn response, @kendtee wrote that she wasn't bothered, because at the end of the day she was a &quot;pretty woman who didn't need to be respected&quot;.\n\n&quot;I didn't get mad or freak out or feel the need to drag his name through the mud. I just calmly informed him that I have a man in my life and that I'm flattered but I don't share photos like that and I'm taken. He didn't like it so he blocked me and I'm okay with that,&quot; she wrote.\n\nALSO ON HUFFPOST AUSTRALIA]" time="0.085"><properties><property name="score" value="0.6766505957179619" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.6766506&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.6766506
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Perhaps one of the most beautiful areas of Southern California is the Pacific Palisades, a neighborhood filled with high rises, luxury estates, and tons of culture. From the obvious to the under the radar, Pacific Palisades has some of the most fascinating museums, gardens, and activities you can visit. When you need to get out of the house and explore a new part of Los Angeles, the Pacific Palisades is a fantastic choice.\n\nThe Getty Villa\n\nThe Getty Villa is a sight to behold, and is a must see when you are in the Pacific Palisades. You are welcome to explore the museum for free on the first Thursday of every month. They offer a free trolley that will take you to the Getty Villa, so you don\u2019t have to worry about navigating through the Pacific Palisades on a crowded day. This is the perfect time to visit the Pacific Palisades.\n\nThere are two events going on at the Getty Villa on Thursdays: Story Time and Pacific Voices. Story time will give you a better understanding of ancient Greek and Roman culture. It is a good idea to purchase your tickets in advance, so you can get a discount. The Pacific Voices event is a singing performance, and a great way to enjoy your time at the Getty Villa.\n\nSanta Monica Mountains National Recreation Area\n\nThe Santa Monica Mountains are a beautiful way to spend a day in the Pacific Palisades. The National Recreation Area has a fantastic number of trails you can explore, and is a great way to get out of the house and spend time with your family. Take the trail to Leo Carillo State Beach, or go for a hike along Rustic Canyon. Whichever trail you take, you are sure to enjoy your time in the Pacific Palisades.\n\nSanta Monica Baykeeper\n\nThe Santa Monica Baykeeper is a fantastic organization that works hard to clean up the Santa Monica Bay. This is a great way to spend time in the Pacific Palisades and make a real difference. Check their website for upcoming volunteer opportunities. You will learn more about the Santa Monica Bay, and get a chance to spend time with some interesting people.\n\nPacific Palisades Center for the Arts\n\nThe Pacific Palisades Center for the Arts is an excellent way to spend an afternoon. This isn\u2019t a museum, so you don\u2019t need to book in advance. It is free to walk around the grounds, and to watch a performance. They have everything from jazz concerts to visual arts. If you want to learn more about the Pacific Palisades, then this is a great place to start.\n\nPacific Palisades Farmers Market\n\nThe Pacific Palisades Farmers Market is an excellent way to spend your Saturday in the Pacific Palisades. From fruits and vegetables, to homemade items and baked goods, there are plenty of delicious items to choose from. This is a great place to bring the kids, as they have tons of activities.\n\nPacific Palisades Art Walk\n\nThe Pacific Palisades Art Walk is a fantastic opportunity to view some of the beautiful artwork created by locals. It is a fun evening filled with laughter and good food. You are welcome to bring your own picnic basket and enjoy it in the beautiful Pacific Palisades.]" time="0.089"><properties><property name="score" value="0.032126342986086666" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[We have been very pleased with the support, workmanship and professionalism of the team at WeFix. They have been able to provide quick and effective solutions for any problems that have arisen. The attention to detail, willingness to understand and understand the importance of high quality work have been important to us. The flexibility of being able to accommodate and adapt to any situations has been very helpful in terms of our team and their working styles.\n\n- Rick Phillips, Site Manager \u2013 Maersk Oil]" time="0.026"><properties><property name="score" value="0.08490881611356027" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Mastering financial knowledge and planning for a good life is a lifelong process. But when it comes to money, how many people have the knowledge to plan for a good life? And what if you had this information, would you use it? How would you live? Would your life change for the better? Would your relationships improve? Would your life be happier?\n\nIn most cases, we don\u2019t have enough financial knowledge to make the right decisions. We don\u2019t have the knowledge and therefore, we end up making bad decisions. The result of these bad decisions is that we lose more than we gain. We lose money and we also lose our peace of mind.\n\nI have experienced both sides of the coin and I know the emotional, mental and physical impact of being financially unstable. Money makes and breaks people. It has broken me in the past and it took me years to recover. But I am better now.\n\nIf you too want to make peace with your money]" time="0.022"><properties><property name="score" value="0.05856926899799886" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Tom Shanks (writer)\n\nTom Shanks (born 1961) is a Scottish writer, journalist, and musician. He is the author of eight novels, including &quot;The Hidden&quot;, the &quot;Vampyrrhic Times&quot; trilogy, and &quot;Eve&quot;, an &quot;Eve&quot; (comics) spin-off novel.\n\nShanks has been living in France since 1983 and is married to the French writer Marie H\xe9l\xe8ne Poitras.\n\nShanks was born in Edinburgh, Scotland in 1961.\n\nShanks has been writing since the mid-1990s and has published several novels.\n\nShanks has also been writing for the British comic book magazine &quot;2000 AD&quot;. His first script for them was the short story &quot;Trevor's Tale&quot;, drawn by Simon Harrison, in &quot;2000 AD&quot; #1704, published in March 2008.\n\nThe comic book &quot;Eve&quot; was written by Shanks. It was released by Black House Comics in 2011 and the second volume was published in 2012. In 2012, IDW Publishing published a comic book prequel to &quot;Eve&quot;, &quot;Eve: The Awakening&quot;. The book was written by Shanks and Marie H\xe9l\xe8ne Poitras and illustrated by Enrica Angiolini.\n\nShanks has also written the script for &quot;Apocalypse Cow&quot;, which was originally published in French and German as &quot;La Vache Apocalypse&quot;. It was adapted as a comic book by Tony Lee.\n\nIn March 2010, the novel &quot;Vampyrrhic Times&quot; was published by Gollancz in the UK and by Roc in the US. It was illustrated by Tony Harwood. The second volume of the trilogy, &quot;The Wisdom of Dead Men&quot;, was published in the UK in 2011 and the US in 2012. The third volume, &quot;The Trade of Queens&quot;, was published in the UK in 2013 and in the US in 2014.\n\n\n\n\n\n\n\nShanks and Poitras run the Black House Comics publishing house.\n\nShanks is a member of the band The Bassholes, along with co-writer of Eve, Marie H\xe9l\xe8ne Poitras. The band is composed of Shanks and Poitras on vocals and bass, Poitras' husband, French artist St\xe9phane Paitreau, on guitar, and drummer &quot;Chino&quot; Gonzales. They released their first album, &quot;No More Mr Nice Guy&quot;, on May 1, 2010, on Stuck on a Name Records. The album includes covers of classic rock songs, such as &quot;Jailhouse Rock&quot; and &quot;Born to be Wild&quot;. In 2013, the band released a second album, &quot;Sex Lies &amp; Basslines&quot;.\n]" time="0.047"><properties><property name="score" value="0.7859141802576686" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.78591418&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.78591418
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Rick Lake\n\nRichard G. Lake II (born April 7, 1979) is an American mixed martial artist who most recently competed in the Welterweight division of the Ultimate Fighting Championship. A professional competitor since 2004, he has also competed for the Quad Cities Silverbacks of the IFL, and was a competitor on Spike TV's &quot;&quot;.\n\nBorn and raised in the Quad Cities area, Lake attended East Moline High School.\n\nAfter graduating from high school, Lake worked for two years as a full-time welder, before deciding to pursue a career in mixed martial arts.\n\nLake made his professional mixed martial arts debut on October 30, 2004 for the Quad Cities Silverbacks of the International Fight League. He won his first three fights for the promotion, before suffering his first loss to Ian Loveland on April 8, 2005.\n\nAfter picking up three wins in the IFL, Lake was signed by the UFC. He was expected to make his debut at UFC Fight Night 7 against Jason Gilliam. However, Gilliam was forced from the card with an injury, and was replaced by Jorge Rivera. Lake lost to Rivera via first-round KO on August 23, 2007.\n\nFor his second fight in the promotion, Lake dropped down to Lightweight and faced Kyle Bradley at UFC Fight Night 11 on November 17, 2007. He won the fight via unanimous decision (30-27, 30-27, 29-28).\n\nIn his third fight for the promotion, Lake dropped to Featherweight to take on Mike Brown at UFC Fight Night 13 on June 16, 2008. He lost the fight via unanimous decision (29-28, 29-28, 29-28).\n\nFor his fourth fight in the promotion, Lake dropped to Bantamweight and faced Brad Pickett at UFC 95 on February 7, 2009. He lost the fight via TKO in the first round.\n\nFor his fifth fight in the promotion, Lake faced Tommy Hayden at UFC 100 on July 11, 2009. He won the fight via submission in the first round, earning &quot;Submission of the Night&quot; honors.\n\nFor his sixth fight in the promotion, Lake faced George Roop on March 21, 2010 at . He won the fight via unanimous decision.\n\nFor his seventh fight in the promotion, Lake fought Roland Delorme on June 19, 2010 at . He lost the fight via submission in the first round.\n\nFor his eighth fight in the promotion, Lake faced John Gunderson on November 20, 2010 at UFC 122. He won the fight via split decision.\n\nFor his ninth fight in the promotion, Lake fought Jason Reinhardt on February 5, 2011 at UFC 126. He won the fight via TKO in the first round.\n\nFor his tenth fight in the promotion, Lake faced newcomer Darren Uyenoyama on June 11, 2011 at . He won the fight via unanimous decision (30-27, 30-27, 30-27).\n\nFor his eleventh fight in the promotion, Lake faced former WEC Featherweight Champion Mike Brown on October 29, 2011 at UFC on Versus 6. He lost the fight via unanimous decision (29-28, 29-28, 29-28).\n\nFor his twelfth fight in the promotion, Lake faced Nam Phan on May 15, 2012 at UFC on FX 4. He lost the fight via unanimous decision.\n\nFor his thirteenth fight in the promotion, Lake faced Sam Sicilia on July 11, 2012 at UFC 149. He won the fight via unanimous decision.\n\nFor his fourteenth fight in the promotion, Lake faced Ivan Menjivar on December 15, 2012 at UFC 155. He won the fight via split decision.\n\nFor his fifteenth fight in the promotion, Lake faced Jeremy Larsen on June 15, 2013 at UFC 161. He won the fight via TKO in the second round.\n\nFor his sixteenth fight in the promotion, Lake faced promotional newcomer Justin Salas on November 16, 2013 at UFC 167. He lost the fight via split decision.\n\nFor his seventeenth fight in the promotion, Lake faced fellow newcomer Max Holloway on February 15, 2014 at UFC 169. He lost the fight via TKO in the first round.\n\nAfter his release from the UFC, Lake dropped down to Bantamweight and made his debut on April 11, 2015 at IFC: International Fighting Championship 2 in Des Moines, Iowa against Rafael &quot;Morcego&quot; da Silva. He lost the fight via knockout in the first round.\n\nLake faced Gabriel Checco on May 28, 2015 at CXF 5. He lost the fight via split decision.\n\nLake faced Daniel Weichel on August 14, 2015 at Superior Challenge 12. He lost the fight via submission in the first round.\n\nIn his fourth fight since returning from the UFC, Lake faced Ryan Quinn on October 23, 2015 at CXF 8. He won the fight via knockout in the first round.\n\nLake faced Jim Alers on January 8, 2016 at CXF 10. He won the fight via TKO in the second round.\n\nLake faced Martin Stapleton on February 19, 2016 at MFW: A Night of Champions. He lost the fight via split decision.\n\nLake next faced Dylan Logan on September 2, 2016 at Cage Warriors 82. He won the fight via TKO in the second round.\n\nOn February 28, 2017, it was announced that Lake had signed with the UFC and was expected to face Alessio Di Chirico on March 18, 2017 at UFC Fight Night 107. However, on March 7, Di Chirico pulled out from the event, citing injury, and was replaced by promotional newcomer Marcin Prachnio. Lake won the fight via TKO in the second round.\n\nLake faced Gregor Gillespie on June 25, 2017 at UFC Fight Night 112. He lost the fight via split decision.\n\nLake faced Jarred Brooks on December 9, 2017 at UFC Fight Night 123. He lost the fight via submission in the first round.\n\nLake faced Drakkar Klose on February 24, 2018 at UFC on Fox 28. He lost the fight via TKO in the second round.\n\nLake faced Andre Fili on August 25, 2018 at UFC Fight Night 135. He lost the fight via TKO in the first round.\n\nLake faced Desmond Green on February 17, 2019 at UFC on ESPN 2. He lost the fight via technical knock out.\n\nIn July 2007, Lake was diagnosed with Crohn's disease, an inflammatory bowel disease that causes ulcers in the digestive tract. He began experiencing symptoms shortly after moving to California to train full-time. His Crohn's was brought under control through medication and diet, and it is currently in remission.\n\n\n\n]" time="0.091"><properties><property name="score" value="0.10590908994341046" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Fenix Funk\n\nFenix Funk is the debut studio album by German electronic music group Telepopmusik. It was released on 9 October 2000 by Columbia Records. It was a success in Japan and peaked at #50 on the Oricon Albums Chart.\n]" time="0.025"><properties><property name="score" value="0.28873503619730734" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.28873504&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.28873504
-    assert 'Human' == 'AI'
-      - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i0]" time="0.077"><properties><property name="score" value="0.5150885098313221" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i1]" time="0.078"><properties><property name="score" value="0.11843365577805336" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i2]" time="0.047"><properties><property name="score" value="0.4060278206221568" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i3]" time="0.044"><properties><property name="score" value="0.211564956720256" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i4]" time="0.046"><properties><property name="score" value="0.47883049812011413" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.4788305&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.4788305
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i354]" time="0.096"><properties><property name="score" value="0.04814062465572828" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i355]" time="0.035"><properties><property name="score" value="0.10794713429921005" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i356]" time="0.017"><properties><property name="score" value="0.05811424640022045" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i357]" time="0.017"><properties><property name="score" value="0.017794853658087417" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i358]" time="0.088"><properties><property name="score" value="0.15031319050105355" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i359]" time="0.036"><properties><property name="score" value="0.04670991987121374" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i360]" time="0.017"><properties><property name="score" value="0.029198669388791343" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i361]" time="0.084"><properties><property name="score" value="0.07864620205418965" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i362]" time="0.018"><properties><property name="score" value="0.07571187157572878" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i363]" time="0.086"><properties><property name="score" value="0.12176820185431772" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i364]" time="0.016"><properties><property name="score" value="0.07842104532147509" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i365]" time="0.090"><properties><property name="score" value="0.09701099885831796" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i366]" time="0.036"><properties><property name="score" value="0.05157020292430404" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i367]" time="0.088"><properties><property name="score" value="0.060180925678193206" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i368]" time="0.085"><properties><property name="score" value="0.10121852350229289" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i369]" time="0.042"><properties><property name="score" value="0.05497163543111265" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i370]" time="0.016"><properties><property name="score" value="0.05855387103061438" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i371]" time="0.084"><properties><property name="score" value="0.071919610568649" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i372]" time="0.034"><properties><property name="score" value="0.059595678096975246" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i373]" time="0.092"><properties><property name="score" value="0.0916661125061234" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i374]" time="0.031"><properties><property name="score" value="0.08389255295419398" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i375]" time="0.015"><properties><property name="score" value="0.046191622409247346" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i376]" time="0.020"><properties><property name="score" value="0.030817053841131203" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255379 (len: 293) is a human-generated sample, misclassified as AI-generated with confidence 0.03081705&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255379 (len: 293) is a human-generated sample, misclassified as AI-generated with confidence 0.03081705
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i5]" time="0.043"><properties><property name="score" value="0.2764518165348784" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i6]" time="0.044"><properties><property name="score" value="0.8021831102545118" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i7]" time="0.043"><properties><property name="score" value="0.10564327337700408" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i8]" time="0.080"><properties><property name="score" value="0.3212529152743504" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i9]" time="0.079"><properties><property name="score" value="0.30264623525501816" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i10]" time="0.044"><properties><property name="score" value="0.0612832114803008" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i11]" time="0.079"><properties><property name="score" value="0.4413940985101468" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i12]" time="0.050"><properties><property name="score" value="0.08847342273077263" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i13]" time="0.081"><properties><property name="score" value="0.21695004566243034" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i14]" time="0.042"><properties><property name="score" value="0.3658530871562815" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i15]" time="0.078"><properties><property name="score" value="0.7835989597065699" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i16]" time="0.079"><properties><property name="score" value="0.6246294715876824" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i17]" time="0.051"><properties><property name="score" value="0.37438603239453583" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.37438603&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.37438603
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i377]" time="0.015"><properties><property name="score" value="0.030523100589469054" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i378]" time="0.029"><properties><property name="score" value="0.09126591721552754" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i379]" time="0.028"><properties><property name="score" value="0.06866283819203234" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i380]" time="0.028"><properties><property name="score" value="0.08349490795474498" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i381]" time="0.075"><properties><property name="score" value="0.06567725197962504" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i382]" time="0.015"><properties><property name="score" value="0.009756390168053075" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255385 (len: 59) is a human-generated sample, misclassified as AI-generated with confidence 0.00975639&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255385 (len: 59) is a human-generated sample, misclassified as AI-generated with confidence 0.00975639
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i18]" time="0.075"><properties><property name="score" value="0.04329430920328262" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i19]" time="0.079"><properties><property name="score" value="0.8365876258744631" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i20]" time="0.030"><properties><property name="score" value="0.5725679449456456" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i21]" time="0.078"><properties><property name="score" value="0.14745490437608386" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i22]" time="0.023"><properties><property name="score" value="0.8019563889197778" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i23]" time="0.073"><properties><property name="score" value="0.836204511415232" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i24]" time="0.046"><properties><property name="score" value="0.255574550480947" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i25]" time="0.043"><properties><property name="score" value="0.2871008314347667" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.28710083&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.28710083
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i383]" time="0.081"><properties><property name="score" value="0.10308135228909965" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i384]" time="0.087"><properties><property name="score" value="0.04636523474985589" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i385]" time="0.099"><properties><property name="score" value="0.04321382686545139" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i386]" time="0.020"><properties><property name="score" value="0.03816017326200294" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i387]" time="0.075"><properties><property name="score" value="0.04782412978681531" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i388]" time="0.017"><properties><property name="score" value="0.03367399083379907" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i389]" time="0.031"><properties><property name="score" value="0.08756692053962213" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i390]" time="0.076"><properties><property name="score" value="0.1648132728572507" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i391]" time="0.090"><properties><property name="score" value="0.0987344846783646" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i392]" time="0.087"><properties><property name="score" value="0.03755271156607985" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i393]" time="0.034"><properties><property name="score" value="0.03236409535822251" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i394]" time="0.078"><properties><property name="score" value="0.003993527340813197" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255397 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.00399353&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255397 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.00399353
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i26]" time="0.076"><properties><property name="score" value="0.14715791336916695" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i27]" time="0.096"><properties><property name="score" value="0.43350122985143696" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i28]" time="0.048"><properties><property name="score" value="0.14288988674229786" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.14288989&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.14288989
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i395]" time="0.070"><properties><property name="score" value="0.07990953717648572" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i396]" time="0.029"><properties><property name="score" value="0.12588195944642253" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i397]" time="0.015"><properties><property name="score" value="0.07227458059381187" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i398]" time="0.088"><properties><property name="score" value="0.1259803230446005" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i399]" time="0.015"><properties><property name="score" value="0.060210802828936094" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i400]" time="0.034"><properties><property name="score" value="0.10504202179735282" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i401]" time="0.027"><properties><property name="score" value="0.008289651686095936" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i402]" time="0.025"><properties><property name="score" value="0.06349607755011993" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i403]" time="0.024"><properties><property name="score" value="0.11119397557286181" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i404]" time="0.024"><properties><property name="score" value="0.09507251236555662" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i405]" time="0.024"><properties><property name="score" value="0.07285755439879514" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i406]" time="0.025"><properties><property name="score" value="0.11681545313810895" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i407]" time="0.015"><properties><property name="score" value="0.041586376472851905" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i408]" time="0.013"><properties><property name="score" value="0.1088073572196091" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i409]" time="0.025"><properties><property name="score" value="0.07572710800385626" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i410]" time="0.025"><properties><property name="score" value="0.0858703811223102" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i411]" time="0.024"><properties><property name="score" value="0.10970850149594066" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i412]" time="0.069"><properties><property name="score" value="0.03988331685584727" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i413]" time="0.024"><properties><property name="score" value="0.08835807309049148" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i414]" time="0.014"><properties><property name="score" value="0.08761653999272423" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i415]" time="0.013"><properties><property name="score" value="0.04279726799708006" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i416]" time="0.013"><properties><property name="score" value="0.04680086427211183" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i417]" time="0.066"><properties><property name="score" value="0.03981918139948222" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i418]" time="0.013"><properties><property name="score" value="0.14035352588276373" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i419]" time="0.013"><properties><property name="score" value="0.03921973199316753" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i420]" time="0.071"><properties><property name="score" value="0.08633259593756293" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i421]" time="0.065"><properties><property name="score" value="0.03836274802289801" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i422]" time="0.027"><properties><property name="score" value="0.12291234809043025" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i423]" time="0.014"><properties><property name="score" value="0.07024297822019188" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i424]" time="0.014"><properties><property name="score" value="0.034430534312673844" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i425]" time="0.075"><properties><property name="score" value="0.04666102743955407" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i426]" time="0.016"><properties><property name="score" value="0.09385350444067742" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i427]" time="0.028"><properties><property name="score" value="0.07498076922844338" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i428]" time="0.027"><properties><property name="score" value="0.08644285342096081" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i429]" time="0.013"><properties><property name="score" value="0.06489646729588294" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i430]" time="0.017"><properties><property name="score" value="0.07903136937538191" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i431]" time="0.030"><properties><property name="score" value="0.0857897364715308" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i432]" time="0.013"><properties><property name="score" value="0.05590667864972532" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i433]" time="0.070"><properties><property name="score" value="0.0978070913266124" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i434]" time="0.027"><properties><property name="score" value="0.0720321953926944" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i435]" time="0.014"><properties><property name="score" value="0.05205873951186213" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i436]" time="0.072"><properties><property name="score" value="0.050292771870896584" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i437]" time="0.029"><properties><property name="score" value="0.11135833001935426" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i438]" time="0.125"><properties><property name="score" value="0.07865459337259147" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i439]" time="0.016"><properties><property name="score" value="0.09202952796041886" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i440]" time="0.078"><properties><property name="score" value="0.09661040462666436" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i441]" time="0.082"><properties><property name="score" value="0.05725167260243741" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i442]" time="0.075"><properties><property name="score" value="0.040887995099585316" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255445 (len: 777) is a human-generated sample, misclassified as AI-generated with confidence 0.040888&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255445 (len: 777) is a human-generated sample, misclassified as AI-generated with confidence 0.040888
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i29]" time="0.042"><properties><property name="score" value="0.45614550726530423" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i30]" time="0.041"><properties><property name="score" value="0.7312638386875459" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i31]" time="0.022"><properties><property name="score" value="0.48846595347146193" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.48846595&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.48846595
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i443]" time="0.086"><properties><property name="score" value="0.07650432689545766" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i444]" time="0.094"><properties><property name="score" value="0.04553570710674554" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i445]" time="0.079"><properties><property name="score" value="0.037724420575059814" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i446]" time="0.039"><properties><property name="score" value="0.03589616162001459" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i447]" time="0.072"><properties><property name="score" value="0.07534858453593579" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i448]" time="0.090"><properties><property name="score" value="0.06515390525729571" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i449]" time="0.027"><properties><property name="score" value="0.11056434278863791" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i450]" time="0.079"><properties><property name="score" value="0.09281360374337588" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i451]" time="0.085"><properties><property name="score" value="0.1447829665964288" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i452]" time="0.080"><properties><property name="score" value="0.027123661787126618" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255456 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.02712366&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255456 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.02712366
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i32]" time="0.070"><properties><property name="score" value="0.5172232311929278" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i33]" time="0.076"><properties><property name="score" value="0.021086478711166794" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.02108648&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.02108648
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i453]" time="0.017"><properties><property name="score" value="0.06521628248180034" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i454]" time="0.017"><properties><property name="score" value="0.01915551553562378" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i455]" time="0.037"><properties><property name="score" value="0.033526060231483176" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i456]" time="0.020"><properties><property name="score" value="0.102096326698764" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i457]" time="0.081"><properties><property name="score" value="0.12239721495856719" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i458]" time="0.020"><properties><property name="score" value="0.08798995215951733" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i459]" time="0.031"><properties><property name="score" value="0.13530657818802122" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i460]" time="0.032"><properties><property name="score" value="0.10198378866095163" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i461]" time="0.076"><properties><property name="score" value="0.20173680972928418" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i462]" time="0.080"><properties><property name="score" value="0.07260578794953715" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i463]" time="0.081"><properties><property name="score" value="0.030725450262465384" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i464]" time="0.088"><properties><property name="score" value="0.11536549740354007" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i465]" time="0.017"><properties><property name="score" value="0.01634675190235746" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i466]" time="0.109"><properties><property name="score" value="0.031006056794248317" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i467]" time="0.020"><properties><property name="score" value="0.017074519452769943" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255471 (len: 296) is a human-generated sample, misclassified as AI-generated with confidence 0.01707452&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255471 (len: 296) is a human-generated sample, misclassified as AI-generated with confidence 0.01707452
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i34]" time="0.070"><properties><property name="score" value="0.7383911060452886" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i35]" time="0.044"><properties><property name="score" value="0.6210401351245953" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i36]" time="0.080"><properties><property name="score" value="0.13742248866865925" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.13742249&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.13742249
-    assert 'AI' == 'Human'
-      - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i37]" time="0.093"><properties><property name="score" value="0.02341469720444171" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.0234147&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.0234147
-    assert 'AI' == 'Human'
-      - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i38]" time="0.081"><properties><property name="score" value="0.11871897353937788" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.11871897&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.11871897
-    assert 'AI' == 'Human'
-      - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i39]" time="0.076"><properties><property name="score" value="0.47608421582007177" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i40]" time="0.074"><properties><property name="score" value="0.05517735401266988" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i41]" time="0.045"><properties><property name="score" value="0.11488925438585262" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.11488925&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.11488925
-    assert 'AI' == 'Human'
-      - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i42]" time="0.081"><properties><property name="score" value="0.5334465225592852" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i43]" time="0.050"><properties><property name="score" value="0.41831894586752405" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.41831895&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.41831895
-    assert 'AI' == 'Human'
-      - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i44]" time="0.077"><properties><property name="score" value="0.09906092012879064" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.09906092&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.09906092
-    assert 'AI' == 'Human'
-      - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i45]" time="0.082"><properties><property name="score" value="0.464935379765367" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i46]" time="0.073"><properties><property name="score" value="0.2559222454237009" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.25592225&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.25592225
-    assert 'AI' == 'Human'
-      - Human
-      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i47]" time="0.073"><properties><property name="score" value="0.3565831527422847" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i48]" time="0.044"><properties><property name="score" value="0.7929543097041081" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i49]" time="0.042"><properties><property name="score" value="0.7437670877539215" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i0]" time="0.073"><properties><property name="score" value="0.2566616652503832" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i1]" time="0.077"><properties><property name="score" value="0.14039542505633884" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i2]" time="0.041"><properties><property name="score" value="0.16731139594272637" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i3]" time="0.039"><properties><property name="score" value="0.10182582784121208" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i4]" time="0.039"><properties><property name="score" value="0.22801555267026874" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i5]" time="0.041"><properties><property name="score" value="0.2602393551909554" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i6]" time="0.038"><properties><property name="score" value="0.046174661689546426" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04617466&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04617466
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i468]" time="0.015"><properties><property name="score" value="0.03494943346383228" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i469]" time="0.017"><properties><property name="score" value="0.07268168871928382" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i470]" time="0.086"><properties><property name="score" value="0.11052142469546371" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i471]" time="0.089"><properties><property name="score" value="0.058504996709617715" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i472]" time="0.035"><properties><property name="score" value="0.05194986435658993" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i473]" time="0.031"><properties><property name="score" value="0.06575473884097523" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i474]" time="0.026"><properties><property name="score" value="0.04751535228796814" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i475]" time="0.075"><properties><property name="score" value="0.12283763143141896" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i476]" time="0.091"><properties><property name="score" value="0.09588575946724305" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i477]" time="0.083"><properties><property name="score" value="0.09346742727536811" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i478]" time="0.092"><properties><property name="score" value="0.09872425883788133" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i479]" time="0.024"><properties><property name="score" value="0.07698613550055078" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i480]" time="0.093"><properties><property name="score" value="0.08718612526083824" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i481]" time="0.014"><properties><property name="score" value="0.0506194963951101" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i482]" time="0.074"><properties><property name="score" value="0.055444328116385985" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i483]" time="0.029"><properties><property name="score" value="0.07794056485863492" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i484]" time="0.017"><properties><property name="score" value="0.03537926585499007" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i485]" time="0.087"><properties><property name="score" value="0.13352028249684772" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i486]" time="0.018"><properties><property name="score" value="0.0417035753297601" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i487]" time="0.087"><properties><property name="score" value="0.05045638235916818" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i488]" time="0.037"><properties><property name="score" value="0.08201142686459388" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i489]" time="0.015"><properties><property name="score" value="0.09225571788019638" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i490]" time="0.098"><properties><property name="score" value="0.11188868166791749" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i491]" time="0.071"><properties><property name="score" value="0.07240720331909846" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i492]" time="0.026"><properties><property name="score" value="0.049205405570832465" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i493]" time="0.012"><properties><property name="score" value="0.051884424680248564" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i494]" time="0.075"><properties><property name="score" value="0.012147952893380855" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i495]" time="0.081"><properties><property name="score" value="0.0750489319991103" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i496]" time="0.084"><properties><property name="score" value="0.012754462336221526" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i497]" time="0.080"><properties><property name="score" value="0.03486518037567782" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i498]" time="0.015"><properties><property name="score" value="0.028511504056588644" /></properties></testcase><testcase classname="test_lzma_detect" name="test_human_jsonl[i499]" time="0.071"><properties><property name="score" value="0.07740961203534781" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i0]" time="0.014"><properties><property name="score" value="0.05516835155210398" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255001 (text: What is this?  This is a neat little plugin to giv) is an LLM-generated sample, misclassified as human-generated with confidence 0.05516835&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255001 (text: What is this?  This is a neat little plugin to giv) is an LLM-generated sample, misclassified as human-generated with confidence 0.05516835
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i7]" time="0.039"><properties><property name="score" value="0.17161818356084546" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i8]" time="0.072"><properties><property name="score" value="0.19998546535297348" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i9]" time="0.042"><properties><property name="score" value="0.2034660593386861" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i10]" time="0.035"><properties><property name="score" value="0.14110637799338854" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i11]" time="0.070"><properties><property name="score" value="0.05264408277925551" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05264408&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05264408
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i1]" time="0.025"><properties><property name="score" value="0.11440710196061674" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255002 (text: People carry a memorial sign to honor Robert Paxto) is an LLM-generated sample, misclassified as human-generated with confidence 0.1144071&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255002 (text: People carry a memorial sign to honor Robert Paxto) is an LLM-generated sample, misclassified as human-generated with confidence 0.1144071
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i12]" time="0.037"><properties><property name="score" value="0.22440463001321964" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i13]" time="0.035"><properties><property name="score" value="0.26382050776749266" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i14]" time="0.036"><properties><property name="score" value="0.26958045278616116" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i15]" time="0.066"><properties><property name="score" value="0.15965864161698873" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i16]" time="0.078"><properties><property name="score" value="0.24014616285820503" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i17]" time="0.041"><properties><property name="score" value="0.07548340382368401" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i18]" time="0.077"><properties><property name="score" value="0.16956558831926483" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i19]" time="0.068"><properties><property name="score" value="0.02311939253483534" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i20]" time="0.027"><properties><property name="score" value="0.6011619541668733" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.60116195&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.60116195
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i2]" time="0.024"><properties><property name="score" value="0.04042116017044489" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255003 (text: Google is launching a new scholarship campaign for) is an LLM-generated sample, misclassified as human-generated with confidence 0.04042116&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255003 (text: Google is launching a new scholarship campaign for) is an LLM-generated sample, misclassified as human-generated with confidence 0.04042116
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i21]" time="0.071"><properties><property name="score" value="0.16145732202147803" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.16145732&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.16145732
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i3]" time="0.075"><properties><property name="score" value="0.12230287514175717" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255004 (text: Authored by: Kristin Scavuzzo, Managing Director o) is an LLM-generated sample, misclassified as human-generated with confidence 0.12230288&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255004 (text: Authored by: Kristin Scavuzzo, Managing Director o) is an LLM-generated sample, misclassified as human-generated with confidence 0.12230288
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i22]" time="0.043"><properties><property name="score" value="0.0633278566554707" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i23]" time="0.074"><properties><property name="score" value="0.019817938024135895" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i24]" time="0.045"><properties><property name="score" value="0.07755169371190251" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i25]" time="0.044"><properties><property name="score" value="0.31888696545837036" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i26]" time="0.039"><properties><property name="score" value="0.21260855478610896" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.21260855&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.21260855
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i4]" time="0.025"><properties><property name="score" value="0.2503487335472128" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255005 (text: &gt;at some point the MPP team stop standardizing pee) is an LLM-generated sample, misclassified as human-generated with confidence 0.25034873&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255005 (text: &gt;at some point the MPP team stop standardizing pee) is an LLM-generated sample, misclassified as human-generated with confidence 0.25034873
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i27]" time="0.068"><properties><property name="score" value="0.5354631147933183" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.53546311&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.53546311
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i5]" time="0.013"><properties><property name="score" value="0.07520896396797094" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255006 (text: ISIS is breeding from among the refugees forced ou) is an LLM-generated sample, misclassified as human-generated with confidence 0.07520896&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255006 (text: ISIS is breeding from among the refugees forced ou) is an LLM-generated sample, misclassified as human-generated with confidence 0.07520896
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i28]" time="0.037"><properties><property name="score" value="0.10759808161389661" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i29]" time="0.041"><properties><property name="score" value="0.13344409730775553" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i30]" time="0.036"><properties><property name="score" value="0.09549167243527723" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.09549167&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.09549167
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i6]" time="0.069"><properties><property name="score" value="0.12701120649006747" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255007 (text: More than a third of population as of recently, Ma) is an LLM-generated sample, misclassified as human-generated with confidence 0.12701121&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255007 (text: More than a third of population as of recently, Ma) is an LLM-generated sample, misclassified as human-generated with confidence 0.12701121
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i31]" time="0.019"><properties><property name="score" value="0.6601123987920454" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.6601124&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.6601124
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i7]" time="0.029"><properties><property name="score" value="0.11665605937876189" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255008 (text: © Provided by Texas State Historical Association' ) is an LLM-generated sample, misclassified as human-generated with confidence 0.11665606&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255008 (text: © Provided by Texas State Historical Association' ) is an LLM-generated sample, misclassified as human-generated with confidence 0.11665606
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i32]" time="0.065"><properties><property name="score" value="0.1046578692525298" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i33]" time="0.068"><properties><property name="score" value="0.12191993541029363" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i34]" time="0.067"><properties><property name="score" value="0.21243825796725865" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i35]" time="0.041"><properties><property name="score" value="0.10674995223169947" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i36]" time="0.073"><properties><property name="score" value="0.1373583224706191" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i37]" time="0.076"><properties><property name="score" value="0.24261896264978397" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i38]" time="0.077"><properties><property name="score" value="0.22693800520616017" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i39]" time="0.040"><properties><property name="score" value="0.20520459783254785" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i40]" time="0.068"><properties><property name="score" value="0.08878936658004988" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i41]" time="0.039"><properties><property name="score" value="0.1106819975531248" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i42]" time="0.038"><properties><property name="score" value="0.11818998761567445" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i43]" time="0.036"><properties><property name="score" value="0.1866741426345686" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i44]" time="0.076"><properties><property name="score" value="0.18474931295131558" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i45]" time="0.074"><properties><property name="score" value="0.1441712300017476" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i46]" time="0.064"><properties><property name="score" value="0.0349737579182959" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i47]" time="0.067"><properties><property name="score" value="0.062163550633764775" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i48]" time="0.043"><properties><property name="score" value="0.39267657023828173" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i49]" time="0.048"><properties><property name="score" value="0.20705274813229146" /></properties></testcase></testsuite></testsuites>
\ No newline at end of file
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i8]" time="0.071"><properties><property name="score" value="0.10377785561847819" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255009 (text: Mongolia has banned the sale and consumption of me) is an LLM-generated sample, misclassified as human-generated with confidence 0.10377786&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255009 (text: Mongolia has banned the sale and consumption of me) is an LLM-generated sample, misclassified as human-generated with confidence 0.10377786
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i9]" time="0.031"><properties><property name="score" value="0.140969027638907" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255010 (text: Bots and AI are everywhere these days, and it's no) is an LLM-generated sample, misclassified as human-generated with confidence 0.14096903&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255010 (text: Bots and AI are everywhere these days, and it's no) is an LLM-generated sample, misclassified as human-generated with confidence 0.14096903
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i10]" time="0.014"><properties><property name="score" value="0.21900297980154138" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i11]" time="0.028"><properties><property name="score" value="0.11263997401257794" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255012 (text: Dr. Karl Widerquist says his study &quot;opens up an in) is an LLM-generated sample, misclassified as human-generated with confidence 0.11263997&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255012 (text: Dr. Karl Widerquist says his study "opens up an in) is an LLM-generated sample, misclassified as human-generated with confidence 0.11263997
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i12]" time="0.072"><properties><property name="score" value="0.15820551450019998" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255013 (text: 2. Every nature, forest, even in the front pasture) is an LLM-generated sample, misclassified as human-generated with confidence 0.15820551&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255013 (text: 2. Every nature, forest, even in the front pasture) is an LLM-generated sample, misclassified as human-generated with confidence 0.15820551
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i13]" time="0.019"><properties><property name="score" value="0.06723511303219909" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255014 (text: Fact: The Justice 4 Steph showing can be good tele) is an LLM-generated sample, misclassified as human-generated with confidence 0.06723511&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255014 (text: Fact: The Justice 4 Steph showing can be good tele) is an LLM-generated sample, misclassified as human-generated with confidence 0.06723511
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i14]" time="0.027"><properties><property name="score" value="0.06813602841278155" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255015 (text: CALGARY — Tory MP Mac Harb resigned from the Conse) is an LLM-generated sample, misclassified as human-generated with confidence 0.06813603&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255015 (text: CALGARY — Tory MP Mac Harb resigned from the Conse) is an LLM-generated sample, misclassified as human-generated with confidence 0.06813603
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i15]" time="0.078"><properties><property name="score" value="0.13399581048007875" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255016 (text: SERVO, Jupiter - Since the discovery of the swirli) is an LLM-generated sample, misclassified as human-generated with confidence 0.13399581&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255016 (text: SERVO, Jupiter - Since the discovery of the swirli) is an LLM-generated sample, misclassified as human-generated with confidence 0.13399581
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i16]" time="0.015"><properties><property name="score" value="0.06722574414058657" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255017 (text: Five of my eight Ragtime stage hands have Bastards) is an LLM-generated sample, misclassified as human-generated with confidence 0.06722574&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255017 (text: Five of my eight Ragtime stage hands have Bastards) is an LLM-generated sample, misclassified as human-generated with confidence 0.06722574
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i17]" time="0.074"><properties><property name="score" value="0.2540382417549755" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255018 (text: Cipei-Rochel  Installerete Generasseanunterzeichnu) is an LLM-generated sample, misclassified as human-generated with confidence 0.25403824&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255018 (text: Cipei-Rochel  Installerete Generasseanunterzeichnu) is an LLM-generated sample, misclassified as human-generated with confidence 0.25403824
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i18]" time="0.026"><properties><property name="score" value="0.06910393570968804" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255019 (text: PITTSBURGH, Aug. 15, 2013 /PRNewswire/ -- The Nati) is an LLM-generated sample, misclassified as human-generated with confidence 0.06910394&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255019 (text: PITTSBURGH, Aug. 15, 2013 /PRNewswire/ -- The Nati) is an LLM-generated sample, misclassified as human-generated with confidence 0.06910394
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i19]" time="0.079"><properties><property name="score" value="0.1156680044067018" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255020 (text: Albert Einstein, one of the world's most famous sc) is an LLM-generated sample, misclassified as human-generated with confidence 0.115668&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255020 (text: Albert Einstein, one of the world's most famous sc) is an LLM-generated sample, misclassified as human-generated with confidence 0.115668
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i20]" time="0.073"><properties><property name="score" value="0.12358454680546577" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255021 (text: Barney: They weren't really that opposed to me if ) is an LLM-generated sample, misclassified as human-generated with confidence 0.12358455&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255021 (text: Barney: They weren't really that opposed to me if ) is an LLM-generated sample, misclassified as human-generated with confidence 0.12358455
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i21]" time="0.081"><properties><property name="score" value="0.13454514174043863" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i22]" time="0.083"><properties><property name="score" value="0.16334227779697974" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255023 (text: My Flamin' Hot Chocolate Rd2o   ❮❮ Newer Download ) is an LLM-generated sample, misclassified as human-generated with confidence 0.16334228&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255023 (text: My Flamin' Hot Chocolate Rd2o   ❮❮ Newer Download ) is an LLM-generated sample, misclassified as human-generated with confidence 0.16334228
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i23]" time="0.085"><properties><property name="score" value="0.11840157231150242" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255024 (text: I wake up with a fever. This isn't good. No, no, n) is an LLM-generated sample, misclassified as human-generated with confidence 0.11840157&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255024 (text: I wake up with a fever. This isn't good. No, no, n) is an LLM-generated sample, misclassified as human-generated with confidence 0.11840157
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i24]" time="0.020"><properties><property name="score" value="0.0948412562541967" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255025 (text: Convicted cop killer should spend more time in jai) is an LLM-generated sample, misclassified as human-generated with confidence 0.09484126&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255025 (text: Convicted cop killer should spend more time in jai) is an LLM-generated sample, misclassified as human-generated with confidence 0.09484126
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i25]" time="0.104"><properties><property name="score" value="0.13603836559253796" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255026 (text: Issue Width: 500px  Posted on: 01/25/2013  Twitter) is an LLM-generated sample, misclassified as human-generated with confidence 0.13603837&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255026 (text: Issue Width: 500px  Posted on: 01/25/2013  Twitter) is an LLM-generated sample, misclassified as human-generated with confidence 0.13603837
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i26]" time="0.016"><properties><property name="score" value="0.05915011422563676" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255027 (text: Unique if not unique!  I bought this rug for my tw) is an LLM-generated sample, misclassified as human-generated with confidence 0.05915011&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255027 (text: Unique if not unique!  I bought this rug for my tw) is an LLM-generated sample, misclassified as human-generated with confidence 0.05915011
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i27]" time="0.069"><properties><property name="score" value="0.12942657758489046" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255028 (text: I know this is a book about Holocaust survivors, b) is an LLM-generated sample, misclassified as human-generated with confidence 0.12942658&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255028 (text: I know this is a book about Holocaust survivors, b) is an LLM-generated sample, misclassified as human-generated with confidence 0.12942658
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i28]" time="0.015"><properties><property name="score" value="0.07833496717293276" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255029 (text: We are pleased to announce that our previously ann) is an LLM-generated sample, misclassified as human-generated with confidence 0.07833497&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255029 (text: We are pleased to announce that our previously ann) is an LLM-generated sample, misclassified as human-generated with confidence 0.07833497
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i29]" time="0.014"><properties><property name="score" value="0.21900764424140906" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255030 (text: Stitch Fuse Box by Kink Novaphor Supporting Kinky ) is an LLM-generated sample, misclassified as human-generated with confidence 0.21900764&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255030 (text: Stitch Fuse Box by Kink Novaphor Supporting Kinky ) is an LLM-generated sample, misclassified as human-generated with confidence 0.21900764
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i30]" time="0.094"><properties><property name="score" value="0.1573755246219531" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255031 (text: A small group of Soviet &quot;Oh-On's&quot; fled North Vietn) is an LLM-generated sample, misclassified as human-generated with confidence 0.15737552&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255031 (text: A small group of Soviet "Oh-On's" fled North Vietn) is an LLM-generated sample, misclassified as human-generated with confidence 0.15737552
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i31]" time="0.016"><properties><property name="score" value="0.022246781545676164" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255032 (text: Some of you dislike Jetpacks. You don't like them,) is an LLM-generated sample, misclassified as human-generated with confidence 0.02224678&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255032 (text: Some of you dislike Jetpacks. You don't like them,) is an LLM-generated sample, misclassified as human-generated with confidence 0.02224678
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i32]" time="0.024"><properties><property name="score" value="0.09558319761657955" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255033 (text: PALLBRIDGE, Mass. -- The New Hampshire Department ) is an LLM-generated sample, misclassified as human-generated with confidence 0.0955832&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255033 (text: PALLBRIDGE, Mass. -- The New Hampshire Department ) is an LLM-generated sample, misclassified as human-generated with confidence 0.0955832
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i33]" time="0.014"><properties><property name="score" value="0.10673118501483003" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255034 (text: Education K-12  World Affairs Middle East &amp; North ) is an LLM-generated sample, misclassified as human-generated with confidence 0.10673119&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255034 (text: Education K-12  World Affairs Middle East &amp; North ) is an LLM-generated sample, misclassified as human-generated with confidence 0.10673119
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i34]" time="0.024"><properties><property name="score" value="0.10711397560808636" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255035 (text: Speaker Paul Ryan Paul Davis RyanHouse passes reso) is an LLM-generated sample, misclassified as human-generated with confidence 0.10711398&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255035 (text: Speaker Paul Ryan Paul Davis RyanHouse passes reso) is an LLM-generated sample, misclassified as human-generated with confidence 0.10711398
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i35]" time="0.066"><properties><property name="score" value="0.15778005219379712" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255036 (text: To revisit an idea that a writer once put forth in) is an LLM-generated sample, misclassified as human-generated with confidence 0.15778005&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255036 (text: To revisit an idea that a writer once put forth in) is an LLM-generated sample, misclassified as human-generated with confidence 0.15778005
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i36]" time="0.081"><properties><property name="score" value="0.10044049894541582" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255037 (text: Jimmy Pastore worked his way up from the lightweig) is an LLM-generated sample, misclassified as human-generated with confidence 0.1004405&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255037 (text: Jimmy Pastore worked his way up from the lightweig) is an LLM-generated sample, misclassified as human-generated with confidence 0.1004405
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i37]" time="0.021"><properties><property name="score" value="0.04889361188814001" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255038 (text: Copyright by WAVY - All rights reserved Associated) is an LLM-generated sample, misclassified as human-generated with confidence 0.04889361&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255038 (text: Copyright by WAVY - All rights reserved Associated) is an LLM-generated sample, misclassified as human-generated with confidence 0.04889361
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i38]" time="0.014"><properties><property name="score" value="0.044344203617108224" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255039 (text: The Republican-led legislature will not appoint a ) is an LLM-generated sample, misclassified as human-generated with confidence 0.0443442&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255039 (text: The Republican-led legislature will not appoint a ) is an LLM-generated sample, misclassified as human-generated with confidence 0.0443442
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i39]" time="0.077"><properties><property name="score" value="0.14337769271843812" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255040 (text: http://tvtropes.org/pmwiki/pmwiki.php/Main/Dunes  ) is an LLM-generated sample, misclassified as human-generated with confidence 0.14337769&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255040 (text: http://tvtropes.org/pmwiki/pmwiki.php/Main/Dunes  ) is an LLM-generated sample, misclassified as human-generated with confidence 0.14337769
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i40]" time="0.017"><properties><property name="score" value="0.24122317817462102" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255041 (text: You have a choice when your four year old gets to ) is an LLM-generated sample, misclassified as human-generated with confidence 0.24122318&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255041 (text: You have a choice when your four year old gets to ) is an LLM-generated sample, misclassified as human-generated with confidence 0.24122318
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i41]" time="0.070"><properties><property name="score" value="0.13271103449938515" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255042 (text: So why the disregard for danger? Contrary to popul) is an LLM-generated sample, misclassified as human-generated with confidence 0.13271103&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255042 (text: So why the disregard for danger? Contrary to popul) is an LLM-generated sample, misclassified as human-generated with confidence 0.13271103
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i42]" time="0.018"><properties><property name="score" value="0.08576293028559889" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255043 (text: review by Carter118 on 6/26/2017  Recommend  This ) is an LLM-generated sample, misclassified as human-generated with confidence 0.08576293&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255043 (text: review by Carter118 on 6/26/2017  Recommend  This ) is an LLM-generated sample, misclassified as human-generated with confidence 0.08576293
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i43]" time="0.071"><properties><property name="score" value="0.09101269373035992" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255044 (text: AryanTheFascist writes...  I am sorry to hear it M) is an LLM-generated sample, misclassified as human-generated with confidence 0.09101269&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255044 (text: AryanTheFascist writes...  I am sorry to hear it M) is an LLM-generated sample, misclassified as human-generated with confidence 0.09101269
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i44]" time="0.074"><properties><property name="score" value="0.07795877933026037" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255045 (text: Fear that mirrors a changing political dynamic in ) is an LLM-generated sample, misclassified as human-generated with confidence 0.07795878&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255045 (text: Fear that mirrors a changing political dynamic in ) is an LLM-generated sample, misclassified as human-generated with confidence 0.07795878
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i45]" time="0.031"><properties><property name="score" value="0.14440228581232709" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255046 (text: By Vigilant Citizen Staff   Just a restaurant that) is an LLM-generated sample, misclassified as human-generated with confidence 0.14440229&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255046 (text: By Vigilant Citizen Staff   Just a restaurant that) is an LLM-generated sample, misclassified as human-generated with confidence 0.14440229
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i46]" time="0.014"><properties><property name="score" value="0.1413942604870766" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255047 (text: 16 #geekyvimcovers submitted after watching this v) is an LLM-generated sample, misclassified as human-generated with confidence 0.14139426&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255047 (text: 16 #geekyvimcovers submitted after watching this v) is an LLM-generated sample, misclassified as human-generated with confidence 0.14139426
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i47]" time="0.068"><properties><property name="score" value="0.185664338521456" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255048 (text: Watch a video recording from the Oct. 29 La Plata ) is an LLM-generated sample, misclassified as human-generated with confidence 0.18566434&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255048 (text: Watch a video recording from the Oct. 29 La Plata ) is an LLM-generated sample, misclassified as human-generated with confidence 0.18566434
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i48]" time="0.030"><properties><property name="score" value="0.09659990140686393" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255049 (text: Watch Rick Santorum's full interview with Glenn Be) is an LLM-generated sample, misclassified as human-generated with confidence 0.0965999&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255049 (text: Watch Rick Santorum's full interview with Glenn Be) is an LLM-generated sample, misclassified as human-generated with confidence 0.0965999
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i49]" time="0.025"><properties><property name="score" value="0.14434082135513082" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255050 (text: Passengers and regulators round up illegal drugs d) is an LLM-generated sample, misclassified as human-generated with confidence 0.14434082&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255050 (text: Passengers and regulators round up illegal drugs d) is an LLM-generated sample, misclassified as human-generated with confidence 0.14434082
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i50]" time="0.013"><properties><property name="score" value="0.06216642510185055" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255051 (text: Former Texas governor Rick Perry will sit down wit) is an LLM-generated sample, misclassified as human-generated with confidence 0.06216643&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255051 (text: Former Texas governor Rick Perry will sit down wit) is an LLM-generated sample, misclassified as human-generated with confidence 0.06216643
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i51]" time="0.073"><properties><property name="score" value="0.17077245111569328" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255052 (text: Beloved writer and poet J.R.R. Tolkien once sugges) is an LLM-generated sample, misclassified as human-generated with confidence 0.17077245&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255052 (text: Beloved writer and poet J.R.R. Tolkien once sugges) is an LLM-generated sample, misclassified as human-generated with confidence 0.17077245
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i52]" time="0.084"><properties><property name="score" value="0.12080523491510084" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255053 (text: 2000 Year Oro Jackson  The tomb was previously ide) is an LLM-generated sample, misclassified as human-generated with confidence 0.12080523&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255053 (text: 2000 Year Oro Jackson  The tomb was previously ide) is an LLM-generated sample, misclassified as human-generated with confidence 0.12080523
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i53]" time="0.020"><properties><property name="score" value="0.07666166995952683" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255054 (text: Terror alerts in 13 eastern states remained strong) is an LLM-generated sample, misclassified as human-generated with confidence 0.07666167&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255054 (text: Terror alerts in 13 eastern states remained strong) is an LLM-generated sample, misclassified as human-generated with confidence 0.07666167
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i54]" time="0.029"><properties><property name="score" value="0.07280889258184857" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255055 (text: Despite offering three levels of adaptive service,) is an LLM-generated sample, misclassified as human-generated with confidence 0.07280889&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255055 (text: Despite offering three levels of adaptive service,) is an LLM-generated sample, misclassified as human-generated with confidence 0.07280889
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i55]" time="0.027"><properties><property name="score" value="0.0711540716207647" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255056 (text: The third and final African adventure. The first t) is an LLM-generated sample, misclassified as human-generated with confidence 0.07115407&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255056 (text: The third and final African adventure. The first t) is an LLM-generated sample, misclassified as human-generated with confidence 0.07115407
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i56]" time="0.027"><properties><property name="score" value="0.0577520659976788" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255057 (text: Danforth East is a unique community in the city of) is an LLM-generated sample, misclassified as human-generated with confidence 0.05775207&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255057 (text: Danforth East is a unique community in the city of) is an LLM-generated sample, misclassified as human-generated with confidence 0.05775207
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i57]" time="0.013"><properties><property name="score" value="0.026005825223329726" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255058 (text: layout_container:Genotion 2 - HD select  Focus: Ba) is an LLM-generated sample, misclassified as human-generated with confidence 0.02600583&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255058 (text: layout_container:Genotion 2 - HD select  Focus: Ba) is an LLM-generated sample, misclassified as human-generated with confidence 0.02600583
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i58]" time="0.012"><properties><property name="score" value="0.20567052300982103" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255059 (text: Australia's favorite anime twins, SEGA and Bandai ) is an LLM-generated sample, misclassified as human-generated with confidence 0.20567052&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255059 (text: Australia's favorite anime twins, SEGA and Bandai ) is an LLM-generated sample, misclassified as human-generated with confidence 0.20567052
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i59]" time="0.064"><properties><property name="score" value="0.08903462903595662" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255060 (text:  All of these are in opposition to the redistribut) is an LLM-generated sample, misclassified as human-generated with confidence 0.08903463&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255060 (text:  All of these are in opposition to the redistribut) is an LLM-generated sample, misclassified as human-generated with confidence 0.08903463
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i60]" time="0.013"><properties><property name="score" value="0.055127379722219594" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255061 (text: In a statement to The Daily Caller, Erickson denie) is an LLM-generated sample, misclassified as human-generated with confidence 0.05512738&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255061 (text: In a statement to The Daily Caller, Erickson denie) is an LLM-generated sample, misclassified as human-generated with confidence 0.05512738
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i61]" time="0.023"><properties><property name="score" value="0.09414768307945046" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255062 (text: Saturday, 13 February 2015   (CNW Group/Black Wit ) is an LLM-generated sample, misclassified as human-generated with confidence 0.09414768&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255062 (text: Saturday, 13 February 2015   (CNW Group/Black Wit ) is an LLM-generated sample, misclassified as human-generated with confidence 0.09414768
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i62]" time="0.012"><properties><property name="score" value="0.07831859926079043" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255063 (text: Sorting, Pricing, and Reviews  iOS UK: who product) is an LLM-generated sample, misclassified as human-generated with confidence 0.0783186&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255063 (text: Sorting, Pricing, and Reviews  iOS UK: who product) is an LLM-generated sample, misclassified as human-generated with confidence 0.0783186
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i63]" time="0.013"><properties><property name="score" value="0.08770000150386537" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255064 (text: This revelation was even more startling than the E) is an LLM-generated sample, misclassified as human-generated with confidence 0.0877&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255064 (text: This revelation was even more startling than the E) is an LLM-generated sample, misclassified as human-generated with confidence 0.0877
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i64]" time="0.064"><properties><property name="score" value="0.16995444322594613" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255065 (text: Friday, March 20, 2013  Note: This week's article ) is an LLM-generated sample, misclassified as human-generated with confidence 0.16995444&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255065 (text: Friday, March 20, 2013  Note: This week's article ) is an LLM-generated sample, misclassified as human-generated with confidence 0.16995444
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i65]" time="0.064"><properties><property name="score" value="0.14355240345391385" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255066 (text: Arthur Milan Victor was an urologist from Bala and) is an LLM-generated sample, misclassified as human-generated with confidence 0.1435524&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255066 (text: Arthur Milan Victor was an urologist from Bala and) is an LLM-generated sample, misclassified as human-generated with confidence 0.1435524
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i66]" time="0.067"><properties><property name="score" value="0.07759465924344172" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255067 (text: A few years ago, I was working 16 hours a day, sev) is an LLM-generated sample, misclassified as human-generated with confidence 0.07759466&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255067 (text: A few years ago, I was working 16 hours a day, sev) is an LLM-generated sample, misclassified as human-generated with confidence 0.07759466
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i67]" time="0.077"><properties><property name="score" value="0.09799511948530268" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255068 (text: Generic and other online electric curbside meter s) is an LLM-generated sample, misclassified as human-generated with confidence 0.09799512&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255068 (text: Generic and other online electric curbside meter s) is an LLM-generated sample, misclassified as human-generated with confidence 0.09799512
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i68]" time="0.089"><properties><property name="score" value="0.09854003484935904" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255069 (text: The Legend of Heroes: Trails in the Sky SC is a ga) is an LLM-generated sample, misclassified as human-generated with confidence 0.09854003&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255069 (text: The Legend of Heroes: Trails in the Sky SC is a ga) is an LLM-generated sample, misclassified as human-generated with confidence 0.09854003
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i69]" time="0.023"><properties><property name="score" value="0.12122949263302929" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255070 (text: Product Description  What would the FBI do if it w) is an LLM-generated sample, misclassified as human-generated with confidence 0.12122949&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255070 (text: Product Description  What would the FBI do if it w) is an LLM-generated sample, misclassified as human-generated with confidence 0.12122949
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i70]" time="0.065"><properties><property name="score" value="0.07760204882823982" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255071 (text: Caught on videotape: New DOJ agents claim they wip) is an LLM-generated sample, misclassified as human-generated with confidence 0.07760205&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255071 (text: Caught on videotape: New DOJ agents claim they wip) is an LLM-generated sample, misclassified as human-generated with confidence 0.07760205
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i71]" time="0.086"><properties><property name="score" value="0.09013889711173581" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255072 (text: Russia is banning imports of chicken and pork from) is an LLM-generated sample, misclassified as human-generated with confidence 0.0901389&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255072 (text: Russia is banning imports of chicken and pork from) is an LLM-generated sample, misclassified as human-generated with confidence 0.0901389
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i72]" time="0.093"><properties><property name="score" value="0.10473514245912947" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i73]" time="0.033"><properties><property name="score" value="0.15567028375834158" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255074 (text: 'Ball' Bin:  (Inside the History of CWF 'Ball') Yo) is an LLM-generated sample, misclassified as human-generated with confidence 0.15567028&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255074 (text: 'Ball' Bin:  (Inside the History of CWF 'Ball') Yo) is an LLM-generated sample, misclassified as human-generated with confidence 0.15567028
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i74]" time="0.026"><properties><property name="score" value="0.12188488063344641" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255075 (text: Mission Summary  While collaborating on the Androi) is an LLM-generated sample, misclassified as human-generated with confidence 0.12188488&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255075 (text: Mission Summary  While collaborating on the Androi) is an LLM-generated sample, misclassified as human-generated with confidence 0.12188488
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i75]" time="0.026"><properties><property name="score" value="0.03696975580113587" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255076 (text: Steve touches a computer and is immediately greete) is an LLM-generated sample, misclassified as human-generated with confidence 0.03696976&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255076 (text: Steve touches a computer and is immediately greete) is an LLM-generated sample, misclassified as human-generated with confidence 0.03696976
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i76]" time="0.013"><properties><property name="score" value="0.06933030045243593" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255077 (text: Mall Santmorn; a study of the art district in Bang) is an LLM-generated sample, misclassified as human-generated with confidence 0.0693303&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255077 (text: Mall Santmorn; a study of the art district in Bang) is an LLM-generated sample, misclassified as human-generated with confidence 0.0693303
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i77]" time="0.013"><properties><property name="score" value="0.06947273045128655" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255078 (text: Hurst Productions' love of Ken Burns documentary s) is an LLM-generated sample, misclassified as human-generated with confidence 0.06947273&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255078 (text: Hurst Productions' love of Ken Burns documentary s) is an LLM-generated sample, misclassified as human-generated with confidence 0.06947273
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i78]" time="0.013"><properties><property name="score" value="0.052260736842629774" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255079 (text: https://www.mirror.co.uk › News › World  Jun 22, 2) is an LLM-generated sample, misclassified as human-generated with confidence 0.05226074&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255079 (text: https://www.mirror.co.uk › News › World  Jun 22, 2) is an LLM-generated sample, misclassified as human-generated with confidence 0.05226074
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i79]" time="0.014"><properties><property name="score" value="0.05368344790945834" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255080 (text: I am new to Linux amd64 development.  Dell Inspiro) is an LLM-generated sample, misclassified as human-generated with confidence 0.05368345&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255080 (text: I am new to Linux amd64 development.  Dell Inspiro) is an LLM-generated sample, misclassified as human-generated with confidence 0.05368345
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i80]" time="0.025"><properties><property name="score" value="0.062246203506382616" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255081 (text: The Guo's had been waiting for about 7 hours in th) is an LLM-generated sample, misclassified as human-generated with confidence 0.0622462&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255081 (text: The Guo's had been waiting for about 7 hours in th) is an LLM-generated sample, misclassified as human-generated with confidence 0.0622462
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i81]" time="0.023"><properties><property name="score" value="0.1306811779152299" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255082 (text: Method 1: Get the Pro graphic card rollover report) is an LLM-generated sample, misclassified as human-generated with confidence 0.13068118&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255082 (text: Method 1: Get the Pro graphic card rollover report) is an LLM-generated sample, misclassified as human-generated with confidence 0.13068118
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i82]" time="0.060"><properties><property name="score" value="0.1432713127146124" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255083 (text: The Conservation Volunteer Volunteers are Food Pan) is an LLM-generated sample, misclassified as human-generated with confidence 0.14327131&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255083 (text: The Conservation Volunteer Volunteers are Food Pan) is an LLM-generated sample, misclassified as human-generated with confidence 0.14327131
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i83]" time="0.024"><properties><property name="score" value="0.11744124257876054" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255084 (text: A UK recording artist and celebrity chef has been ) is an LLM-generated sample, misclassified as human-generated with confidence 0.11744124&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255084 (text: A UK recording artist and celebrity chef has been ) is an LLM-generated sample, misclassified as human-generated with confidence 0.11744124
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i84]" time="0.013"><properties><property name="score" value="0.059365808818195864" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255085 (text: Let's go national! (CBC) The country's first-ever ) is an LLM-generated sample, misclassified as human-generated with confidence 0.05936581&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255085 (text: Let's go national! (CBC) The country's first-ever ) is an LLM-generated sample, misclassified as human-generated with confidence 0.05936581
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i85]" time="0.064"><properties><property name="score" value="0.2440808930468855" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255086 (text: Like the others, I started with a Retina MacBook P) is an LLM-generated sample, misclassified as human-generated with confidence 0.24408089&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255086 (text: Like the others, I started with a Retina MacBook P) is an LLM-generated sample, misclassified as human-generated with confidence 0.24408089
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i86]" time="0.062"><properties><property name="score" value="0.19029696664434242" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255087 (text: %static name=&quot;order_is_quota_indicator&quot; expr=&quot;Orde) is an LLM-generated sample, misclassified as human-generated with confidence 0.19029697&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255087 (text: %static name="order_is_quota_indicator" expr="Orde) is an LLM-generated sample, misclassified as human-generated with confidence 0.19029697
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i87]" time="0.067"><properties><property name="score" value="0.11867464985005197" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255088 (text: Share. From games to Funko and MORE. From games to) is an LLM-generated sample, misclassified as human-generated with confidence 0.11867465&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255088 (text: Share. From games to Funko and MORE. From games to) is an LLM-generated sample, misclassified as human-generated with confidence 0.11867465
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i88]" time="0.031"><properties><property name="score" value="0.12037884386365919" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255089 (text: The music video for Nirvana's &quot;Smells Like Teen Sp) is an LLM-generated sample, misclassified as human-generated with confidence 0.12037884&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255089 (text: The music video for Nirvana's "Smells Like Teen Sp) is an LLM-generated sample, misclassified as human-generated with confidence 0.12037884
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i89]" time="0.076"><properties><property name="score" value="0.12805071301460572" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255090 (text: On any given Sunday, scientists go to the museum, ) is an LLM-generated sample, misclassified as human-generated with confidence 0.12805071&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255090 (text: On any given Sunday, scientists go to the museum, ) is an LLM-generated sample, misclassified as human-generated with confidence 0.12805071
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i90]" time="0.014"><properties><property name="score" value="0.01983028278876864" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255091 (text: At the very least you can target everything on the) is an LLM-generated sample, misclassified as human-generated with confidence 0.01983028&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255091 (text: At the very least you can target everything on the) is an LLM-generated sample, misclassified as human-generated with confidence 0.01983028
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i91]" time="0.062"><properties><property name="score" value="0.13160398635913712" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255092 (text: [image-36]  NASA's Mars rover Curiosity has taken ) is an LLM-generated sample, misclassified as human-generated with confidence 0.13160399&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255092 (text: [image-36]  NASA's Mars rover Curiosity has taken ) is an LLM-generated sample, misclassified as human-generated with confidence 0.13160399
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i92]" time="0.067"><properties><property name="score" value="0.16230394872043077" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255093 (text: Infowars.com  December 16, 2015  John McCain react) is an LLM-generated sample, misclassified as human-generated with confidence 0.16230395&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255093 (text: Infowars.com  December 16, 2015  John McCain react) is an LLM-generated sample, misclassified as human-generated with confidence 0.16230395
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i93]" time="0.013"><properties><property name="score" value="0.012528406386652646" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i94]" time="0.026"><properties><property name="score" value="0.12961933267391554" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255095 (text: ROME (Reuters) - Italian Prime Minister Matteo Ren) is an LLM-generated sample, misclassified as human-generated with confidence 0.12961933&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255095 (text: ROME (Reuters) - Italian Prime Minister Matteo Ren) is an LLM-generated sample, misclassified as human-generated with confidence 0.12961933
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i95]" time="0.018"><properties><property name="score" value="0.036662779878998686" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255096 (text: Finally, the most basic systemd unit file for Cano) is an LLM-generated sample, misclassified as human-generated with confidence 0.03666278&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255096 (text: Finally, the most basic systemd unit file for Cano) is an LLM-generated sample, misclassified as human-generated with confidence 0.03666278
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i96]" time="0.071"><properties><property name="score" value="0.1286339378450703" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255097 (text: 4 Part Liver, 3 Part Pancreas Autauga Indians, Uni) is an LLM-generated sample, misclassified as human-generated with confidence 0.12863394&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255097 (text: 4 Part Liver, 3 Part Pancreas Autauga Indians, Uni) is an LLM-generated sample, misclassified as human-generated with confidence 0.12863394
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i97]" time="0.078"><properties><property name="score" value="0.09150872243669941" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255099 (text: Another local UFO sighting that only a couple of w) is an LLM-generated sample, misclassified as human-generated with confidence 0.09150872&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255099 (text: Another local UFO sighting that only a couple of w) is an LLM-generated sample, misclassified as human-generated with confidence 0.09150872
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i98]" time="0.079"><properties><property name="score" value="0.1379870222532861" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255100 (text: Recovery initiated after Opwell has been confirmed) is an LLM-generated sample, misclassified as human-generated with confidence 0.13798702&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255100 (text: Recovery initiated after Opwell has been confirmed) is an LLM-generated sample, misclassified as human-generated with confidence 0.13798702
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i99]" time="0.088"><properties><property name="score" value="0.06750404222007281" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255101 (text: Burbank, CA  Duration of studio: 9 - 14 hours.  Co) is an LLM-generated sample, misclassified as human-generated with confidence 0.06750404&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255101 (text: Burbank, CA  Duration of studio: 9 - 14 hours.  Co) is an LLM-generated sample, misclassified as human-generated with confidence 0.06750404
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i100]" time="0.097"><properties><property name="score" value="0.15671370526031567" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255103 (text: Purchasing options / latest deals  Contact in Sego) is an LLM-generated sample, misclassified as human-generated with confidence 0.15671371&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255103 (text: Purchasing options / latest deals  Contact in Sego) is an LLM-generated sample, misclassified as human-generated with confidence 0.15671371
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i101]" time="0.023"><properties><property name="score" value="0.06855287913709596" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255104 (text: adj.  &quot;deadly, dreadful, utterly abject,&quot; 1775, fr) is an LLM-generated sample, misclassified as human-generated with confidence 0.06855288&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255104 (text: adj.  "deadly, dreadful, utterly abject," 1775, fr) is an LLM-generated sample, misclassified as human-generated with confidence 0.06855288
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i102]" time="0.017"><properties><property name="score" value="0.04034986040563071" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255105 (text: ES5 iterations can be omitted. PDOException and Bl) is an LLM-generated sample, misclassified as human-generated with confidence 0.04034986&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255105 (text: ES5 iterations can be omitted. PDOException and Bl) is an LLM-generated sample, misclassified as human-generated with confidence 0.04034986
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i103]" time="0.026"><properties><property name="score" value="0.06741430017771549" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255106 (text: The entrance to the Werribee apartment block in we) is an LLM-generated sample, misclassified as human-generated with confidence 0.0674143&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255106 (text: The entrance to the Werribee apartment block in we) is an LLM-generated sample, misclassified as human-generated with confidence 0.0674143
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i104]" time="0.063"><properties><property name="score" value="0.08615812476426266" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255107 (text: Acclaimed biomedical researcher Michael Stanton, P) is an LLM-generated sample, misclassified as human-generated with confidence 0.08615812&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255107 (text: Acclaimed biomedical researcher Michael Stanton, P) is an LLM-generated sample, misclassified as human-generated with confidence 0.08615812
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i105]" time="0.064"><properties><property name="score" value="0.08422312758029435" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255108 (text: The principle of plagiarism dates back to Roman an) is an LLM-generated sample, misclassified as human-generated with confidence 0.08422313&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255108 (text: The principle of plagiarism dates back to Roman an) is an LLM-generated sample, misclassified as human-generated with confidence 0.08422313
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i106]" time="0.025"><properties><property name="score" value="0.0923382039195073" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255109 (text: Baby will make $740 per week during summer holiday) is an LLM-generated sample, misclassified as human-generated with confidence 0.0923382&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255109 (text: Baby will make $740 per week during summer holiday) is an LLM-generated sample, misclassified as human-generated with confidence 0.0923382
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i107]" time="0.026"><properties><property name="score" value="0.13547306657162922" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255110 (text: That 1993 election always weighs on me. For the fi) is an LLM-generated sample, misclassified as human-generated with confidence 0.13547307&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255110 (text: That 1993 election always weighs on me. For the fi) is an LLM-generated sample, misclassified as human-generated with confidence 0.13547307
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i108]" time="0.024"><properties><property name="score" value="0.15065251104721566" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255111 (text: Everton transfer target Victor Wanyama ruffled Che) is an LLM-generated sample, misclassified as human-generated with confidence 0.15065251&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255111 (text: Everton transfer target Victor Wanyama ruffled Che) is an LLM-generated sample, misclassified as human-generated with confidence 0.15065251
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i109]" time="0.012"><properties><property name="score" value="0.07537445773441176" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255112 (text: ULTRAMETALLA is known for it's unique blend of tra) is an LLM-generated sample, misclassified as human-generated with confidence 0.07537446&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255112 (text: ULTRAMETALLA is known for it's unique blend of tra) is an LLM-generated sample, misclassified as human-generated with confidence 0.07537446
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i110]" time="0.064"><properties><property name="score" value="0.21821255610915954" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255113 (text: The forgiveness Glaswegians are capriciously expro) is an LLM-generated sample, misclassified as human-generated with confidence 0.21821256&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255113 (text: The forgiveness Glaswegians are capriciously expro) is an LLM-generated sample, misclassified as human-generated with confidence 0.21821256
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i111]" time="0.023"><properties><property name="score" value="0.04183297701745815" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255114 (text: Page Content  Democrats share common interests ran) is an LLM-generated sample, misclassified as human-generated with confidence 0.04183298&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255114 (text: Page Content  Democrats share common interests ran) is an LLM-generated sample, misclassified as human-generated with confidence 0.04183298
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i112]" time="0.023"><properties><property name="score" value="0.1325699559790796" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255115 (text: Kareem Abdul-Jabbar was one of the most lauded mem) is an LLM-generated sample, misclassified as human-generated with confidence 0.13256996&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255115 (text: Kareem Abdul-Jabbar was one of the most lauded mem) is an LLM-generated sample, misclassified as human-generated with confidence 0.13256996
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i113]" time="0.063"><properties><property name="score" value="0.18426305510748456" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255116 (text: Words and photos by Sonia Heerveld Photography  (D) is an LLM-generated sample, misclassified as human-generated with confidence 0.18426306&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255116 (text: Words and photos by Sonia Heerveld Photography  (D) is an LLM-generated sample, misclassified as human-generated with confidence 0.18426306
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i114]" time="0.026"><properties><property name="score" value="0.07930856816478495" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255117 (text: There's been much talk and speculation about wheth) is an LLM-generated sample, misclassified as human-generated with confidence 0.07930857&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255117 (text: There's been much talk and speculation about wheth) is an LLM-generated sample, misclassified as human-generated with confidence 0.07930857
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i115]" time="0.065"><properties><property name="score" value="0.10343717392318881" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255118 (text: When things go horribly wrong in an emergency, the) is an LLM-generated sample, misclassified as human-generated with confidence 0.10343717&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255118 (text: When things go horribly wrong in an emergency, the) is an LLM-generated sample, misclassified as human-generated with confidence 0.10343717
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i116]" time="0.024"><properties><property name="score" value="0.10104427177429433" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255119 (text: Pep Guardiola has hinted that the Manchester City ) is an LLM-generated sample, misclassified as human-generated with confidence 0.10104427&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255119 (text: Pep Guardiola has hinted that the Manchester City ) is an LLM-generated sample, misclassified as human-generated with confidence 0.10104427
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i117]" time="0.013"><properties><property name="score" value="0.09901593202084058" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255120 (text: To be slugged on deductibles of over $6,500 annual) is an LLM-generated sample, misclassified as human-generated with confidence 0.09901593&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255120 (text: To be slugged on deductibles of over $6,500 annual) is an LLM-generated sample, misclassified as human-generated with confidence 0.09901593
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i118]" time="0.012"><properties><property name="score" value="0.39154126505792086" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i119]" time="0.066"><properties><property name="score" value="0.11408234495299702" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255122 (text: Wilson's comments about gay relationships also inj) is an LLM-generated sample, misclassified as human-generated with confidence 0.11408234&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255122 (text: Wilson's comments about gay relationships also inj) is an LLM-generated sample, misclassified as human-generated with confidence 0.11408234
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i120]" time="0.070"><properties><property name="score" value="0.1049855633352291" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255123 (text: Turing Pharmaceuticals has agreed in principle to ) is an LLM-generated sample, misclassified as human-generated with confidence 0.10498556&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255123 (text: Turing Pharmaceuticals has agreed in principle to ) is an LLM-generated sample, misclassified as human-generated with confidence 0.10498556
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i121]" time="0.013"><properties><property name="score" value="0.03836829067943026" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255124 (text: I'm not out to bash the community or call attentio) is an LLM-generated sample, misclassified as human-generated with confidence 0.03836829&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255124 (text: I'm not out to bash the community or call attentio) is an LLM-generated sample, misclassified as human-generated with confidence 0.03836829
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i122]" time="0.013"><properties><property name="score" value="0.05372224103825096" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255125 (text: Jackson has advised Clinton for years, and aides s) is an LLM-generated sample, misclassified as human-generated with confidence 0.05372224&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255125 (text: Jackson has advised Clinton for years, and aides s) is an LLM-generated sample, misclassified as human-generated with confidence 0.05372224
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i123]" time="0.017"><properties><property name="score" value="0.03601975088104448" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255126 (text: Instant  Dwarven Blast deals 3 damage to target cr) is an LLM-generated sample, misclassified as human-generated with confidence 0.03601975&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255126 (text: Instant  Dwarven Blast deals 3 damage to target cr) is an LLM-generated sample, misclassified as human-generated with confidence 0.03601975
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i124]" time="0.068"><properties><property name="score" value="0.14803774673316125" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255127 (text: SAN FRANCISCO — Opportunities to test the effectiv) is an LLM-generated sample, misclassified as human-generated with confidence 0.14803775&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255127 (text: SAN FRANCISCO — Opportunities to test the effectiv) is an LLM-generated sample, misclassified as human-generated with confidence 0.14803775
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i125]" time="0.014"><properties><property name="score" value="0.1194913871642056" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255128 (text: Convocainftvalerie 20  13.9'' L Y  1107g  With Scr) is an LLM-generated sample, misclassified as human-generated with confidence 0.11949139&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255128 (text: Convocainftvalerie 20  13.9'' L Y  1107g  With Scr) is an LLM-generated sample, misclassified as human-generated with confidence 0.11949139
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i126]" time="0.076"><properties><property name="score" value="0.1297131208344253" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255129 (text: &quot;I always say to them: 'Pray for our nation and fa) is an LLM-generated sample, misclassified as human-generated with confidence 0.12971312&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255129 (text: "I always say to them: 'Pray for our nation and fa) is an LLM-generated sample, misclassified as human-generated with confidence 0.12971312
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i127]" time="0.029"><properties><property name="score" value="0.04866499505187401" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255130 (text: CBIS Finance equipped farmers in Mangodong to impl) is an LLM-generated sample, misclassified as human-generated with confidence 0.048665&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255130 (text: CBIS Finance equipped farmers in Mangodong to impl) is an LLM-generated sample, misclassified as human-generated with confidence 0.048665
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i128]" time="0.070"><properties><property name="score" value="0.11657564248574889" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255131 (text: A combination logo for DuPont, Inc. is seen in thi) is an LLM-generated sample, misclassified as human-generated with confidence 0.11657564&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255131 (text: A combination logo for DuPont, Inc. is seen in thi) is an LLM-generated sample, misclassified as human-generated with confidence 0.11657564
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i129]" time="0.072"><properties><property name="score" value="0.13694379766057868" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255132 (text: Harry Kane has reminded us of the fantastically da) is an LLM-generated sample, misclassified as human-generated with confidence 0.1369438&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255132 (text: Harry Kane has reminded us of the fantastically da) is an LLM-generated sample, misclassified as human-generated with confidence 0.1369438
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i130]" time="0.015"><properties><property name="score" value="0.022537020397778162" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i131]" time="0.029"><properties><property name="score" value="0.07662390138967763" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255134 (text: Thousands of women visited a new San Francisco cen) is an LLM-generated sample, misclassified as human-generated with confidence 0.0766239&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255134 (text: Thousands of women visited a new San Francisco cen) is an LLM-generated sample, misclassified as human-generated with confidence 0.0766239
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i132]" time="0.014"><properties><property name="score" value="0.12406720405083171" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255135 (text: BCF GP Pro  ACQ Team $10,000,  220+  Current Camp:) is an LLM-generated sample, misclassified as human-generated with confidence 0.1240672&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255135 (text: BCF GP Pro  ACQ Team $10,000,  220+  Current Camp:) is an LLM-generated sample, misclassified as human-generated with confidence 0.1240672
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i133]" time="0.072"><properties><property name="score" value="0.12927016580827383" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255136 (text: &quot;If you run for the White House and you want peopl) is an LLM-generated sample, misclassified as human-generated with confidence 0.12927017&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255136 (text: "If you run for the White House and you want peopl) is an LLM-generated sample, misclassified as human-generated with confidence 0.12927017
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i134]" time="0.067"><properties><property name="score" value="0.13339481439533118" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255137 (text: Some Tips to Gain a Competitive Advantage  A commo) is an LLM-generated sample, misclassified as human-generated with confidence 0.13339481&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255137 (text: Some Tips to Gain a Competitive Advantage  A commo) is an LLM-generated sample, misclassified as human-generated with confidence 0.13339481
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i135]" time="0.016"><properties><property name="score" value="0.312106430868736" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255138 (text: It seems that set 2 of the SNES classic Skies of A) is an LLM-generated sample, misclassified as human-generated with confidence 0.31210643&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255138 (text: It seems that set 2 of the SNES classic Skies of A) is an LLM-generated sample, misclassified as human-generated with confidence 0.31210643
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i136]" time="0.078"><properties><property name="score" value="0.11368907856359671" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255139 (text: &quot;Spear Hands, I know this is a new duty for you, b) is an LLM-generated sample, misclassified as human-generated with confidence 0.11368908&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255139 (text: "Spear Hands, I know this is a new duty for you, b) is an LLM-generated sample, misclassified as human-generated with confidence 0.11368908
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i137]" time="0.077"><properties><property name="score" value="0.17275010348887015" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255140 (text: Electrify Spring Low-Cost Solar Power To Electrici) is an LLM-generated sample, misclassified as human-generated with confidence 0.1727501&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255140 (text: Electrify Spring Low-Cost Solar Power To Electrici) is an LLM-generated sample, misclassified as human-generated with confidence 0.1727501
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i138]" time="0.027"><properties><property name="score" value="0.10236918899338182" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255141 (text: SpaceX confirms all three have been arrested Copyr) is an LLM-generated sample, misclassified as human-generated with confidence 0.10236919&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255141 (text: SpaceX confirms all three have been arrested Copyr) is an LLM-generated sample, misclassified as human-generated with confidence 0.10236919
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i139]" time="0.084"><properties><property name="score" value="0.1475260197900381" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255142 (text: http://gty.im/621026313  Some can argue the folks ) is an LLM-generated sample, misclassified as human-generated with confidence 0.14752602&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255142 (text: http://gty.im/621026313  Some can argue the folks ) is an LLM-generated sample, misclassified as human-generated with confidence 0.14752602
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i140]" time="0.088"><properties><property name="score" value="0.14904642209834176" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255143 (text: Once the last research work had been completed, I ) is an LLM-generated sample, misclassified as human-generated with confidence 0.14904642&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255143 (text: Once the last research work had been completed, I ) is an LLM-generated sample, misclassified as human-generated with confidence 0.14904642
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i141]" time="0.029"><properties><property name="score" value="0.11048467893150968" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255144 (text: Don't trust fake news  Help journalists by signing) is an LLM-generated sample, misclassified as human-generated with confidence 0.11048468&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255144 (text: Don't trust fake news  Help journalists by signing) is an LLM-generated sample, misclassified as human-generated with confidence 0.11048468
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i142]" time="0.013"><properties><property name="score" value="0.06390490434858842" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255145 (text: Aha!Could be his questioned tongue!Huehuehue Dogec) is an LLM-generated sample, misclassified as human-generated with confidence 0.0639049&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255145 (text: Aha!Could be his questioned tongue!Huehuehue Dogec) is an LLM-generated sample, misclassified as human-generated with confidence 0.0639049
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i143]" time="0.082"><properties><property name="score" value="0.2947737579372821" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255146 (text: Madden 18 Patch Notes - November 9th, 2018  Patch ) is an LLM-generated sample, misclassified as human-generated with confidence 0.29477376&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255146 (text: Madden 18 Patch Notes - November 9th, 2018  Patch ) is an LLM-generated sample, misclassified as human-generated with confidence 0.29477376
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i144]" time="0.018"><properties><property name="score" value="0.03526046730917831" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255147 (text: favorite favorite favorite favorite favorite  - 4 ) is an LLM-generated sample, misclassified as human-generated with confidence 0.03526047&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255147 (text: favorite favorite favorite favorite favorite  - 4 ) is an LLM-generated sample, misclassified as human-generated with confidence 0.03526047
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i145]" time="0.082"><properties><property name="score" value="0.1546600645417154" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255148 (text: The film begins with Champagne in the cloud. Upon ) is an LLM-generated sample, misclassified as human-generated with confidence 0.15466006&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255148 (text: The film begins with Champagne in the cloud. Upon ) is an LLM-generated sample, misclassified as human-generated with confidence 0.15466006
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i146]" time="0.071"><properties><property name="score" value="0.16945301717663946" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255149 (text: Chamberlain rescued the fans with his picture  Liv) is an LLM-generated sample, misclassified as human-generated with confidence 0.16945302&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255149 (text: Chamberlain rescued the fans with his picture  Liv) is an LLM-generated sample, misclassified as human-generated with confidence 0.16945302
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i147]" time="0.086"><properties><property name="score" value="0.08204334829381488" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255150 (text: On eight occasions since 2011, U.S. companies that) is an LLM-generated sample, misclassified as human-generated with confidence 0.08204335&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255150 (text: On eight occasions since 2011, U.S. companies that) is an LLM-generated sample, misclassified as human-generated with confidence 0.08204335
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i148]" time="0.031"><properties><property name="score" value="0.12140458324524457" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255151 (text: Keyslinging. Driving fast on one side of a state l) is an LLM-generated sample, misclassified as human-generated with confidence 0.12140458&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255151 (text: Keyslinging. Driving fast on one side of a state l) is an LLM-generated sample, misclassified as human-generated with confidence 0.12140458
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i149]" time="0.019"><properties><property name="score" value="0.04077338118378959" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255152 (text: 2014/07/16 02:23  SEOUL, July 15 (Yonhap) -- North) is an LLM-generated sample, misclassified as human-generated with confidence 0.04077338&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255152 (text: 2014/07/16 02:23  SEOUL, July 15 (Yonhap) -- North) is an LLM-generated sample, misclassified as human-generated with confidence 0.04077338
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i150]" time="0.107"><properties><property name="score" value="0.10122255701384077" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255153 (text: They recommend stopping normal exercise 45 minutes) is an LLM-generated sample, misclassified as human-generated with confidence 0.10122256&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255153 (text: They recommend stopping normal exercise 45 minutes) is an LLM-generated sample, misclassified as human-generated with confidence 0.10122256
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i151]" time="0.086"><properties><property name="score" value="0.22559684144219222" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255154 (text: Was coddled when he was a kid, indifference that c) is an LLM-generated sample, misclassified as human-generated with confidence 0.22559684&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255154 (text: Was coddled when he was a kid, indifference that c) is an LLM-generated sample, misclassified as human-generated with confidence 0.22559684
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i152]" time="0.071"><properties><property name="score" value="0.12465253457831899" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255155 (text: TL: I sure am. I will try to edit this slightly an) is an LLM-generated sample, misclassified as human-generated with confidence 0.12465253&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255155 (text: TL: I sure am. I will try to edit this slightly an) is an LLM-generated sample, misclassified as human-generated with confidence 0.12465253
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i153]" time="0.091"><properties><property name="score" value="0.11565365578560526" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255156 (text: Sweden's at a crossroads. Within the next two year) is an LLM-generated sample, misclassified as human-generated with confidence 0.11565366&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255156 (text: Sweden's at a crossroads. Within the next two year) is an LLM-generated sample, misclassified as human-generated with confidence 0.11565366
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i154]" time="0.021"><properties><property name="score" value="0.0283124895711917" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255157 (text: Lethe crocodiles are commonly found in Australia. ) is an LLM-generated sample, misclassified as human-generated with confidence 0.02831249&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255157 (text: Lethe crocodiles are commonly found in Australia. ) is an LLM-generated sample, misclassified as human-generated with confidence 0.02831249
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i155]" time="0.071"><properties><property name="score" value="0.10671603250855337" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255158 (text: Sig-Ape Laser Engraved Build  Sig-Ape mailed us th) is an LLM-generated sample, misclassified as human-generated with confidence 0.10671603&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255158 (text: Sig-Ape Laser Engraved Build  Sig-Ape mailed us th) is an LLM-generated sample, misclassified as human-generated with confidence 0.10671603
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i156]" time="0.014"><properties><property name="score" value="0.08539779456719354" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255159 (text: Meet The New York Rangers  The New York Rangers su) is an LLM-generated sample, misclassified as human-generated with confidence 0.08539779&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255159 (text: Meet The New York Rangers  The New York Rangers su) is an LLM-generated sample, misclassified as human-generated with confidence 0.08539779
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i157]" time="0.102"><properties><property name="score" value="0.12041101438030777" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255160 (text: Revealed: The Men Behind the Pain Dispensers You D) is an LLM-generated sample, misclassified as human-generated with confidence 0.12041101&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255160 (text: Revealed: The Men Behind the Pain Dispensers You D) is an LLM-generated sample, misclassified as human-generated with confidence 0.12041101
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i158]" time="0.073"><properties><property name="score" value="0.09753187458286265" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255161 (text: [W]hat used to be only a few space ships a generat) is an LLM-generated sample, misclassified as human-generated with confidence 0.09753187&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255161 (text: [W]hat used to be only a few space ships a generat) is an LLM-generated sample, misclassified as human-generated with confidence 0.09753187
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i159]" time="0.016"><properties><property name="score" value="0.020074146875548315" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255162 (text: At the April 23, 2016, state, federal and local he) is an LLM-generated sample, misclassified as human-generated with confidence 0.02007415&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255162 (text: At the April 23, 2016, state, federal and local he) is an LLM-generated sample, misclassified as human-generated with confidence 0.02007415
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i160]" time="0.066"><properties><property name="score" value="0.018045982611767795" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255163 (text: Vericon Performance System consists of dual T400 s) is an LLM-generated sample, misclassified as human-generated with confidence 0.01804598&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255163 (text: Vericon Performance System consists of dual T400 s) is an LLM-generated sample, misclassified as human-generated with confidence 0.01804598
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i161]" time="0.080"><properties><property name="score" value="0.09775612169295946" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255164 (text: BetaClient module describes how to send or receive) is an LLM-generated sample, misclassified as human-generated with confidence 0.09775612&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255164 (text: BetaClient module describes how to send or receive) is an LLM-generated sample, misclassified as human-generated with confidence 0.09775612
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i162]" time="0.091"><properties><property name="score" value="0.37659632361458223" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255165 (text: ZCZC MIATCPAT5 ALL TTAA00 KNHC DDHHMM BULLETIN Tro) is an LLM-generated sample, misclassified as human-generated with confidence 0.37659632&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255165 (text: ZCZC MIATCPAT5 ALL TTAA00 KNHC DDHHMM BULLETIN Tro) is an LLM-generated sample, misclassified as human-generated with confidence 0.37659632
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i163]" time="0.029"><properties><property name="score" value="0.13555803091892532" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255166 (text: If you've got day one of summer aboard a train or ) is an LLM-generated sample, misclassified as human-generated with confidence 0.13555803&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255166 (text: If you've got day one of summer aboard a train or ) is an LLM-generated sample, misclassified as human-generated with confidence 0.13555803
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i164]" time="0.077"><properties><property name="score" value="0.11354066791298198" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255167 (text: American Freedom Network's film Renegade Heffron w) is an LLM-generated sample, misclassified as human-generated with confidence 0.11354067&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255167 (text: American Freedom Network's film Renegade Heffron w) is an LLM-generated sample, misclassified as human-generated with confidence 0.11354067
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i165]" time="0.079"><properties><property name="score" value="0.17044066755861162" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255168 (text: Obama didn't plaster his office in Amsterdam with ) is an LLM-generated sample, misclassified as human-generated with confidence 0.17044067&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255168 (text: Obama didn't plaster his office in Amsterdam with ) is an LLM-generated sample, misclassified as human-generated with confidence 0.17044067
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i166]" time="0.027"><properties><property name="score" value="0.03950115336295028" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255169 (text: IGF1 and PGC1α are key signaling targets during de) is an LLM-generated sample, misclassified as human-generated with confidence 0.03950115&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255169 (text: IGF1 and PGC1α are key signaling targets during de) is an LLM-generated sample, misclassified as human-generated with confidence 0.03950115
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i167]" time="0.092"><properties><property name="score" value="0.13274080543792152" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255170 (text: One of the most frequently asked questions we rece) is an LLM-generated sample, misclassified as human-generated with confidence 0.13274081&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255170 (text: One of the most frequently asked questions we rece) is an LLM-generated sample, misclassified as human-generated with confidence 0.13274081
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i168]" time="0.027"><properties><property name="score" value="0.11914944129566357" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255171 (text: DEMO TURF CATEGORIES:  - Cyclo-cross  - Double tra) is an LLM-generated sample, misclassified as human-generated with confidence 0.11914944&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255171 (text: DEMO TURF CATEGORIES:  - Cyclo-cross  - Double tra) is an LLM-generated sample, misclassified as human-generated with confidence 0.11914944
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i169]" time="0.066"><properties><property name="score" value="0.2106273522631169" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255172 (text: Overview (5)  Mini Bio (1)  Randy Quaid was origin) is an LLM-generated sample, misclassified as human-generated with confidence 0.21062735&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255172 (text: Overview (5)  Mini Bio (1)  Randy Quaid was origin) is an LLM-generated sample, misclassified as human-generated with confidence 0.21062735
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i170]" time="0.075"><properties><property name="score" value="0.09910349020767567" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255173 (text: Would you like to dream about being appreciated in) is an LLM-generated sample, misclassified as human-generated with confidence 0.09910349&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255173 (text: Would you like to dream about being appreciated in) is an LLM-generated sample, misclassified as human-generated with confidence 0.09910349
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i171]" time="0.084"><properties><property name="score" value="0.1086631874883054" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255174 (text: What made Eva Marie what she was wasn't thinking o) is an LLM-generated sample, misclassified as human-generated with confidence 0.10866319&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255174 (text: What made Eva Marie what she was wasn't thinking o) is an LLM-generated sample, misclassified as human-generated with confidence 0.10866319
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i172]" time="0.027"><properties><property name="score" value="0.13545743535765112" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255175 (text: The Cabinet has agreed to back the introduction of) is an LLM-generated sample, misclassified as human-generated with confidence 0.13545744&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255175 (text: The Cabinet has agreed to back the introduction of) is an LLM-generated sample, misclassified as human-generated with confidence 0.13545744
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i173]" time="0.015"><properties><property name="score" value="0.18027831405245665" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255176 (text: Despite Emily Topp's sex scenes, she is disinteres) is an LLM-generated sample, misclassified as human-generated with confidence 0.18027831&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255176 (text: Despite Emily Topp's sex scenes, she is disinteres) is an LLM-generated sample, misclassified as human-generated with confidence 0.18027831
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i174]" time="0.014"><properties><property name="score" value="0.03129459969213322" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255177 (text: However, even without their access to renewable so) is an LLM-generated sample, misclassified as human-generated with confidence 0.0312946&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255177 (text: However, even without their access to renewable so) is an LLM-generated sample, misclassified as human-generated with confidence 0.0312946
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i175]" time="0.083"><properties><property name="score" value="0.11584990193404214" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255178 (text: Contents show]  Skins Edit  Thewas a warrior facti) is an LLM-generated sample, misclassified as human-generated with confidence 0.1158499&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255178 (text: Contents show]  Skins Edit  Thewas a warrior facti) is an LLM-generated sample, misclassified as human-generated with confidence 0.1158499
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i176]" time="0.030"><properties><property name="score" value="0.14549832130939636" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255179 (text: Job seekers applying to work in the UK under the S) is an LLM-generated sample, misclassified as human-generated with confidence 0.14549832&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255179 (text: Job seekers applying to work in the UK under the S) is an LLM-generated sample, misclassified as human-generated with confidence 0.14549832
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i177]" time="0.086"><properties><property name="score" value="0.11487146338867382" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255180 (text: ImageMoney  Note: The measly weed market happiness) is an LLM-generated sample, misclassified as human-generated with confidence 0.11487146&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255180 (text: ImageMoney  Note: The measly weed market happiness) is an LLM-generated sample, misclassified as human-generated with confidence 0.11487146
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i178]" time="0.078"><properties><property name="score" value="0.14520097791651385" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255181 (text: The early years for the Dallas Stars were not a gr) is an LLM-generated sample, misclassified as human-generated with confidence 0.14520098&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255181 (text: The early years for the Dallas Stars were not a gr) is an LLM-generated sample, misclassified as human-generated with confidence 0.14520098
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i179]" time="0.016"><properties><property name="score" value="0.028057010320819975" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255182 (text: In the predawn darkness a massive world breathes f) is an LLM-generated sample, misclassified as human-generated with confidence 0.02805701&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255182 (text: In the predawn darkness a massive world breathes f) is an LLM-generated sample, misclassified as human-generated with confidence 0.02805701
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i180]" time="0.082"><properties><property name="score" value="0.17251339647611702" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255183 (text: BOSTON (CBS) — Brandon McManus was just getting st) is an LLM-generated sample, misclassified as human-generated with confidence 0.1725134&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255183 (text: BOSTON (CBS) — Brandon McManus was just getting st) is an LLM-generated sample, misclassified as human-generated with confidence 0.1725134
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i181]" time="0.077"><properties><property name="score" value="0.1361529709767023" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255184 (text: A morning workout of exercises that emphasizes cal) is an LLM-generated sample, misclassified as human-generated with confidence 0.13615297&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255184 (text: A morning workout of exercises that emphasizes cal) is an LLM-generated sample, misclassified as human-generated with confidence 0.13615297
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i182]" time="0.018"><properties><property name="score" value="0.11645511174728229" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255185 (text: Despite the fact that you operate at a statistical) is an LLM-generated sample, misclassified as human-generated with confidence 0.11645511&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255185 (text: Despite the fact that you operate at a statistical) is an LLM-generated sample, misclassified as human-generated with confidence 0.11645511
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i183]" time="0.013"><properties><property name="score" value="0.08875927063399236" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255186 (text: Whistleblower vant bracelet  It is a warm day. The) is an LLM-generated sample, misclassified as human-generated with confidence 0.08875927&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255186 (text: Whistleblower vant bracelet  It is a warm day. The) is an LLM-generated sample, misclassified as human-generated with confidence 0.08875927
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i184]" time="0.025"><properties><property name="score" value="0.18531374345406876" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255187 (text: Chris Evans is in talks for the Hannibal role of M) is an LLM-generated sample, misclassified as human-generated with confidence 0.18531374&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255187 (text: Chris Evans is in talks for the Hannibal role of M) is an LLM-generated sample, misclassified as human-generated with confidence 0.18531374
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i185]" time="0.013"><properties><property name="score" value="0.14761191393095063" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255188 (text: In an effort to rectify the problem, the FA has no) is an LLM-generated sample, misclassified as human-generated with confidence 0.14761191&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255188 (text: In an effort to rectify the problem, the FA has no) is an LLM-generated sample, misclassified as human-generated with confidence 0.14761191
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i186]" time="0.068"><properties><property name="score" value="0.15372197977999902" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255189 (text: Chicken wings are a meal I'm perfectly happy eatin) is an LLM-generated sample, misclassified as human-generated with confidence 0.15372198&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255189 (text: Chicken wings are a meal I'm perfectly happy eatin) is an LLM-generated sample, misclassified as human-generated with confidence 0.15372198
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i187]" time="0.075"><properties><property name="score" value="0.12537363223442957" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255190 (text: Salutations, from the external assets most opposed) is an LLM-generated sample, misclassified as human-generated with confidence 0.12537363&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255190 (text: Salutations, from the external assets most opposed) is an LLM-generated sample, misclassified as human-generated with confidence 0.12537363
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i188]" time="0.018"><properties><property name="score" value="0.010559408551702454" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255191 (text: Once needed a government for from the security of ) is an LLM-generated sample, misclassified as human-generated with confidence 0.01055941&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255191 (text: Once needed a government for from the security of ) is an LLM-generated sample, misclassified as human-generated with confidence 0.01055941
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i189]" time="0.013"><properties><property name="score" value="0.01200573523242876" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255192 (text: This is Knotted Sexy… 3D Braille Booklet   This is) is an LLM-generated sample, misclassified as human-generated with confidence 0.01200574&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255192 (text: This is Knotted Sexy… 3D Braille Booklet   This is) is an LLM-generated sample, misclassified as human-generated with confidence 0.01200574
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i190]" time="0.073"><properties><property name="score" value="0.13759221112913542" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255193 (text: On Saturday, President Donald Trump proposed a cha) is an LLM-generated sample, misclassified as human-generated with confidence 0.13759221&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255193 (text: On Saturday, President Donald Trump proposed a cha) is an LLM-generated sample, misclassified as human-generated with confidence 0.13759221
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i191]" time="0.033"><properties><property name="score" value="0.18561669653424956" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255194 (text: By Hilbert Hagedoorn on: #2071  Hiroshi Matsuyama ) is an LLM-generated sample, misclassified as human-generated with confidence 0.1856167&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255194 (text: By Hilbert Hagedoorn on: #2071  Hiroshi Matsuyama ) is an LLM-generated sample, misclassified as human-generated with confidence 0.1856167
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i192]" time="0.018"><properties><property name="score" value="0.07006096417996877" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255195 (text: The report said the town could do with enhancing i) is an LLM-generated sample, misclassified as human-generated with confidence 0.07006096&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255195 (text: The report said the town could do with enhancing i) is an LLM-generated sample, misclassified as human-generated with confidence 0.07006096
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i193]" time="0.094"><properties><property name="score" value="0.07937540264972902" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255196 (text: Sheldon Adelson, a powerful figure in Republican p) is an LLM-generated sample, misclassified as human-generated with confidence 0.0793754&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255196 (text: Sheldon Adelson, a powerful figure in Republican p) is an LLM-generated sample, misclassified as human-generated with confidence 0.0793754
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i194]" time="0.111"><properties><property name="score" value="0.19513950949101533" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255197 (text: Water Damage to Sea floor: As the unique environme) is an LLM-generated sample, misclassified as human-generated with confidence 0.19513951&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255197 (text: Water Damage to Sea floor: As the unique environme) is an LLM-generated sample, misclassified as human-generated with confidence 0.19513951
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i195]" time="0.017"><properties><property name="score" value="0.07297855032050715" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255198 (text: Overall 7 Story 6 Animation 8 Sound 10 Character 7) is an LLM-generated sample, misclassified as human-generated with confidence 0.07297855&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255198 (text: Overall 7 Story 6 Animation 8 Sound 10 Character 7) is an LLM-generated sample, misclassified as human-generated with confidence 0.07297855
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i196]" time="0.077"><properties><property name="score" value="0.14073282673285126" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255199 (text: A Public Service of Akihiko Nakamatsuri Translatio) is an LLM-generated sample, misclassified as human-generated with confidence 0.14073283&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255199 (text: A Public Service of Akihiko Nakamatsuri Translatio) is an LLM-generated sample, misclassified as human-generated with confidence 0.14073283
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i197]" time="0.027"><properties><property name="score" value="0.12086492244290625" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255200 (text: It's every reader's dream to work at a major newsp) is an LLM-generated sample, misclassified as human-generated with confidence 0.12086492&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255200 (text: It's every reader's dream to work at a major newsp) is an LLM-generated sample, misclassified as human-generated with confidence 0.12086492
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i198]" time="0.084"><properties><property name="score" value="0.04883509715052921" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i199]" time="0.021"><properties><property name="score" value="0.09855095036416794" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255202 (text: A Minnesota woman was arrested after she lied abou) is an LLM-generated sample, misclassified as human-generated with confidence 0.09855095&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255202 (text: A Minnesota woman was arrested after she lied abou) is an LLM-generated sample, misclassified as human-generated with confidence 0.09855095
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i200]" time="0.075"><properties><property name="score" value="0.14191843800696344" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255203 (text: A worker clears the intersection of Palmetto and C) is an LLM-generated sample, misclassified as human-generated with confidence 0.14191844&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255203 (text: A worker clears the intersection of Palmetto and C) is an LLM-generated sample, misclassified as human-generated with confidence 0.14191844
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i201]" time="0.028"><properties><property name="score" value="0.16907980890630164" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255204 (text: How do you like dragons with your black tea?  Aust) is an LLM-generated sample, misclassified as human-generated with confidence 0.16907981&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255204 (text: How do you like dragons with your black tea?  Aust) is an LLM-generated sample, misclassified as human-generated with confidence 0.16907981
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i202]" time="0.014"><properties><property name="score" value="0.16505349693618498" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255205 (text: While on vacation in Europe, over his lunch, CEO S) is an LLM-generated sample, misclassified as human-generated with confidence 0.1650535&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255205 (text: While on vacation in Europe, over his lunch, CEO S) is an LLM-generated sample, misclassified as human-generated with confidence 0.1650535
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i203]" time="0.073"><properties><property name="score" value="0.08214274134947938" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255206 (text: Editors' note: Andrew Ortega debunks a common ster) is an LLM-generated sample, misclassified as human-generated with confidence 0.08214274&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255206 (text: Editors' note: Andrew Ortega debunks a common ster) is an LLM-generated sample, misclassified as human-generated with confidence 0.08214274
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i204]" time="0.014"><properties><property name="score" value="0.1582897613216061" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255207 (text: The Twins-Red Sox series opens Monday evening, and) is an LLM-generated sample, misclassified as human-generated with confidence 0.15828976&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255207 (text: The Twins-Red Sox series opens Monday evening, and) is an LLM-generated sample, misclassified as human-generated with confidence 0.15828976
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i205]" time="0.030"><properties><property name="score" value="0.10205480971651855" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255208 (text: Illustrating the latest advancements in plant rese) is an LLM-generated sample, misclassified as human-generated with confidence 0.10205481&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255208 (text: Illustrating the latest advancements in plant rese) is an LLM-generated sample, misclassified as human-generated with confidence 0.10205481
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i206]" time="0.068"><properties><property name="score" value="0.10973232135239855" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255209 (text: This article first appeared on the Atlantic Counci) is an LLM-generated sample, misclassified as human-generated with confidence 0.10973232&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255209 (text: This article first appeared on the Atlantic Counci) is an LLM-generated sample, misclassified as human-generated with confidence 0.10973232
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i207]" time="0.015"><properties><property name="score" value="0.08626281654509071" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255210 (text: GIANTS forward Will Hayward is on track to return ) is an LLM-generated sample, misclassified as human-generated with confidence 0.08626282&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255210 (text: GIANTS forward Will Hayward is on track to return ) is an LLM-generated sample, misclassified as human-generated with confidence 0.08626282
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i208]" time="0.026"><properties><property name="score" value="0.17222026542087626" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255211 (text: Walter [wiki] is the Co-Leader of Passion DWS. His) is an LLM-generated sample, misclassified as human-generated with confidence 0.17222027&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255211 (text: Walter [wiki] is the Co-Leader of Passion DWS. His) is an LLM-generated sample, misclassified as human-generated with confidence 0.17222027
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i209]" time="0.014"><properties><property name="score" value="0.029381159096714127" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255212 (text: n.  1550s, from Latin aegaeius &quot;he surmounted,&quot; of) is an LLM-generated sample, misclassified as human-generated with confidence 0.02938116&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255212 (text: n.  1550s, from Latin aegaeius "he surmounted," of) is an LLM-generated sample, misclassified as human-generated with confidence 0.02938116
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i210]" time="0.069"><properties><property name="score" value="0.004108414674460461" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i211]" time="0.073"><properties><property name="score" value="0.10930563939106126" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255214 (text: 'We are in a cycle, going through all the cycles o) is an LLM-generated sample, misclassified as human-generated with confidence 0.10930564&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255214 (text: 'We are in a cycle, going through all the cycles o) is an LLM-generated sample, misclassified as human-generated with confidence 0.10930564
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i212]" time="0.021"><properties><property name="score" value="0.029569518959488006" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255215 (text: Police say that investigators have seized cannabis) is an LLM-generated sample, misclassified as human-generated with confidence 0.02956952&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255215 (text: Police say that investigators have seized cannabis) is an LLM-generated sample, misclassified as human-generated with confidence 0.02956952
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i213]" time="0.077"><properties><property name="score" value="0.10081828052775005" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255216 (text: Jay Gruden started his tenure as head coach of the) is an LLM-generated sample, misclassified as human-generated with confidence 0.10081828&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255216 (text: Jay Gruden started his tenure as head coach of the) is an LLM-generated sample, misclassified as human-generated with confidence 0.10081828
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i214]" time="0.068"><properties><property name="score" value="0.1277983834367248" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255217 (text: Google's most lavish corridor to date - the cactus) is an LLM-generated sample, misclassified as human-generated with confidence 0.12779838&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255217 (text: Google's most lavish corridor to date - the cactus) is an LLM-generated sample, misclassified as human-generated with confidence 0.12779838
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i215]" time="0.038"><properties><property name="score" value="0.09930888635800147" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255218 (text: Rita Redstone, the head of Viacom (which owns MTV,) is an LLM-generated sample, misclassified as human-generated with confidence 0.09930889&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255218 (text: Rita Redstone, the head of Viacom (which owns MTV,) is an LLM-generated sample, misclassified as human-generated with confidence 0.09930889
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i216]" time="0.013"><properties><property name="score" value="0.11018088395189363" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255219 (text: Dilla, the late hip-hop star who passed away last ) is an LLM-generated sample, misclassified as human-generated with confidence 0.11018088&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255219 (text: Dilla, the late hip-hop star who passed away last ) is an LLM-generated sample, misclassified as human-generated with confidence 0.11018088
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i217]" time="0.072"><properties><property name="score" value="0.1277197289388584" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255220 (text: ARC Pilot: 'I just felt like I had to say somethin) is an LLM-generated sample, misclassified as human-generated with confidence 0.12771973&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255220 (text: ARC Pilot: 'I just felt like I had to say somethin) is an LLM-generated sample, misclassified as human-generated with confidence 0.12771973
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i218]" time="0.085"><properties><property name="score" value="0.1528297182908539" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255221 (text: It's a tough new season for the Seattle Seahawks, ) is an LLM-generated sample, misclassified as human-generated with confidence 0.15282972&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255221 (text: It's a tough new season for the Seattle Seahawks, ) is an LLM-generated sample, misclassified as human-generated with confidence 0.15282972
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i219]" time="0.080"><properties><property name="score" value="0.09340079967607356" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255222 (text: 3.5.2 · Fixes a bug that was stopping some users f) is an LLM-generated sample, misclassified as human-generated with confidence 0.0934008&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255222 (text: 3.5.2 · Fixes a bug that was stopping some users f) is an LLM-generated sample, misclassified as human-generated with confidence 0.0934008
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i220]" time="0.071"><properties><property name="score" value="0.1556669567633086" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255223 (text: Satyakgir  Satyakgir are a herd of bovid-like, hor) is an LLM-generated sample, misclassified as human-generated with confidence 0.15566696&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255223 (text: Satyakgir  Satyakgir are a herd of bovid-like, hor) is an LLM-generated sample, misclassified as human-generated with confidence 0.15566696
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i221]" time="0.015"><properties><property name="score" value="0.02840077513507344" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255224 (text: /w SmartConverter 6.5's main interface (photo back) is an LLM-generated sample, misclassified as human-generated with confidence 0.02840078&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255224 (text: /w SmartConverter 6.5's main interface (photo back) is an LLM-generated sample, misclassified as human-generated with confidence 0.02840078
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i222]" time="0.014"><properties><property name="score" value="0.12377158860670812" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255225 (text: A great factoid about the Toronto Blue Jays is tha) is an LLM-generated sample, misclassified as human-generated with confidence 0.12377159&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255225 (text: A great factoid about the Toronto Blue Jays is tha) is an LLM-generated sample, misclassified as human-generated with confidence 0.12377159
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i223]" time="0.016"><properties><property name="score" value="0.08494686902172921" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255226 (text:  Decorative Morning Combs, aka Mohalla Curls or Dr) is an LLM-generated sample, misclassified as human-generated with confidence 0.08494687&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255226 (text:  Decorative Morning Combs, aka Mohalla Curls or Dr) is an LLM-generated sample, misclassified as human-generated with confidence 0.08494687
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i224]" time="0.098"><properties><property name="score" value="0.12508665609328642" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255227 (text: The Temple and Factom Foundation are jointly annou) is an LLM-generated sample, misclassified as human-generated with confidence 0.12508666&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255227 (text: The Temple and Factom Foundation are jointly annou) is an LLM-generated sample, misclassified as human-generated with confidence 0.12508666
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i225]" time="0.014"><properties><property name="score" value="0.04988516237735552" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255228 (text: Verne Troyer is an American actor, voice actor, an) is an LLM-generated sample, misclassified as human-generated with confidence 0.04988516&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255228 (text: Verne Troyer is an American actor, voice actor, an) is an LLM-generated sample, misclassified as human-generated with confidence 0.04988516
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i226]" time="0.066"><properties><property name="score" value="0.09625398167745197" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255229 (text: Another day, another cheer for Harry Potter and th) is an LLM-generated sample, misclassified as human-generated with confidence 0.09625398&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255229 (text: Another day, another cheer for Harry Potter and th) is an LLM-generated sample, misclassified as human-generated with confidence 0.09625398
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i227]" time="0.028"><properties><property name="score" value="0.09142135147803125" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255230 (text: Sedona Designs  Sedona Designs &quot;Veil Star Foam&quot; Ha) is an LLM-generated sample, misclassified as human-generated with confidence 0.09142135&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255230 (text: Sedona Designs  Sedona Designs "Veil Star Foam" Ha) is an LLM-generated sample, misclassified as human-generated with confidence 0.09142135
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i228]" time="0.014"><properties><property name="score" value="0.0701900228543606" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255231 (text: DISCLAIMER: I am doing this for fun only and do no) is an LLM-generated sample, misclassified as human-generated with confidence 0.07019002&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255231 (text: DISCLAIMER: I am doing this for fun only and do no) is an LLM-generated sample, misclassified as human-generated with confidence 0.07019002
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i229]" time="0.024"><properties><property name="score" value="0.08518897151898344" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255232 (text: Ronda Rousey is getting ready for UFC 200. She'll ) is an LLM-generated sample, misclassified as human-generated with confidence 0.08518897&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255232 (text: Ronda Rousey is getting ready for UFC 200. She'll ) is an LLM-generated sample, misclassified as human-generated with confidence 0.08518897
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i230]" time="0.012"><properties><property name="score" value="0.12457777808826953" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255233 (text: Heard it said at the White House: &quot;Republicans and) is an LLM-generated sample, misclassified as human-generated with confidence 0.12457778&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255233 (text: Heard it said at the White House: "Republicans and) is an LLM-generated sample, misclassified as human-generated with confidence 0.12457778
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i231]" time="0.024"><properties><property name="score" value="0.16053825852734188" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255234 (text: I want to use the dare theft,� the Ouija board say) is an LLM-generated sample, misclassified as human-generated with confidence 0.16053826&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255234 (text: I want to use the dare theft,� the Ouija board say) is an LLM-generated sample, misclassified as human-generated with confidence 0.16053826
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i232]" time="0.065"><properties><property name="score" value="0.06711042883029672" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255235 (text: Story highlights Staffers plan to sign convention ) is an LLM-generated sample, misclassified as human-generated with confidence 0.06711043&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255235 (text: Story highlights Staffers plan to sign convention ) is an LLM-generated sample, misclassified as human-generated with confidence 0.06711043
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i233]" time="0.014"><properties><property name="score" value="0.056527266147304855" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255236 (text: No, I don't hate kitchen memberships. Bread buns i) is an LLM-generated sample, misclassified as human-generated with confidence 0.05652727&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255236 (text: No, I don't hate kitchen memberships. Bread buns i) is an LLM-generated sample, misclassified as human-generated with confidence 0.05652727
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i234]" time="0.071"><properties><property name="score" value="0.17526610944545226" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255237 (text: William Ichigo Ragnarok stunned the New York Red B) is an LLM-generated sample, misclassified as human-generated with confidence 0.17526611&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255237 (text: William Ichigo Ragnarok stunned the New York Red B) is an LLM-generated sample, misclassified as human-generated with confidence 0.17526611
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i235]" time="0.070"><properties><property name="score" value="0.06516136168259265" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255238 (text: South Korean official confirms North Korea may lau) is an LLM-generated sample, misclassified as human-generated with confidence 0.06516136&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255238 (text: South Korean official confirms North Korea may lau) is an LLM-generated sample, misclassified as human-generated with confidence 0.06516136
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i236]" time="0.028"><properties><property name="score" value="0.04423010544790335" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255239 (text: One unfortunate mother received a text message tel) is an LLM-generated sample, misclassified as human-generated with confidence 0.04423011&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255239 (text: One unfortunate mother received a text message tel) is an LLM-generated sample, misclassified as human-generated with confidence 0.04423011
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i237]" time="0.074"><properties><property name="score" value="0.17241287693578194" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255240 (text: PATHOGEN SAFETY DATA SHEET - INFECTIOUS SUBSTANCES) is an LLM-generated sample, misclassified as human-generated with confidence 0.17241288&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255240 (text: PATHOGEN SAFETY DATA SHEET - INFECTIOUS SUBSTANCES) is an LLM-generated sample, misclassified as human-generated with confidence 0.17241288
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i238]" time="0.032"><properties><property name="score" value="0.10588035277055192" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255241 (text: Jadon Roethlisberger's rehab from eight surgeries ) is an LLM-generated sample, misclassified as human-generated with confidence 0.10588035&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255241 (text: Jadon Roethlisberger's rehab from eight surgeries ) is an LLM-generated sample, misclassified as human-generated with confidence 0.10588035
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i239]" time="0.018"><properties><property name="score" value="0.04966593501842809" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255242 (text: As they did in Iowa, Michigan and Idaho, the PACs ) is an LLM-generated sample, misclassified as human-generated with confidence 0.04966594&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255242 (text: As they did in Iowa, Michigan and Idaho, the PACs ) is an LLM-generated sample, misclassified as human-generated with confidence 0.04966594
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i240]" time="0.017"><properties><property name="score" value="0.07363130768008874" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255243 (text: The Real OK vs. Snoop Dogg Debate Continues: Is th) is an LLM-generated sample, misclassified as human-generated with confidence 0.07363131&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255243 (text: The Real OK vs. Snoop Dogg Debate Continues: Is th) is an LLM-generated sample, misclassified as human-generated with confidence 0.07363131
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i241]" time="0.071"><properties><property name="score" value="0.14239662098701522" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255244 (text: NANAIMO — Water-skiing security is a risky proposi) is an LLM-generated sample, misclassified as human-generated with confidence 0.14239662&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255244 (text: NANAIMO — Water-skiing security is a risky proposi) is an LLM-generated sample, misclassified as human-generated with confidence 0.14239662
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i242]" time="0.027"><properties><property name="score" value="0.23765471368416113" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255245 (text: The chronological order of magazine pages (left to) is an LLM-generated sample, misclassified as human-generated with confidence 0.23765471&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255245 (text: The chronological order of magazine pages (left to) is an LLM-generated sample, misclassified as human-generated with confidence 0.23765471
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i243]" time="0.065"><properties><property name="score" value="0.09849096005029845" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255246 (text: Changes in 1964  Thomas goverment introduced compu) is an LLM-generated sample, misclassified as human-generated with confidence 0.09849096&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255246 (text: Changes in 1964  Thomas goverment introduced compu) is an LLM-generated sample, misclassified as human-generated with confidence 0.09849096
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i244]" time="0.013"><properties><property name="score" value="0.012727311059318458" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255247 (text: More:  More:  Family name:  Given name:  Given nam) is an LLM-generated sample, misclassified as human-generated with confidence 0.01272731&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255247 (text: More:  More:  Family name:  Given name:  Given nam) is an LLM-generated sample, misclassified as human-generated with confidence 0.01272731
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i245]" time="0.075"><properties><property name="score" value="0.2583621183897951" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255248 (text: A resident of the Pokedex neighborhood, who is dea) is an LLM-generated sample, misclassified as human-generated with confidence 0.25836212&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255248 (text: A resident of the Pokedex neighborhood, who is dea) is an LLM-generated sample, misclassified as human-generated with confidence 0.25836212
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i246]" time="0.072"><properties><property name="score" value="0.15900073706567378" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255249 (text: Tall and thin, from the red cap to the light brown) is an LLM-generated sample, misclassified as human-generated with confidence 0.15900074&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255249 (text: Tall and thin, from the red cap to the light brown) is an LLM-generated sample, misclassified as human-generated with confidence 0.15900074
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i247]" time="0.091"><properties><property name="score" value="0.08494314588112184" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255250 (text: How do you get a list of net gainers and losses?  ) is an LLM-generated sample, misclassified as human-generated with confidence 0.08494315&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255250 (text: How do you get a list of net gainers and losses?  ) is an LLM-generated sample, misclassified as human-generated with confidence 0.08494315
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i248]" time="0.090"><properties><property name="score" value="0.11721716833153542" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255251 (text: DW: A word of advice to Indonesia with Akhmed Akhm) is an LLM-generated sample, misclassified as human-generated with confidence 0.11721717&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255251 (text: DW: A word of advice to Indonesia with Akhmed Akhm) is an LLM-generated sample, misclassified as human-generated with confidence 0.11721717
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i249]" time="0.019"><properties><property name="score" value="0.06892373942347652" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255252 (text: Question  What is the exact point at which it's a ) is an LLM-generated sample, misclassified as human-generated with confidence 0.06892374&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255252 (text: Question  What is the exact point at which it's a ) is an LLM-generated sample, misclassified as human-generated with confidence 0.06892374
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i250]" time="0.018"><properties><property name="score" value="0.06562568945533975" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255253 (text: And so the pages turn, and I find myself at a loss) is an LLM-generated sample, misclassified as human-generated with confidence 0.06562569&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255253 (text: And so the pages turn, and I find myself at a loss) is an LLM-generated sample, misclassified as human-generated with confidence 0.06562569
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i251]" time="0.014"><properties><property name="score" value="0.12902118900011694" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255254 (text: fb 256×176 with Sheena Shaw and Darryl Jenkins. Fl) is an LLM-generated sample, misclassified as human-generated with confidence 0.12902119&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255254 (text: fb 256×176 with Sheena Shaw and Darryl Jenkins. Fl) is an LLM-generated sample, misclassified as human-generated with confidence 0.12902119
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i252]" time="0.061"><properties><property name="score" value="0.11662321276081444" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255255 (text: PERRY, Texas — Not long after the scheduled start ) is an LLM-generated sample, misclassified as human-generated with confidence 0.11662321&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255255 (text: PERRY, Texas — Not long after the scheduled start ) is an LLM-generated sample, misclassified as human-generated with confidence 0.11662321
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i253]" time="0.064"><properties><property name="score" value="0.12821656403712473" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255256 (text: Monthly Scouting Notes  October 2017  Jonathan Eis) is an LLM-generated sample, misclassified as human-generated with confidence 0.12821656&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255256 (text: Monthly Scouting Notes  October 2017  Jonathan Eis) is an LLM-generated sample, misclassified as human-generated with confidence 0.12821656
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i254]" time="0.064"><properties><property name="score" value="0.2015958971305431" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255257 (text: Work southern Megalopolis Sector Three, and there') is an LLM-generated sample, misclassified as human-generated with confidence 0.2015959&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255257 (text: Work southern Megalopolis Sector Three, and there') is an LLM-generated sample, misclassified as human-generated with confidence 0.2015959
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i255]" time="0.074"><properties><property name="score" value="0.1258642082159347" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255258 (text: In order to complete Book Trailer email Sign-up No) is an LLM-generated sample, misclassified as human-generated with confidence 0.12586421&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255258 (text: In order to complete Book Trailer email Sign-up No) is an LLM-generated sample, misclassified as human-generated with confidence 0.12586421
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i256]" time="0.069"><properties><property name="score" value="0.1796415914926091" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255259 (text: An image of a critical energy source will be displ) is an LLM-generated sample, misclassified as human-generated with confidence 0.17964159&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255259 (text: An image of a critical energy source will be displ) is an LLM-generated sample, misclassified as human-generated with confidence 0.17964159
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i257]" time="0.015"><properties><property name="score" value="0.10053683822200954" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255260 (text: 6 Peer-to-Peer Device  Once you identify the probl) is an LLM-generated sample, misclassified as human-generated with confidence 0.10053684&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255260 (text: 6 Peer-to-Peer Device  Once you identify the probl) is an LLM-generated sample, misclassified as human-generated with confidence 0.10053684
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i258]" time="0.072"><properties><property name="score" value="0.16330828796902597" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255261 (text: The funeral procession carrying the coffin for Col) is an LLM-generated sample, misclassified as human-generated with confidence 0.16330829&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255261 (text: The funeral procession carrying the coffin for Col) is an LLM-generated sample, misclassified as human-generated with confidence 0.16330829
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i259]" time="0.031"><properties><property name="score" value="0.10019084683068091" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255262 (text: Nice to see your interest in rolling the GoBlock, ) is an LLM-generated sample, misclassified as human-generated with confidence 0.10019085&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255262 (text: Nice to see your interest in rolling the GoBlock, ) is an LLM-generated sample, misclassified as human-generated with confidence 0.10019085
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i260]" time="0.066"><properties><property name="score" value="0.09414895397671605" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255263 (text: The Legend Of Heroes: Trails In The Sky SC Main Fe) is an LLM-generated sample, misclassified as human-generated with confidence 0.09414895&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255263 (text: The Legend Of Heroes: Trails In The Sky SC Main Fe) is an LLM-generated sample, misclassified as human-generated with confidence 0.09414895
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i261]" time="0.068"><properties><property name="score" value="0.11599401518853517" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255264 (text: The size of Indian Prime Minister Narendra Modi's ) is an LLM-generated sample, misclassified as human-generated with confidence 0.11599402&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255264 (text: The size of Indian Prime Minister Narendra Modi's ) is an LLM-generated sample, misclassified as human-generated with confidence 0.11599402
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i262]" time="0.068"><properties><property name="score" value="0.08389312173378272" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255265 (text: Paul Murray, then chief executive of Elswick Petro) is an LLM-generated sample, misclassified as human-generated with confidence 0.08389312&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255265 (text: Paul Murray, then chief executive of Elswick Petro) is an LLM-generated sample, misclassified as human-generated with confidence 0.08389312
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i263]" time="0.069"><properties><property name="score" value="0.14951546422874307" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255266 (text: Ellen Pao during trial. (Photo by Edmond Su / Staf) is an LLM-generated sample, misclassified as human-generated with confidence 0.14951546&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255266 (text: Ellen Pao during trial. (Photo by Edmond Su / Staf) is an LLM-generated sample, misclassified as human-generated with confidence 0.14951546
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i264]" time="0.014"><properties><property name="score" value="0.038412379482827275" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255267 (text: THE 10 most inspirational stories of the year  Aft) is an LLM-generated sample, misclassified as human-generated with confidence 0.03841238&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255267 (text: THE 10 most inspirational stories of the year  Aft) is an LLM-generated sample, misclassified as human-generated with confidence 0.03841238
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i265]" time="0.016"><properties><property name="score" value="0.12261964843393591" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255268 (text: The Sampedro Wingtip triple medallion Alexander Mc) is an LLM-generated sample, misclassified as human-generated with confidence 0.12261965&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255268 (text: The Sampedro Wingtip triple medallion Alexander Mc) is an LLM-generated sample, misclassified as human-generated with confidence 0.12261965
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i266]" time="0.029"><properties><property name="score" value="0.16669523213451742" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255269 (text: To philosophise, I follow the advice of the aforem) is an LLM-generated sample, misclassified as human-generated with confidence 0.16669523&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255269 (text: To philosophise, I follow the advice of the aforem) is an LLM-generated sample, misclassified as human-generated with confidence 0.16669523
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i267]" time="0.013"><properties><property name="score" value="0.052227353298966994" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255270 (text: CHARLOTTE, N.C. (AP) — North Carolina's top electi) is an LLM-generated sample, misclassified as human-generated with confidence 0.05222735&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255270 (text: CHARLOTTE, N.C. (AP) — North Carolina's top electi) is an LLM-generated sample, misclassified as human-generated with confidence 0.05222735
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i268]" time="0.013"><properties><property name="score" value="0.055363298154215146" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255271 (text: Guz. Ahhhh I forgot to post my last entry last wee) is an LLM-generated sample, misclassified as human-generated with confidence 0.0553633&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255271 (text: Guz. Ahhhh I forgot to post my last entry last wee) is an LLM-generated sample, misclassified as human-generated with confidence 0.0553633
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i269]" time="0.065"><properties><property name="score" value="0.21037205046580248" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255272 (text: The Texture Bundle  The Texture Bundle contains th) is an LLM-generated sample, misclassified as human-generated with confidence 0.21037205&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255272 (text: The Texture Bundle  The Texture Bundle contains th) is an LLM-generated sample, misclassified as human-generated with confidence 0.21037205
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i270]" time="0.026"><properties><property name="score" value="0.04999998376868122" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255273 (text: RAW IMAGES | IMAGES | INTERACTIVES | VIDEOS | MARS) is an LLM-generated sample, misclassified as human-generated with confidence 0.04999998&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255273 (text: RAW IMAGES | IMAGES | INTERACTIVES | VIDEOS | MARS) is an LLM-generated sample, misclassified as human-generated with confidence 0.04999998
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i271]" time="0.075"><properties><property name="score" value="0.10567365058166367" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255274 (text: By Kim Key  CALGARY – A blaze that spread across a) is an LLM-generated sample, misclassified as human-generated with confidence 0.10567365&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255274 (text: By Kim Key  CALGARY – A blaze that spread across a) is an LLM-generated sample, misclassified as human-generated with confidence 0.10567365
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i272]" time="0.014"><properties><property name="score" value="0.008532443205377938" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255275 (text: Subscriber Domains near information  I acquired a ) is an LLM-generated sample, misclassified as human-generated with confidence 0.00853244&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255275 (text: Subscriber Domains near information  I acquired a ) is an LLM-generated sample, misclassified as human-generated with confidence 0.00853244
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i273]" time="0.014"><properties><property name="score" value="0.24742577788096076" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255276 (text: Tracklist  01. Virtual Riot – VLA (extended cut)  ) is an LLM-generated sample, misclassified as human-generated with confidence 0.24742578&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255276 (text: Tracklist  01. Virtual Riot – VLA (extended cut)  ) is an LLM-generated sample, misclassified as human-generated with confidence 0.24742578
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i274]" time="0.013"><properties><property name="score" value="0.04892823644719324" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255277 (text: Dec 20, 2016 - Art MCGILLIAN finds love and they h) is an LLM-generated sample, misclassified as human-generated with confidence 0.04892824&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255277 (text: Dec 20, 2016 - Art MCGILLIAN finds love and they h) is an LLM-generated sample, misclassified as human-generated with confidence 0.04892824
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i275]" time="0.066"><properties><property name="score" value="0.1191246273585331" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255278 (text: reloaded kicks through the abuse threshold  for un) is an LLM-generated sample, misclassified as human-generated with confidence 0.11912463&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255278 (text: reloaded kicks through the abuse threshold  for un) is an LLM-generated sample, misclassified as human-generated with confidence 0.11912463
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i276]" time="0.071"><properties><property name="score" value="0.11276350469395349" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255279 (text: German astronaut Werner Lang, who spent seven mont) is an LLM-generated sample, misclassified as human-generated with confidence 0.1127635&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255279 (text: German astronaut Werner Lang, who spent seven mont) is an LLM-generated sample, misclassified as human-generated with confidence 0.1127635
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i277]" time="0.072"><properties><property name="score" value="0.08973300351287028" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255280 (text: A third Milwaukee Motor Speedway parking spot face) is an LLM-generated sample, misclassified as human-generated with confidence 0.089733&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255280 (text: A third Milwaukee Motor Speedway parking spot face) is an LLM-generated sample, misclassified as human-generated with confidence 0.089733
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i278]" time="0.076"><properties><property name="score" value="0.1119712347015729" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255281 (text: Sexy girls and polygonal subjects over T. She make) is an LLM-generated sample, misclassified as human-generated with confidence 0.11197123&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255281 (text: Sexy girls and polygonal subjects over T. She make) is an LLM-generated sample, misclassified as human-generated with confidence 0.11197123
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i279]" time="0.035"><properties><property name="score" value="0.09274363358041356" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255282 (text: The ATG-ACE 1000-F was an upgrade of the ATG-M10A/) is an LLM-generated sample, misclassified as human-generated with confidence 0.09274363&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255282 (text: The ATG-ACE 1000-F was an upgrade of the ATG-M10A/) is an LLM-generated sample, misclassified as human-generated with confidence 0.09274363
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i280]" time="0.075"><properties><property name="score" value="0.1537113557600865" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255283 (text: &quot;Since we first worked together in 2000 at UINP an) is an LLM-generated sample, misclassified as human-generated with confidence 0.15371136&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255283 (text: "Since we first worked together in 2000 at UINP an) is an LLM-generated sample, misclassified as human-generated with confidence 0.15371136
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i281]" time="0.013"><properties><property name="score" value="0.0698771655760766" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255284 (text: Safer Routers from TSMC  We Now Start the GA-RST18) is an LLM-generated sample, misclassified as human-generated with confidence 0.06987717&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255284 (text: Safer Routers from TSMC  We Now Start the GA-RST18) is an LLM-generated sample, misclassified as human-generated with confidence 0.06987717
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i282]" time="0.069"><properties><property name="score" value="0.11940773461484105" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255285 (text: Today's Kiwi currency thriving after long recessio) is an LLM-generated sample, misclassified as human-generated with confidence 0.11940773&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255285 (text: Today's Kiwi currency thriving after long recessio) is an LLM-generated sample, misclassified as human-generated with confidence 0.11940773
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i283]" time="0.028"><properties><property name="score" value="0.12435924204274275" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255286 (text: Taxi commuters complain that having to pay £10 ext) is an LLM-generated sample, misclassified as human-generated with confidence 0.12435924&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255286 (text: Taxi commuters complain that having to pay £10 ext) is an LLM-generated sample, misclassified as human-generated with confidence 0.12435924
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i284]" time="0.070"><properties><property name="score" value="0.22165544122143443" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255288 (text: Holy shit, you guys showed tremendous support and ) is an LLM-generated sample, misclassified as human-generated with confidence 0.22165544&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255288 (text: Holy shit, you guys showed tremendous support and ) is an LLM-generated sample, misclassified as human-generated with confidence 0.22165544
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i285]" time="0.086"><properties><property name="score" value="0.15464435568024726" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255289 (text: Spurs incur Arsenal injuries  Injuries have been a) is an LLM-generated sample, misclassified as human-generated with confidence 0.15464436&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255289 (text: Spurs incur Arsenal injuries  Injuries have been a) is an LLM-generated sample, misclassified as human-generated with confidence 0.15464436
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i286]" time="0.072"><properties><property name="score" value="0.24120550076575636" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255290 (text: and/or neck design evolved with People to People i) is an LLM-generated sample, misclassified as human-generated with confidence 0.2412055&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255290 (text: and/or neck design evolved with People to People i) is an LLM-generated sample, misclassified as human-generated with confidence 0.2412055
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i287]" time="0.076"><properties><property name="score" value="0.16483202214032894" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255291 (text: Ancient vampire and master of the mythical blood p) is an LLM-generated sample, misclassified as human-generated with confidence 0.16483202&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255291 (text: Ancient vampire and master of the mythical blood p) is an LLM-generated sample, misclassified as human-generated with confidence 0.16483202
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i288]" time="0.087"><properties><property name="score" value="0.14635688497981045" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255292 (text: and so far, they've done a pretty amazing job.  La) is an LLM-generated sample, misclassified as human-generated with confidence 0.14635688&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255292 (text: and so far, they've done a pretty amazing job.  La) is an LLM-generated sample, misclassified as human-generated with confidence 0.14635688
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i289]" time="0.029"><properties><property name="score" value="0.19264066056532486" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255293 (text: One of the novelties in yarn progression is tables) is an LLM-generated sample, misclassified as human-generated with confidence 0.19264066&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255293 (text: One of the novelties in yarn progression is tables) is an LLM-generated sample, misclassified as human-generated with confidence 0.19264066
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i290]" time="0.069"><properties><property name="score" value="0.18657658073046499" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255294 (text: With Denver puck queen Jason Golden every week, we) is an LLM-generated sample, misclassified as human-generated with confidence 0.18657658&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255294 (text: With Denver puck queen Jason Golden every week, we) is an LLM-generated sample, misclassified as human-generated with confidence 0.18657658
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i291]" time="0.016"><properties><property name="score" value="0.06098347511786506" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255295 (text: Need more Star Wars news? Darth Vader is going to ) is an LLM-generated sample, misclassified as human-generated with confidence 0.06098348&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255295 (text: Need more Star Wars news? Darth Vader is going to ) is an LLM-generated sample, misclassified as human-generated with confidence 0.06098348
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i292]" time="0.026"><properties><property name="score" value="0.15082495845412786" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255296 (text: Others at some locations are willing to see multip) is an LLM-generated sample, misclassified as human-generated with confidence 0.15082496&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255296 (text: Others at some locations are willing to see multip) is an LLM-generated sample, misclassified as human-generated with confidence 0.15082496
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i293]" time="0.063"><properties><property name="score" value="0.1424693347143278" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255297 (text: New 3-D printer shoots out bagels without help  A ) is an LLM-generated sample, misclassified as human-generated with confidence 0.14246933&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255297 (text: New 3-D printer shoots out bagels without help  A ) is an LLM-generated sample, misclassified as human-generated with confidence 0.14246933
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i294]" time="0.064"><properties><property name="score" value="0.13902906138746618" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255298 (text: Breaking News Emails Get breaking news alerts and ) is an LLM-generated sample, misclassified as human-generated with confidence 0.13902906&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255298 (text: Breaking News Emails Get breaking news alerts and ) is an LLM-generated sample, misclassified as human-generated with confidence 0.13902906
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i295]" time="0.014"><properties><property name="score" value="0.08410133050150148" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255299 (text: Master Engineer Estel (played by Jude Law) escapes) is an LLM-generated sample, misclassified as human-generated with confidence 0.08410133&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255299 (text: Master Engineer Estel (played by Jude Law) escapes) is an LLM-generated sample, misclassified as human-generated with confidence 0.08410133
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i296]" time="0.068"><properties><property name="score" value="0.11243564530821153" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255300 (text: War at the Top of the Food Chain - Rough Rework (O) is an LLM-generated sample, misclassified as human-generated with confidence 0.11243565&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255300 (text: War at the Top of the Food Chain - Rough Rework (O) is an LLM-generated sample, misclassified as human-generated with confidence 0.11243565
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i297]" time="0.015"><properties><property name="score" value="0.18884417006913568" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255301 (text: Related videos  Type of scandal totally devastatin) is an LLM-generated sample, misclassified as human-generated with confidence 0.18884417&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255301 (text: Related videos  Type of scandal totally devastatin) is an LLM-generated sample, misclassified as human-generated with confidence 0.18884417
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i298]" time="0.015"><properties><property name="score" value="0.027076647330398496" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255302 (text: Basic German Vocabulary for Spanish-English Learne) is an LLM-generated sample, misclassified as human-generated with confidence 0.02707665&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255302 (text: Basic German Vocabulary for Spanish-English Learne) is an LLM-generated sample, misclassified as human-generated with confidence 0.02707665
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i299]" time="0.013"><properties><property name="score" value="0.07268168871928382" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255303 (text: Uber will make all future activities involving rid) is an LLM-generated sample, misclassified as human-generated with confidence 0.07268169&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255303 (text: Uber will make all future activities involving rid) is an LLM-generated sample, misclassified as human-generated with confidence 0.07268169
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i300]" time="0.081"><properties><property name="score" value="0.13868866782812708" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255304 (text: A &quot;slam those little bastards!&quot; (Quoting Theodor H) is an LLM-generated sample, misclassified as human-generated with confidence 0.13868867&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255304 (text: A "slam those little bastards!" (Quoting Theodor H) is an LLM-generated sample, misclassified as human-generated with confidence 0.13868867
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i301]" time="0.017"><properties><property name="score" value="0.1289043161371739" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255305 (text: Replica Tamiya Headlight Set  Ceramic Motor Caps -) is an LLM-generated sample, misclassified as human-generated with confidence 0.12890432&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255305 (text: Replica Tamiya Headlight Set  Ceramic Motor Caps -) is an LLM-generated sample, misclassified as human-generated with confidence 0.12890432
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i302]" time="0.014"><properties><property name="score" value="0.0668602366750648" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255306 (text: Solo Brewing Announces New Brewery in Oklahoma Cit) is an LLM-generated sample, misclassified as human-generated with confidence 0.06686024&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255306 (text: Solo Brewing Announces New Brewery in Oklahoma Cit) is an LLM-generated sample, misclassified as human-generated with confidence 0.06686024
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i303]" time="0.078"><properties><property name="score" value="0.1856364486139954" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255307 (text: a former Royal Navy submarine that survived in Nor) is an LLM-generated sample, misclassified as human-generated with confidence 0.18563645&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255307 (text: a former Royal Navy submarine that survived in Nor) is an LLM-generated sample, misclassified as human-generated with confidence 0.18563645
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i304]" time="0.074"><properties><property name="score" value="0.14832432387846625" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255308 (text: Police chief demands changes after new video appea) is an LLM-generated sample, misclassified as human-generated with confidence 0.14832432&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255308 (text: Police chief demands changes after new video appea) is an LLM-generated sample, misclassified as human-generated with confidence 0.14832432
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i305]" time="0.078"><properties><property name="score" value="0.1909679033918066" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255309 (text: …if he could have the skills, he should focus on a) is an LLM-generated sample, misclassified as human-generated with confidence 0.1909679&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255309 (text: …if he could have the skills, he should focus on a) is an LLM-generated sample, misclassified as human-generated with confidence 0.1909679
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i306]" time="0.097"><properties><property name="score" value="0.14001698172734975" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255310 (text: How hard will you fight for justice?   Parameters ) is an LLM-generated sample, misclassified as human-generated with confidence 0.14001698&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255310 (text: How hard will you fight for justice?   Parameters ) is an LLM-generated sample, misclassified as human-generated with confidence 0.14001698
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i307]" time="0.025"><properties><property name="score" value="0.2016024352211926" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255311 (text: What artistic game superior element presaged the s) is an LLM-generated sample, misclassified as human-generated with confidence 0.20160244&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255311 (text: What artistic game superior element presaged the s) is an LLM-generated sample, misclassified as human-generated with confidence 0.20160244
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i308]" time="0.064"><properties><property name="score" value="0.08591903021259237" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255312 (text: Activists working on smarter scale models are prop) is an LLM-generated sample, misclassified as human-generated with confidence 0.08591903&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255312 (text: Activists working on smarter scale models are prop) is an LLM-generated sample, misclassified as human-generated with confidence 0.08591903
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i309]" time="0.065"><properties><property name="score" value="0.13897119035555466" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255313 (text: Today, Theresa May is suffering the partisan effec) is an LLM-generated sample, misclassified as human-generated with confidence 0.13897119&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255313 (text: Today, Theresa May is suffering the partisan effec) is an LLM-generated sample, misclassified as human-generated with confidence 0.13897119
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i310]" time="0.081"><properties><property name="score" value="0.14367640239007776" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255314 (text: Topic Number 209 - Off-Generation Vehicles  For ap) is an LLM-generated sample, misclassified as human-generated with confidence 0.1436764&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255314 (text: Topic Number 209 - Off-Generation Vehicles  For ap) is an LLM-generated sample, misclassified as human-generated with confidence 0.1436764
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i311]" time="0.084"><properties><property name="score" value="0.07852247395936816" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255315 (text: Food Milk is the most important source of protein ) is an LLM-generated sample, misclassified as human-generated with confidence 0.07852247&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255315 (text: Food Milk is the most important source of protein ) is an LLM-generated sample, misclassified as human-generated with confidence 0.07852247
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i312]" time="0.025"><properties><property name="score" value="0.06427453404161934" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255316 (text: AT&amp;T will try to convince the U.S. Supreme Court t) is an LLM-generated sample, misclassified as human-generated with confidence 0.06427453&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255316 (text: AT&amp;T will try to convince the U.S. Supreme Court t) is an LLM-generated sample, misclassified as human-generated with confidence 0.06427453
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i313]" time="0.079"><properties><property name="score" value="0.13055121837066233" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255317 (text: Facebook Twitter Google+ LinkedIn  It seems almost) is an LLM-generated sample, misclassified as human-generated with confidence 0.13055122&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255317 (text: Facebook Twitter Google+ LinkedIn  It seems almost) is an LLM-generated sample, misclassified as human-generated with confidence 0.13055122
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i314]" time="0.018"><properties><property name="score" value="0.02650584892377461" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255318 (text:  More than one-third of Arizona state university s) is an LLM-generated sample, misclassified as human-generated with confidence 0.02650585&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255318 (text:  More than one-third of Arizona state university s) is an LLM-generated sample, misclassified as human-generated with confidence 0.02650585
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i315]" time="0.085"><properties><property name="score" value="0.14597388497107147" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255319 (text: &quot;That's The Wolf of Wall Street!&quot; People to hear t) is an LLM-generated sample, misclassified as human-generated with confidence 0.14597388&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255319 (text: "That's The Wolf of Wall Street!" People to hear t) is an LLM-generated sample, misclassified as human-generated with confidence 0.14597388
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i316]" time="0.071"><properties><property name="score" value="0.15391037567375032" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255320 (text: Arras Square just happened on Jessie Jolie'sbucket) is an LLM-generated sample, misclassified as human-generated with confidence 0.15391038&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255320 (text: Arras Square just happened on Jessie Jolie'sbucket) is an LLM-generated sample, misclassified as human-generated with confidence 0.15391038
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i317]" time="0.090"><properties><property name="score" value="0.17484087421782551" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255321 (text: Dieting in Animals: Documentary  Communaire Celebr) is an LLM-generated sample, misclassified as human-generated with confidence 0.17484087&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255321 (text: Dieting in Animals: Documentary  Communaire Celebr) is an LLM-generated sample, misclassified as human-generated with confidence 0.17484087
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i318]" time="0.069"><properties><property name="score" value="0.13059161382740092" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255322 (text: Please enable Javascript to watch this video  A cl) is an LLM-generated sample, misclassified as human-generated with confidence 0.13059161&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255322 (text: Please enable Javascript to watch this video  A cl) is an LLM-generated sample, misclassified as human-generated with confidence 0.13059161
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i319]" time="0.073"><properties><property name="score" value="0.1474940204805704" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255323 (text: DIY Weapons+Stuff  Welcome to my Bushcraft website) is an LLM-generated sample, misclassified as human-generated with confidence 0.14749402&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255323 (text: DIY Weapons+Stuff  Welcome to my Bushcraft website) is an LLM-generated sample, misclassified as human-generated with confidence 0.14749402
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i320]" time="0.018"><properties><property name="score" value="0.13710362621475736" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255324 (text: SAANPD (S. A. Thai Police Bureau)  &quot;Aspirational s) is an LLM-generated sample, misclassified as human-generated with confidence 0.13710363&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255324 (text: SAANPD (S. A. Thai Police Bureau)  "Aspirational s) is an LLM-generated sample, misclassified as human-generated with confidence 0.13710363
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i321]" time="0.013"><properties><property name="score" value="0.026907866138020653" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255325 (text: The only reason you might want to create a PDF wit) is an LLM-generated sample, misclassified as human-generated with confidence 0.02690787&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255325 (text: The only reason you might want to create a PDF wit) is an LLM-generated sample, misclassified as human-generated with confidence 0.02690787
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i322]" time="0.065"><properties><property name="score" value="0.14131908298485668" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255326 (text: ***  If you're looking for one thing to get excite) is an LLM-generated sample, misclassified as human-generated with confidence 0.14131908&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255326 (text: ***  If you're looking for one thing to get excite) is an LLM-generated sample, misclassified as human-generated with confidence 0.14131908
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i323]" time="0.088"><properties><property name="score" value="0.13270528531071493" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255327 (text: Part of Star Fleet Academy®  Every child, no matte) is an LLM-generated sample, misclassified as human-generated with confidence 0.13270529&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255327 (text: Part of Star Fleet Academy®  Every child, no matte) is an LLM-generated sample, misclassified as human-generated with confidence 0.13270529
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i324]" time="0.080"><properties><property name="score" value="0.1391387723718489" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255328 (text: Participation card for a Bulgarian national footba) is an LLM-generated sample, misclassified as human-generated with confidence 0.13913877&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255328 (text: Participation card for a Bulgarian national footba) is an LLM-generated sample, misclassified as human-generated with confidence 0.13913877
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i325]" time="0.077"><properties><property name="score" value="0.054935715077596245" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255329 (text: MADISON, Wis. - Polled residents who support legal) is an LLM-generated sample, misclassified as human-generated with confidence 0.05493572&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255329 (text: MADISON, Wis. - Polled residents who support legal) is an LLM-generated sample, misclassified as human-generated with confidence 0.05493572
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i326]" time="0.069"><properties><property name="score" value="0.11690808720135623" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255330 (text: In 'Kanbad: Nights of the Dead', there are many co) is an LLM-generated sample, misclassified as human-generated with confidence 0.11690809&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255330 (text: In 'Kanbad: Nights of the Dead', there are many co) is an LLM-generated sample, misclassified as human-generated with confidence 0.11690809
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i327]" time="0.069"><properties><property name="score" value="0.1479092752762674" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255331 (text: 7th Runner-Up  Third place Venice - $15,000  The p) is an LLM-generated sample, misclassified as human-generated with confidence 0.14790928&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255331 (text: 7th Runner-Up  Third place Venice - $15,000  The p) is an LLM-generated sample, misclassified as human-generated with confidence 0.14790928
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i328]" time="0.029"><properties><property name="score" value="0.9881238652129898" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255332 (text: 35189235392018118258089619225575314975163  5002495) is an LLM-generated sample, misclassified as human-generated with confidence 0.98812387&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255332 (text: 35189235392018118258089619225575314975163  5002495) is an LLM-generated sample, misclassified as human-generated with confidence 0.98812387
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i329]" time="0.013"><properties><property name="score" value="0.052890169928071495" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255333 (text: • About the Sibelius I VGM Collection »  Andrew Gl) is an LLM-generated sample, misclassified as human-generated with confidence 0.05289017&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255333 (text: • About the Sibelius I VGM Collection »  Andrew Gl) is an LLM-generated sample, misclassified as human-generated with confidence 0.05289017
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i330]" time="0.016"><properties><property name="score" value="0.14968231711855462" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255334 (text: You are a pirate, thief, hunter, adventurer, life-) is an LLM-generated sample, misclassified as human-generated with confidence 0.14968232&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255334 (text: You are a pirate, thief, hunter, adventurer, life-) is an LLM-generated sample, misclassified as human-generated with confidence 0.14968232
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i331]" time="0.066"><properties><property name="score" value="0.1290691999379197" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255335 (text: Robert Mercer (R), the billionaire who runs a chai) is an LLM-generated sample, misclassified as human-generated with confidence 0.1290692&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255335 (text: Robert Mercer (R), the billionaire who runs a chai) is an LLM-generated sample, misclassified as human-generated with confidence 0.1290692
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i332]" time="0.066"><properties><property name="score" value="0.08390756059649238" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255336 (text: Over a dozen House Republicans are opposed to Hous) is an LLM-generated sample, misclassified as human-generated with confidence 0.08390756&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255336 (text: Over a dozen House Republicans are opposed to Hous) is an LLM-generated sample, misclassified as human-generated with confidence 0.08390756
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i333]" time="0.086"><properties><property name="score" value="0.1002002697122962" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255337 (text: CAIRO — Egyptian TV news channels are giving viewe) is an LLM-generated sample, misclassified as human-generated with confidence 0.10020027&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255337 (text: CAIRO — Egyptian TV news channels are giving viewe) is an LLM-generated sample, misclassified as human-generated with confidence 0.10020027
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i334]" time="0.017"><properties><property name="score" value="0.1200224695500185" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255338 (text: Netflix has chosen Taiwan TV drama &quot;Healing Camp&quot; ) is an LLM-generated sample, misclassified as human-generated with confidence 0.12002247&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255338 (text: Netflix has chosen Taiwan TV drama "Healing Camp" ) is an LLM-generated sample, misclassified as human-generated with confidence 0.12002247
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i335]" time="0.024"><properties><property name="score" value="0.14470355470904075" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255339 (text: Media playback is used to report on this live broa) is an LLM-generated sample, misclassified as human-generated with confidence 0.14470355&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255339 (text: Media playback is used to report on this live broa) is an LLM-generated sample, misclassified as human-generated with confidence 0.14470355
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i336]" time="0.013"><properties><property name="score" value="0.14113178257863512" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255340 (text: Like this: Like Loading...  Related  FEATURED COVE) is an LLM-generated sample, misclassified as human-generated with confidence 0.14113178&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255340 (text: Like this: Like Loading...  Related  FEATURED COVE) is an LLM-generated sample, misclassified as human-generated with confidence 0.14113178
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i337]" time="0.013"><properties><property name="score" value="0.027696063787996383" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255341 (text: &quot;There's no doubt it's spread since then, with mor) is an LLM-generated sample, misclassified as human-generated with confidence 0.02769606&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255341 (text: "There's no doubt it's spread since then, with mor) is an LLM-generated sample, misclassified as human-generated with confidence 0.02769606
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i338]" time="0.064"><properties><property name="score" value="0.19837648823427673" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255342 (text: Source: Crimescenews.pk  Unregistered oboist / com) is an LLM-generated sample, misclassified as human-generated with confidence 0.19837649&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255342 (text: Source: Crimescenews.pk  Unregistered oboist / com) is an LLM-generated sample, misclassified as human-generated with confidence 0.19837649
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i339]" time="0.014"><properties><property name="score" value="0.1737167813621876" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255343 (text: The card logo is quite charismatic and would make ) is an LLM-generated sample, misclassified as human-generated with confidence 0.17371678&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255343 (text: The card logo is quite charismatic and would make ) is an LLM-generated sample, misclassified as human-generated with confidence 0.17371678
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i340]" time="0.026"><properties><property name="score" value="0.1604633356870455" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255344 (text: There are a variety of Number mount, including F-N) is an LLM-generated sample, misclassified as human-generated with confidence 0.16046334&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255344 (text: There are a variety of Number mount, including F-N) is an LLM-generated sample, misclassified as human-generated with confidence 0.16046334
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i341]" time="0.012"><properties><property name="score" value="0.16088424890740027" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i342]" time="0.068"><properties><property name="score" value="0.1411615646544996" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255346 (text: WE STAND TALKING ABOUT THIS Soda Girl has been a p) is an LLM-generated sample, misclassified as human-generated with confidence 0.14116156&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255346 (text: WE STAND TALKING ABOUT THIS Soda Girl has been a p) is an LLM-generated sample, misclassified as human-generated with confidence 0.14116156
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i343]" time="0.069"><properties><property name="score" value="0.1248044024969711" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255347 (text: 2.09.0-BETA What's New:   • New user interface, wi) is an LLM-generated sample, misclassified as human-generated with confidence 0.1248044&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255347 (text: 2.09.0-BETA What's New:   • New user interface, wi) is an LLM-generated sample, misclassified as human-generated with confidence 0.1248044
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i344]" time="0.014"><properties><property name="score" value="0.10617123102231174" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255348 (text: In all of the urban legend speculation regarding t) is an LLM-generated sample, misclassified as human-generated with confidence 0.10617123&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255348 (text: In all of the urban legend speculation regarding t) is an LLM-generated sample, misclassified as human-generated with confidence 0.10617123
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i345]" time="0.025"><properties><property name="score" value="0.10821364018252706" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255349 (text: Quote from: Abtsis on 05/15/2013 06:52 AM  Large N) is an LLM-generated sample, misclassified as human-generated with confidence 0.10821364&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255349 (text: Quote from: Abtsis on 05/15/2013 06:52 AM  Large N) is an LLM-generated sample, misclassified as human-generated with confidence 0.10821364
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i346]" time="0.076"><properties><property name="score" value="0.07703999330297127" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i347]" time="0.033"><properties><property name="score" value="0.18103821689655605" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255351 (text: Tory LGBTQ correspondent Nadim Kobeissi. (Instagra) is an LLM-generated sample, misclassified as human-generated with confidence 0.18103822&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255351 (text: Tory LGBTQ correspondent Nadim Kobeissi. (Instagra) is an LLM-generated sample, misclassified as human-generated with confidence 0.18103822
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i348]" time="0.019"><properties><property name="score" value="0.1453397184400962" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255352 (text: Text and image capture with 2 cameras  For a long ) is an LLM-generated sample, misclassified as human-generated with confidence 0.14533972&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255352 (text: Text and image capture with 2 cameras  For a long ) is an LLM-generated sample, misclassified as human-generated with confidence 0.14533972
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i349]" time="0.065"><properties><property name="score" value="0.1655140689348392" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255353 (text: A spokesman for the NCAA defended the decision not) is an LLM-generated sample, misclassified as human-generated with confidence 0.16551407&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255353 (text: A spokesman for the NCAA defended the decision not) is an LLM-generated sample, misclassified as human-generated with confidence 0.16551407
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i350]" time="0.016"><properties><property name="score" value="0.10644023551538706" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255354 (text: Scientist Rebecca J. Blewett and her team of resea) is an LLM-generated sample, misclassified as human-generated with confidence 0.10644024&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255354 (text: Scientist Rebecca J. Blewett and her team of resea) is an LLM-generated sample, misclassified as human-generated with confidence 0.10644024
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i351]" time="0.030"><properties><property name="score" value="0.14746446614835715" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255355 (text: Tama is celebrated in autumn as a rebirth of natur) is an LLM-generated sample, misclassified as human-generated with confidence 0.14746447&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255355 (text: Tama is celebrated in autumn as a rebirth of natur) is an LLM-generated sample, misclassified as human-generated with confidence 0.14746447
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i352]" time="0.024"><properties><property name="score" value="0.08120146089683074" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255356 (text: Last season was one that really showed the charact) is an LLM-generated sample, misclassified as human-generated with confidence 0.08120146&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255356 (text: Last season was one that really showed the charact) is an LLM-generated sample, misclassified as human-generated with confidence 0.08120146
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i353]" time="0.013"><properties><property name="score" value="0.10938402148152893" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255357 (text: SCOTTISH LEADERS have approved the country's first) is an LLM-generated sample, misclassified as human-generated with confidence 0.10938402&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255357 (text: SCOTTISH LEADERS have approved the country's first) is an LLM-generated sample, misclassified as human-generated with confidence 0.10938402
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i354]" time="0.012"><properties><property name="score" value="0.06816630666359735" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255358 (text: In interviews published early Saturday, the second) is an LLM-generated sample, misclassified as human-generated with confidence 0.06816631&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255358 (text: In interviews published early Saturday, the second) is an LLM-generated sample, misclassified as human-generated with confidence 0.06816631
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i355]" time="0.012"><properties><property name="score" value="0.042507445354822426" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255359 (text: Bad Android Keyboard Replacement Project For /ph/ ) is an LLM-generated sample, misclassified as human-generated with confidence 0.04250745&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255359 (text: Bad Android Keyboard Replacement Project For /ph/ ) is an LLM-generated sample, misclassified as human-generated with confidence 0.04250745
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i356]" time="0.062"><properties><property name="score" value="0.21501200781564034" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255360 (text: Some stats from the AHL's preseason games. Still, ) is an LLM-generated sample, misclassified as human-generated with confidence 0.21501201&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255360 (text: Some stats from the AHL's preseason games. Still, ) is an LLM-generated sample, misclassified as human-generated with confidence 0.21501201
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i357]" time="0.072"><properties><property name="score" value="0.07006449532556708" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255361 (text: Enhanced Escalation of Hydraulic Fracturing Everyw) is an LLM-generated sample, misclassified as human-generated with confidence 0.0700645&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255361 (text: Enhanced Escalation of Hydraulic Fracturing Everyw) is an LLM-generated sample, misclassified as human-generated with confidence 0.0700645
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i358]" time="0.014"><properties><property name="score" value="0.046359110089477795" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255362 (text: Netflix is blocking people from using two Firefox ) is an LLM-generated sample, misclassified as human-generated with confidence 0.04635911&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255362 (text: Netflix is blocking people from using two Firefox ) is an LLM-generated sample, misclassified as human-generated with confidence 0.04635911
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i359]" time="0.073"><properties><property name="score" value="0.22135373315695958" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255363 (text: L Publicarch, S. Egerton, J. Möller, J. Köhler, C.) is an LLM-generated sample, misclassified as human-generated with confidence 0.22135373&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255363 (text: L Publicarch, S. Egerton, J. Möller, J. Köhler, C.) is an LLM-generated sample, misclassified as human-generated with confidence 0.22135373
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i360]" time="0.030"><properties><property name="score" value="0.15488527550725875" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255364 (text: Green-skin has its origins in China. White skin ar) is an LLM-generated sample, misclassified as human-generated with confidence 0.15488528&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255364 (text: Green-skin has its origins in China. White skin ar) is an LLM-generated sample, misclassified as human-generated with confidence 0.15488528
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i361]" time="0.089"><properties><property name="score" value="0.14421496691970057" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255365 (text: Swiss parties Yvette Clarke (left) and 64-year-old) is an LLM-generated sample, misclassified as human-generated with confidence 0.14421497&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255365 (text: Swiss parties Yvette Clarke (left) and 64-year-old) is an LLM-generated sample, misclassified as human-generated with confidence 0.14421497
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i362]" time="0.094"><properties><property name="score" value="0.11065903321699391" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255366 (text: 「It's pretty cool」(Kazura)  「Yeah, pretty cool…… T) is an LLM-generated sample, misclassified as human-generated with confidence 0.11065903&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255366 (text: 「It's pretty cool」(Kazura)  「Yeah, pretty cool…… T) is an LLM-generated sample, misclassified as human-generated with confidence 0.11065903
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i363]" time="0.079"><properties><property name="score" value="0.1031918623622043" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255367 (text: The Railroad Headquarters (ブレイクユニ十二音ブログフィールド) was ) is an LLM-generated sample, misclassified as human-generated with confidence 0.10319186&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255367 (text: The Railroad Headquarters (ブレイクユニ十二音ブログフィールド) was ) is an LLM-generated sample, misclassified as human-generated with confidence 0.10319186
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i364]" time="0.013"><properties><property name="score" value="0.049776808011653406" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255368 (text: The Z1 is a high-end Zbox-branded wireless gaming ) is an LLM-generated sample, misclassified as human-generated with confidence 0.04977681&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255368 (text: The Z1 is a high-end Zbox-branded wireless gaming ) is an LLM-generated sample, misclassified as human-generated with confidence 0.04977681
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i365]" time="0.121"><properties><property name="score" value="0.10710276426762053" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255369 (text: By John Feffer, Victoria Scofield: Why did the pla) is an LLM-generated sample, misclassified as human-generated with confidence 0.10710276&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255369 (text: By John Feffer, Victoria Scofield: Why did the pla) is an LLM-generated sample, misclassified as human-generated with confidence 0.10710276
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i366]" time="0.078"><properties><property name="score" value="0.10819921244367932" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255370 (text: New UK banknotes/tokens from 2020 to be comprehens) is an LLM-generated sample, misclassified as human-generated with confidence 0.10819921&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255370 (text: New UK banknotes/tokens from 2020 to be comprehens) is an LLM-generated sample, misclassified as human-generated with confidence 0.10819921
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i367]" time="0.078"><properties><property name="score" value="0.1550564159410795" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255371 (text: Jean-Paul Sartre is a great philosopher, critic an) is an LLM-generated sample, misclassified as human-generated with confidence 0.15505642&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255371 (text: Jean-Paul Sartre is a great philosopher, critic an) is an LLM-generated sample, misclassified as human-generated with confidence 0.15505642
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i368]" time="0.087"><properties><property name="score" value="0.13441149532075414" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255372 (text: Oakland Raiders safety Charles Woodson — grouchy, ) is an LLM-generated sample, misclassified as human-generated with confidence 0.1344115&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255372 (text: Oakland Raiders safety Charles Woodson — grouchy, ) is an LLM-generated sample, misclassified as human-generated with confidence 0.1344115
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i369]" time="0.033"><properties><property name="score" value="0.1385726310546967" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255373 (text: Kimi Raikkonen, Lewis Hamilton and Sebastian Vette) is an LLM-generated sample, misclassified as human-generated with confidence 0.13857263&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255373 (text: Kimi Raikkonen, Lewis Hamilton and Sebastian Vette) is an LLM-generated sample, misclassified as human-generated with confidence 0.13857263
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i370]" time="0.027"><properties><property name="score" value="0.051125619637849444" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255374 (text: Decision to sell company upheld by high court: Alm) is an LLM-generated sample, misclassified as human-generated with confidence 0.05112562&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255374 (text: Decision to sell company upheld by high court: Alm) is an LLM-generated sample, misclassified as human-generated with confidence 0.05112562
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i371]" time="0.085"><properties><property name="score" value="0.19186540280121553" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255375 (text: FIRST TEST, THURSDAY  WITH ROB BAGGALO  44th minut) is an LLM-generated sample, misclassified as human-generated with confidence 0.1918654&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255375 (text: FIRST TEST, THURSDAY  WITH ROB BAGGALO  44th minut) is an LLM-generated sample, misclassified as human-generated with confidence 0.1918654
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i372]" time="0.033"><properties><property name="score" value="0.11996336796315404" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255376 (text: My father called me.  &quot;… Nneka Nosyika. Why are yo) is an LLM-generated sample, misclassified as human-generated with confidence 0.11996337&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255376 (text: My father called me.  "… Nneka Nosyika. Why are yo) is an LLM-generated sample, misclassified as human-generated with confidence 0.11996337
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i373]" time="0.016"><properties><property name="score" value="0.10687679627676139" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255377 (text: I just thoroughly enjoy Maurice Conlon's work.....) is an LLM-generated sample, misclassified as human-generated with confidence 0.1068768&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255377 (text: I just thoroughly enjoy Maurice Conlon's work.....) is an LLM-generated sample, misclassified as human-generated with confidence 0.1068768
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i374]" time="0.014"><properties><property name="score" value="0.2867942871446427" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255378 (text: Back to page 1  Week 6's specialist so far wants t) is an LLM-generated sample, misclassified as human-generated with confidence 0.28679429&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255378 (text: Back to page 1  Week 6's specialist so far wants t) is an LLM-generated sample, misclassified as human-generated with confidence 0.28679429
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i375]" time="0.014"><properties><property name="score" value="0.08777926388178203" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255379 (text: Hoping to start a trail of victory under new champ) is an LLM-generated sample, misclassified as human-generated with confidence 0.08777926&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255379 (text: Hoping to start a trail of victory under new champ) is an LLM-generated sample, misclassified as human-generated with confidence 0.08777926
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i376]" time="0.087"><properties><property name="score" value="0.08033828842982459" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255380 (text: Border War is an entirely mod-free CrossFire campa) is an LLM-generated sample, misclassified as human-generated with confidence 0.08033829&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255380 (text: Border War is an entirely mod-free CrossFire campa) is an LLM-generated sample, misclassified as human-generated with confidence 0.08033829
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i377]" time="0.026"><properties><property name="score" value="0.026015432736384916" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255381 (text: We generally assume that women walk more slowly th) is an LLM-generated sample, misclassified as human-generated with confidence 0.02601543&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255381 (text: We generally assume that women walk more slowly th) is an LLM-generated sample, misclassified as human-generated with confidence 0.02601543
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i378]" time="0.013"><properties><property name="score" value="0.14621201364003467" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255382 (text: The 49ers are so scared they won't take their game) is an LLM-generated sample, misclassified as human-generated with confidence 0.14621201&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255382 (text: The 49ers are so scared they won't take their game) is an LLM-generated sample, misclassified as human-generated with confidence 0.14621201
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i379]" time="0.091"><properties><property name="score" value="0.12600508793092205" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255383 (text: Hello, my name is Tim, and I am an entrepreneur, w) is an LLM-generated sample, misclassified as human-generated with confidence 0.12600509&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255383 (text: Hello, my name is Tim, and I am an entrepreneur, w) is an LLM-generated sample, misclassified as human-generated with confidence 0.12600509
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i380]" time="0.036"><properties><property name="score" value="0.09250133448829548" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255384 (text: &quot;We are holding the complete material in our files) is an LLM-generated sample, misclassified as human-generated with confidence 0.09250133&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255384 (text: "We are holding the complete material in our files) is an LLM-generated sample, misclassified as human-generated with confidence 0.09250133
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i381]" time="0.019"><properties><property name="score" value="0.058825171133763954" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255385 (text: Ginaldo Rivera, president of the Miami-Dade County) is an LLM-generated sample, misclassified as human-generated with confidence 0.05882517&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255385 (text: Ginaldo Rivera, president of the Miami-Dade County) is an LLM-generated sample, misclassified as human-generated with confidence 0.05882517
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i382]" time="0.105"><properties><property name="score" value="0.10142292127847538" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255386 (text: Robotics is old hat to enterprise technology compa) is an LLM-generated sample, misclassified as human-generated with confidence 0.10142292&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255386 (text: Robotics is old hat to enterprise technology compa) is an LLM-generated sample, misclassified as human-generated with confidence 0.10142292
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i383]" time="0.020"><properties><property name="score" value="0.1336944619004743" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255388 (text: &quot;Pride comes before a fall&quot; – Prophet Muhammad  wi) is an LLM-generated sample, misclassified as human-generated with confidence 0.13369446&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255388 (text: "Pride comes before a fall" – Prophet Muhammad  wi) is an LLM-generated sample, misclassified as human-generated with confidence 0.13369446
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i384]" time="0.069"><properties><property name="score" value="0.153985610917168" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255389 (text: Raspberry Pi B+ – I have a new home! My wife and I) is an LLM-generated sample, misclassified as human-generated with confidence 0.15398561&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255389 (text: Raspberry Pi B+ – I have a new home! My wife and I) is an LLM-generated sample, misclassified as human-generated with confidence 0.15398561
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i385]" time="0.013"><properties><property name="score" value="0.17164285739774687" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255390 (text: EXFOLIATING THS JG-2 BMC9000 PRO BULF I have machi) is an LLM-generated sample, misclassified as human-generated with confidence 0.17164286&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255390 (text: EXFOLIATING THS JG-2 BMC9000 PRO BULF I have machi) is an LLM-generated sample, misclassified as human-generated with confidence 0.17164286
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i386]" time="0.013"><properties><property name="score" value="0.07882736221715425" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255391 (text: The Gotham Story Two are Original Dead PEOPLE - th) is an LLM-generated sample, misclassified as human-generated with confidence 0.07882736&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255391 (text: The Gotham Story Two are Original Dead PEOPLE - th) is an LLM-generated sample, misclassified as human-generated with confidence 0.07882736
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i387]" time="0.012"><properties><property name="score" value="0.1453478301170541" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255392 (text: Mini Heart Mini Deck  12 Cards  Smallish tape meas) is an LLM-generated sample, misclassified as human-generated with confidence 0.14534783&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255392 (text: Mini Heart Mini Deck  12 Cards  Smallish tape meas) is an LLM-generated sample, misclassified as human-generated with confidence 0.14534783
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i388]" time="0.078"><properties><property name="score" value="0.05151970268570699" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255393 (text: God I just don't understand. Reply  Parent  Thread) is an LLM-generated sample, misclassified as human-generated with confidence 0.0515197&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255393 (text: God I just don't understand. Reply  Parent  Thread) is an LLM-generated sample, misclassified as human-generated with confidence 0.0515197
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i389]" time="0.025"><properties><property name="score" value="0.14233104245756517" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255394 (text: ThinkPad Air  Available configuration  Aluminum Ca) is an LLM-generated sample, misclassified as human-generated with confidence 0.14233104&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255394 (text: ThinkPad Air  Available configuration  Aluminum Ca) is an LLM-generated sample, misclassified as human-generated with confidence 0.14233104
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i390]" time="0.023"><properties><property name="score" value="0.15957063362622065" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255395 (text: Kentucky GOP sponsored bill to save Chafee from tr) is an LLM-generated sample, misclassified as human-generated with confidence 0.15957063&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255395 (text: Kentucky GOP sponsored bill to save Chafee from tr) is an LLM-generated sample, misclassified as human-generated with confidence 0.15957063
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i391]" time="0.013"><properties><property name="score" value="0.07148954991405554" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255396 (text: Inside the Environment Stabilizer Federation  Ther) is an LLM-generated sample, misclassified as human-generated with confidence 0.07148955&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255396 (text: Inside the Environment Stabilizer Federation  Ther) is an LLM-generated sample, misclassified as human-generated with confidence 0.07148955
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i392]" time="0.013"><properties><property name="score" value="0.1443854263391442" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255397 (text: Lloyd's Stone was my Absolute Favorite App for Fam) is an LLM-generated sample, misclassified as human-generated with confidence 0.14438543&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255397 (text: Lloyd's Stone was my Absolute Favorite App for Fam) is an LLM-generated sample, misclassified as human-generated with confidence 0.14438543
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i393]" time="0.064"><properties><property name="score" value="0.15193943509055796" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255398 (text: Friendly Families  GIRLS' GROOMSHOP  &quot;Families Who) is an LLM-generated sample, misclassified as human-generated with confidence 0.15193944&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255398 (text: Friendly Families  GIRLS' GROOMSHOP  "Families Who) is an LLM-generated sample, misclassified as human-generated with confidence 0.15193944
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i394]" time="0.071"><properties><property name="score" value="0.09067932535653646" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255399 (text: SEO is one of the most important elements of any w) is an LLM-generated sample, misclassified as human-generated with confidence 0.09067933&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255399 (text: SEO is one of the most important elements of any w) is an LLM-generated sample, misclassified as human-generated with confidence 0.09067933
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i395]" time="0.074"><properties><property name="score" value="0.11760248583437893" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255400 (text: Americans are furious over the Transportation Secu) is an LLM-generated sample, misclassified as human-generated with confidence 0.11760249&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255400 (text: Americans are furious over the Transportation Secu) is an LLM-generated sample, misclassified as human-generated with confidence 0.11760249
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i396]" time="0.032"><properties><property name="score" value="0.13687807902722773" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255401 (text: A Baki National Park official is covered in a plas) is an LLM-generated sample, misclassified as human-generated with confidence 0.13687808&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255401 (text: A Baki National Park official is covered in a plas) is an LLM-generated sample, misclassified as human-generated with confidence 0.13687808
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i397]" time="0.023"><properties><property name="score" value="0.10521715202378484" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255402 (text: Jared Law (pictured) joins West Ham from Wolves af) is an LLM-generated sample, misclassified as human-generated with confidence 0.10521715&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255402 (text: Jared Law (pictured) joins West Ham from Wolves af) is an LLM-generated sample, misclassified as human-generated with confidence 0.10521715
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i398]" time="0.012"><properties><property name="score" value="0.10943576270400213" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255403 (text: On this date in 1613, with the continuing decline ) is an LLM-generated sample, misclassified as human-generated with confidence 0.10943576&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255403 (text: On this date in 1613, with the continuing decline ) is an LLM-generated sample, misclassified as human-generated with confidence 0.10943576
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i399]" time="0.070"><properties><property name="score" value="0.11313151380854111" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255404 (text: People are gullible and vulnerable to the temptati) is an LLM-generated sample, misclassified as human-generated with confidence 0.11313151&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255404 (text: People are gullible and vulnerable to the temptati) is an LLM-generated sample, misclassified as human-generated with confidence 0.11313151
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i400]" time="0.068"><properties><property name="score" value="0.14605504623681412" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255405 (text: updated: Justin Sodikoff continues his blog series) is an LLM-generated sample, misclassified as human-generated with confidence 0.14605505&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255405 (text: updated: Justin Sodikoff continues his blog series) is an LLM-generated sample, misclassified as human-generated with confidence 0.14605505
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i401]" time="0.015"><properties><property name="score" value="0.03418137881810446" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255406 (text: Hour 21: Michael Griffith. We interviewed Michael ) is an LLM-generated sample, misclassified as human-generated with confidence 0.03418138&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255406 (text: Hour 21: Michael Griffith. We interviewed Michael ) is an LLM-generated sample, misclassified as human-generated with confidence 0.03418138
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i402]" time="0.018"><properties><property name="score" value="0.14021418222815996" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255407 (text: Kanye, after supposedly backing out of doing a sol) is an LLM-generated sample, misclassified as human-generated with confidence 0.14021418&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255407 (text: Kanye, after supposedly backing out of doing a sol) is an LLM-generated sample, misclassified as human-generated with confidence 0.14021418
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i403]" time="0.017"><properties><property name="score" value="0.08474730407514453" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255408 (text: Story highlights Sanders and Clinton have differen) is an LLM-generated sample, misclassified as human-generated with confidence 0.0847473&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255408 (text: Story highlights Sanders and Clinton have differen) is an LLM-generated sample, misclassified as human-generated with confidence 0.0847473
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i404]" time="0.013"><properties><property name="score" value="0.06298509192219082" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255409 (text: Last month and after his earlier inactivity as Cho) is an LLM-generated sample, misclassified as human-generated with confidence 0.06298509&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255409 (text: Last month and after his earlier inactivity as Cho) is an LLM-generated sample, misclassified as human-generated with confidence 0.06298509
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i405]" time="0.012"><properties><property name="score" value="0.0808341727346007" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255410 (text: Savinelli Port and Bulki is an easy-to-use hiking ) is an LLM-generated sample, misclassified as human-generated with confidence 0.08083417&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255410 (text: Savinelli Port and Bulki is an easy-to-use hiking ) is an LLM-generated sample, misclassified as human-generated with confidence 0.08083417
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i406]" time="0.013"><properties><property name="score" value="0.05935670264755055" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255411 (text: As our Liberals continue to ban,&quot; wrote a Trump su) is an LLM-generated sample, misclassified as human-generated with confidence 0.0593567&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255411 (text: As our Liberals continue to ban," wrote a Trump su) is an LLM-generated sample, misclassified as human-generated with confidence 0.0593567
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i407]" time="0.065"><properties><property name="score" value="0.13417322531401757" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255412 (text: Shark Attack: Why Would Someone with an 100-ft Gia) is an LLM-generated sample, misclassified as human-generated with confidence 0.13417323&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255412 (text: Shark Attack: Why Would Someone with an 100-ft Gia) is an LLM-generated sample, misclassified as human-generated with confidence 0.13417323
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i408]" time="0.014"><properties><property name="score" value="0.10615185602558808" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255413 (text: Small bidders to get a broad selection  You can no) is an LLM-generated sample, misclassified as human-generated with confidence 0.10615186&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255413 (text: Small bidders to get a broad selection  You can no) is an LLM-generated sample, misclassified as human-generated with confidence 0.10615186
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i409]" time="0.013"><properties><property name="score" value="0.03140563015537334" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255414 (text: reddit gold gives you extra features and helps kee) is an LLM-generated sample, misclassified as human-generated with confidence 0.03140563&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255414 (text: reddit gold gives you extra features and helps kee) is an LLM-generated sample, misclassified as human-generated with confidence 0.03140563
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i410]" time="0.078"><properties><property name="score" value="0.2625522330498414" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255415 (text: Allan Durand  March  MONTRÉAL – Wei-Chin Peng beat) is an LLM-generated sample, misclassified as human-generated with confidence 0.26255223&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255415 (text: Allan Durand  March  MONTRÉAL – Wei-Chin Peng beat) is an LLM-generated sample, misclassified as human-generated with confidence 0.26255223
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i411]" time="0.034"><properties><property name="score" value="0.11302771443952767" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255416 (text: Our next chapter begins as we continue our explora) is an LLM-generated sample, misclassified as human-generated with confidence 0.11302771&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255416 (text: Our next chapter begins as we continue our explora) is an LLM-generated sample, misclassified as human-generated with confidence 0.11302771
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i412]" time="0.065"><properties><property name="score" value="0.12065777724673961" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255417 (text: Ali Ali (Ala Mubarak) is a high school student wit) is an LLM-generated sample, misclassified as human-generated with confidence 0.12065778&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255417 (text: Ali Ali (Ala Mubarak) is a high school student wit) is an LLM-generated sample, misclassified as human-generated with confidence 0.12065778
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i413]" time="0.017"><properties><property name="score" value="0.055231010757594845" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255418 (text: There are a lot of rumours going around about the ) is an LLM-generated sample, misclassified as human-generated with confidence 0.05523101&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255418 (text: There are a lot of rumours going around about the ) is an LLM-generated sample, misclassified as human-generated with confidence 0.05523101
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i414]" time="0.024"><properties><property name="score" value="0.031747308523486076" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255419 (text: California Governor Jerry Brown has signed legisla) is an LLM-generated sample, misclassified as human-generated with confidence 0.03174731&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255419 (text: California Governor Jerry Brown has signed legisla) is an LLM-generated sample, misclassified as human-generated with confidence 0.03174731
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i415]" time="0.061"><properties><property name="score" value="0.16915485975094538" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255420 (text: &quot;Hey, I keep talking to you while you're fucking g) is an LLM-generated sample, misclassified as human-generated with confidence 0.16915486&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255420 (text: "Hey, I keep talking to you while you're fucking g) is an LLM-generated sample, misclassified as human-generated with confidence 0.16915486
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i416]" time="0.013"><properties><property name="score" value="0.12575649871598982" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255421 (text: So for those of you in need of some training wheel) is an LLM-generated sample, misclassified as human-generated with confidence 0.1257565&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255421 (text: So for those of you in need of some training wheel) is an LLM-generated sample, misclassified as human-generated with confidence 0.1257565
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i417]" time="0.073"><properties><property name="score" value="0.07956655525725319" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255422 (text: Physical Therapy Diagnostic Instrumentation  Sourc) is an LLM-generated sample, misclassified as human-generated with confidence 0.07956656&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255422 (text: Physical Therapy Diagnostic Instrumentation  Sourc) is an LLM-generated sample, misclassified as human-generated with confidence 0.07956656
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i418]" time="0.085"><properties><property name="score" value="0.09947314088485443" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255423 (text: Emirates airline is offering return trips from Ice) is an LLM-generated sample, misclassified as human-generated with confidence 0.09947314&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255423 (text: Emirates airline is offering return trips from Ice) is an LLM-generated sample, misclassified as human-generated with confidence 0.09947314
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i419]" time="0.094"><properties><property name="score" value="0.11784838950396276" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255424 (text: &quot;It's like Fight Club,&quot; says Yuval Ben Zvi, presen) is an LLM-generated sample, misclassified as human-generated with confidence 0.11784839&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255424 (text: "It's like Fight Club," says Yuval Ben Zvi, presen) is an LLM-generated sample, misclassified as human-generated with confidence 0.11784839
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i420]" time="0.028"><properties><property name="score" value="0.17047556236592437" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255425 (text: Steven C. Schwalm Posted in reply to eagleparent  ) is an LLM-generated sample, misclassified as human-generated with confidence 0.17047556&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255425 (text: Steven C. Schwalm Posted in reply to eagleparent  ) is an LLM-generated sample, misclassified as human-generated with confidence 0.17047556
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i421]" time="0.013"><properties><property name="score" value="0.039871665385809996" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i422]" time="0.069"><properties><property name="score" value="0.1149571976677823" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255427 (text: The neocons sprang into action when the pugnacious) is an LLM-generated sample, misclassified as human-generated with confidence 0.1149572&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255427 (text: The neocons sprang into action when the pugnacious) is an LLM-generated sample, misclassified as human-generated with confidence 0.1149572
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i423]" time="0.012"><properties><property name="score" value="0.1652829460780214" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255428 (text: Get business updates directly to your inbox Subscr) is an LLM-generated sample, misclassified as human-generated with confidence 0.16528295&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255428 (text: Get business updates directly to your inbox Subscr) is an LLM-generated sample, misclassified as human-generated with confidence 0.16528295
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i424]" time="0.064"><properties><property name="score" value="0.08411435349002101" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255429 (text: Our comprehensive review of the Federal Court of A) is an LLM-generated sample, misclassified as human-generated with confidence 0.08411435&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255429 (text: Our comprehensive review of the Federal Court of A) is an LLM-generated sample, misclassified as human-generated with confidence 0.08411435
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i425]" time="0.066"><properties><property name="score" value="0.16059916583075862" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255430 (text: So that's it? Why are you still reading now? Proba) is an LLM-generated sample, misclassified as human-generated with confidence 0.16059917&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255430 (text: So that's it? Why are you still reading now? Proba) is an LLM-generated sample, misclassified as human-generated with confidence 0.16059917
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i426]" time="0.068"><properties><property name="score" value="0.30331465402227886" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255431 (text: #21 UST 3RD INDEPENDENCE 4 (NAC) - HIGH HOP - PICK) is an LLM-generated sample, misclassified as human-generated with confidence 0.30331465&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255431 (text: #21 UST 3RD INDEPENDENCE 4 (NAC) - HIGH HOP - PICK) is an LLM-generated sample, misclassified as human-generated with confidence 0.30331465
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i427]" time="0.070"><properties><property name="score" value="0.16671304916423121" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255432 (text: Advanced Threat Analysis: The Town of Ashland  Sum) is an LLM-generated sample, misclassified as human-generated with confidence 0.16671305&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255432 (text: Advanced Threat Analysis: The Town of Ashland  Sum) is an LLM-generated sample, misclassified as human-generated with confidence 0.16671305
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i428]" time="0.013"><properties><property name="score" value="0.06142743146007601" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255433 (text: Jinganjiang/Yunnan Province ― Leaders of China's m) is an LLM-generated sample, misclassified as human-generated with confidence 0.06142743&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255433 (text: Jinganjiang/Yunnan Province ― Leaders of China's m) is an LLM-generated sample, misclassified as human-generated with confidence 0.06142743
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i429]" time="0.012"><properties><property name="score" value="0.13492986231006254" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255434 (text: CALGARY, AB -- The Calgary Flames announced today ) is an LLM-generated sample, misclassified as human-generated with confidence 0.13492986&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255434 (text: CALGARY, AB -- The Calgary Flames announced today ) is an LLM-generated sample, misclassified as human-generated with confidence 0.13492986
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i430]" time="0.016"><properties><property name="score" value="0.038602088675332435" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255435 (text: 2015-08-20T13:38:59Z (GMT) William Frommer by (GMT) is an LLM-generated sample, misclassified as human-generated with confidence 0.03860209&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255435 (text: 2015-08-20T13:38:59Z (GMT) William Frommer by (GMT) is an LLM-generated sample, misclassified as human-generated with confidence 0.03860209
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i431]" time="0.073"><properties><property name="score" value="0.14249137339701012" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255436 (text: Thank you to every single author who has been kind) is an LLM-generated sample, misclassified as human-generated with confidence 0.14249137&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255436 (text: Thank you to every single author who has been kind) is an LLM-generated sample, misclassified as human-generated with confidence 0.14249137
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i432]" time="0.069"><properties><property name="score" value="0.002723729057006744" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255437 (text: The Title Person Recording a Acceptable Order from) is an LLM-generated sample, misclassified as human-generated with confidence 0.00272373&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255437 (text: The Title Person Recording a Acceptable Order from) is an LLM-generated sample, misclassified as human-generated with confidence 0.00272373
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i433]" time="0.079"><properties><property name="score" value="0.13971227229276864" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255438 (text: In ordinary circumstances, the veterans of politic) is an LLM-generated sample, misclassified as human-generated with confidence 0.13971227&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255438 (text: In ordinary circumstances, the veterans of politic) is an LLM-generated sample, misclassified as human-generated with confidence 0.13971227
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i434]" time="0.029"><properties><property name="score" value="0.12852225750738433" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255439 (text: North America based Manjaro Linux has been one of ) is an LLM-generated sample, misclassified as human-generated with confidence 0.12852226&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255439 (text: North America based Manjaro Linux has been one of ) is an LLM-generated sample, misclassified as human-generated with confidence 0.12852226
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i435]" time="0.029"><properties><property name="score" value="0.016439274162013545" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255440 (text: In the latest release to the Java platform, Java S) is an LLM-generated sample, misclassified as human-generated with confidence 0.01643927&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255440 (text: In the latest release to the Java platform, Java S) is an LLM-generated sample, misclassified as human-generated with confidence 0.01643927
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i436]" time="0.012"><properties><property name="score" value="0.09377485294930765" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255441 (text: Go The Distance To Write Your Novel  By: Duane Cha) is an LLM-generated sample, misclassified as human-generated with confidence 0.09377485&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255441 (text: Go The Distance To Write Your Novel  By: Duane Cha) is an LLM-generated sample, misclassified as human-generated with confidence 0.09377485
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i437]" time="0.022"><properties><property name="score" value="0.17500298339346784" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255443 (text: thelands with sheets  88 real beds  on the right  ) is an LLM-generated sample, misclassified as human-generated with confidence 0.17500298&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255443 (text: thelands with sheets  88 real beds  on the right  ) is an LLM-generated sample, misclassified as human-generated with confidence 0.17500298
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i438]" time="0.011"><properties><property name="score" value="0.04871590845632601" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255444 (text: Adrian Peterson is one of the greatest running bac) is an LLM-generated sample, misclassified as human-generated with confidence 0.04871591&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255444 (text: Adrian Peterson is one of the greatest running bac) is an LLM-generated sample, misclassified as human-generated with confidence 0.04871591
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i439]" time="0.063"><properties><property name="score" value="0.08008288997656805" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255445 (text: By Han Choi  North Korea sees the United States as) is an LLM-generated sample, misclassified as human-generated with confidence 0.08008289&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255445 (text: By Han Choi  North Korea sees the United States as) is an LLM-generated sample, misclassified as human-generated with confidence 0.08008289
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i440]" time="0.069"><properties><property name="score" value="0.14633068912418007" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255446 (text: Background:  JAMA ART EXPOSURE is the number one j) is an LLM-generated sample, misclassified as human-generated with confidence 0.14633069&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255446 (text: Background:  JAMA ART EXPOSURE is the number one j) is an LLM-generated sample, misclassified as human-generated with confidence 0.14633069
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i441]" time="0.022"><properties><property name="score" value="0.17159119936654535" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255447 (text: Cream of the crop [aka JD]  Using Jenkins to compi) is an LLM-generated sample, misclassified as human-generated with confidence 0.1715912&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255447 (text: Cream of the crop [aka JD]  Using Jenkins to compi) is an LLM-generated sample, misclassified as human-generated with confidence 0.1715912
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i442]" time="0.012"><properties><property name="score" value="0.057517064517803584" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255448 (text: There are two main options for performance in an O) is an LLM-generated sample, misclassified as human-generated with confidence 0.05751706&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255448 (text: There are two main options for performance in an O) is an LLM-generated sample, misclassified as human-generated with confidence 0.05751706
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i443]" time="0.069"><properties><property name="score" value="0.17111011047287766" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255449 (text: Listen, before you start writing to your local pap) is an LLM-generated sample, misclassified as human-generated with confidence 0.17111011&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255449 (text: Listen, before you start writing to your local pap) is an LLM-generated sample, misclassified as human-generated with confidence 0.17111011
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i444]" time="0.027"><properties><property name="score" value="0.12494709977662223" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255450 (text: http://youtu.be/Kp09yP36UBw  A postal worker who h) is an LLM-generated sample, misclassified as human-generated with confidence 0.1249471&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255450 (text: http://youtu.be/Kp09yP36UBw  A postal worker who h) is an LLM-generated sample, misclassified as human-generated with confidence 0.1249471
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i445]" time="0.026"><properties><property name="score" value="0.13975173015488976" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255451 (text: In a great feat of seduction, Bill Gothard had his) is an LLM-generated sample, misclassified as human-generated with confidence 0.13975173&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255451 (text: In a great feat of seduction, Bill Gothard had his) is an LLM-generated sample, misclassified as human-generated with confidence 0.13975173
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i446]" time="0.013"><properties><property name="score" value="0.01897629244955734" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255452 (text: Markets for running shoes are skewed, according to) is an LLM-generated sample, misclassified as human-generated with confidence 0.01897629&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255452 (text: Markets for running shoes are skewed, according to) is an LLM-generated sample, misclassified as human-generated with confidence 0.01897629
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i447]" time="0.063"><properties><property name="score" value="0.11942073494704279" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255453 (text: At least 40,000 people were evacuated on Sunday ni) is an LLM-generated sample, misclassified as human-generated with confidence 0.11942073&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255453 (text: At least 40,000 people were evacuated on Sunday ni) is an LLM-generated sample, misclassified as human-generated with confidence 0.11942073
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i448]" time="0.016"><properties><property name="score" value="0.014296851627165763" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255454 (text: Purchase Pincover College All-American Game Pincov) is an LLM-generated sample, misclassified as human-generated with confidence 0.01429685&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255454 (text: Purchase Pincover College All-American Game Pincov) is an LLM-generated sample, misclassified as human-generated with confidence 0.01429685
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i449]" time="0.065"><properties><property name="score" value="0.13301173062212743" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255455 (text: That would seem to contradict the S&amp;P Global Marke) is an LLM-generated sample, misclassified as human-generated with confidence 0.13301173&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255455 (text: That would seem to contradict the S&amp;P Global Marke) is an LLM-generated sample, misclassified as human-generated with confidence 0.13301173
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i450]" time="0.022"><properties><property name="score" value="0.09667350301464794" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255456 (text: BRASILIA (Reuters) - Brazil said Stockholm's non-b) is an LLM-generated sample, misclassified as human-generated with confidence 0.0966735&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255456 (text: BRASILIA (Reuters) - Brazil said Stockholm's non-b) is an LLM-generated sample, misclassified as human-generated with confidence 0.0966735
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i451]" time="0.066"><properties><property name="score" value="0.06950023173222405" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255457 (text: Decade 3: the contest between information and user) is an LLM-generated sample, misclassified as human-generated with confidence 0.06950023&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255457 (text: Decade 3: the contest between information and user) is an LLM-generated sample, misclassified as human-generated with confidence 0.06950023
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i452]" time="0.026"><properties><property name="score" value="0.1859078609403175" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255458 (text: Mark Composer  mother to two  years, THREE childre) is an LLM-generated sample, misclassified as human-generated with confidence 0.18590786&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255458 (text: Mark Composer  mother to two  years, THREE childre) is an LLM-generated sample, misclassified as human-generated with confidence 0.18590786
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i453]" time="0.027"><properties><property name="score" value="0.11840316442641573" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255459 (text: To help people trust your business are you posting) is an LLM-generated sample, misclassified as human-generated with confidence 0.11840316&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255459 (text: To help people trust your business are you posting) is an LLM-generated sample, misclassified as human-generated with confidence 0.11840316
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i454]" time="0.013"><properties><property name="score" value="0.09597431970172488" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255460 (text: Now that many people have read my Saints' testing ) is an LLM-generated sample, misclassified as human-generated with confidence 0.09597432&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255460 (text: Now that many people have read my Saints' testing ) is an LLM-generated sample, misclassified as human-generated with confidence 0.09597432
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i455]" time="0.064"><properties><property name="score" value="0.18193256241574732" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255461 (text: Benedict Cumberbatch plays James Bond in the film ) is an LLM-generated sample, misclassified as human-generated with confidence 0.18193256&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255461 (text: Benedict Cumberbatch plays James Bond in the film ) is an LLM-generated sample, misclassified as human-generated with confidence 0.18193256
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i456]" time="0.067"><properties><property name="score" value="0.1325928424712889" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255462 (text: Defection versus providing cover  Machiavelli dire) is an LLM-generated sample, misclassified as human-generated with confidence 0.13259284&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255462 (text: Defection versus providing cover  Machiavelli dire) is an LLM-generated sample, misclassified as human-generated with confidence 0.13259284
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i457]" time="0.015"><properties><property name="score" value="0.061102223352660356" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255463 (text: Ponnus  A wacky new platforming puzzle game with t) is an LLM-generated sample, misclassified as human-generated with confidence 0.06110222&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255463 (text: Ponnus  A wacky new platforming puzzle game with t) is an LLM-generated sample, misclassified as human-generated with confidence 0.06110222
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i458]" time="0.090"><properties><property name="score" value="0.1533148374228897" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255464 (text: A poll conducted by YouGov for the Guardian sugges) is an LLM-generated sample, misclassified as human-generated with confidence 0.15331484&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255464 (text: A poll conducted by YouGov for the Guardian sugges) is an LLM-generated sample, misclassified as human-generated with confidence 0.15331484
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i459]" time="0.079"><properties><property name="score" value="0.11784727863474849" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255465 (text: • Freedom estimated at $3.6bn • President also pre) is an LLM-generated sample, misclassified as human-generated with confidence 0.11784728&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255465 (text: • Freedom estimated at $3.6bn • President also pre) is an LLM-generated sample, misclassified as human-generated with confidence 0.11784728
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i460]" time="0.016"><properties><property name="score" value="0.03581294411772462" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255466 (text: Winning a gold medal is a huge accomplishment in o) is an LLM-generated sample, misclassified as human-generated with confidence 0.03581294&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255466 (text: Winning a gold medal is a huge accomplishment in o) is an LLM-generated sample, misclassified as human-generated with confidence 0.03581294
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i461]" time="0.062"><properties><property name="score" value="0.1383975159769983" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255467 (text: Survivors of the Syrian civil war have given their) is an LLM-generated sample, misclassified as human-generated with confidence 0.13839752&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255467 (text: Survivors of the Syrian civil war have given their) is an LLM-generated sample, misclassified as human-generated with confidence 0.13839752
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i462]" time="0.015"><properties><property name="score" value="0.13240564611228245" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255468 (text: England tighthead prop Owen Farrell is in the race) is an LLM-generated sample, misclassified as human-generated with confidence 0.13240565&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255468 (text: England tighthead prop Owen Farrell is in the race) is an LLM-generated sample, misclassified as human-generated with confidence 0.13240565
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i463]" time="0.090"><properties><property name="score" value="0.08330990588752321" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255469 (text: As you may have heard, the Clinton Foundation has ) is an LLM-generated sample, misclassified as human-generated with confidence 0.08330991&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255469 (text: As you may have heard, the Clinton Foundation has ) is an LLM-generated sample, misclassified as human-generated with confidence 0.08330991
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i464]" time="0.083"><properties><property name="score" value="0.1380719339494438" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255470 (text: Match date: 22 July 2016  In the end, qualificatio) is an LLM-generated sample, misclassified as human-generated with confidence 0.13807193&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255470 (text: Match date: 22 July 2016  In the end, qualificatio) is an LLM-generated sample, misclassified as human-generated with confidence 0.13807193
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i465]" time="0.025"><properties><property name="score" value="0.10322662785605674" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255471 (text: DISCORD WHO ARE YOU?  Who Am I?  Are You Still For) is an LLM-generated sample, misclassified as human-generated with confidence 0.10322663&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255471 (text: DISCORD WHO ARE YOU?  Who Am I?  Are You Still For) is an LLM-generated sample, misclassified as human-generated with confidence 0.10322663
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i466]" time="0.017"><properties><property name="score" value="0.09194553033609609" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255472 (text: Barry and Me: Sonia DDD Jacqiek Kalan, DP Kishan J) is an LLM-generated sample, misclassified as human-generated with confidence 0.09194553&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255472 (text: Barry and Me: Sonia DDD Jacqiek Kalan, DP Kishan J) is an LLM-generated sample, misclassified as human-generated with confidence 0.09194553
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i467]" time="0.022"><properties><property name="score" value="0.03436166311682198" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255473 (text: Stinking Skull  Statue  Grahka tribe Summoning Art) is an LLM-generated sample, misclassified as human-generated with confidence 0.03436166&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255473 (text: Stinking Skull  Statue  Grahka tribe Summoning Art) is an LLM-generated sample, misclassified as human-generated with confidence 0.03436166
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i468]" time="0.100"><properties><property name="score" value="0.07893534183199274" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255474 (text: With an unrestricted market to search for any perf) is an LLM-generated sample, misclassified as human-generated with confidence 0.07893534&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255474 (text: With an unrestricted market to search for any perf) is an LLM-generated sample, misclassified as human-generated with confidence 0.07893534
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i469]" time="0.103"><properties><property name="score" value="0.1335558406171356" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255475 (text: In this little-noticed chapter, Darmstadt, Germany) is an LLM-generated sample, misclassified as human-generated with confidence 0.13355584&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255475 (text: In this little-noticed chapter, Darmstadt, Germany) is an LLM-generated sample, misclassified as human-generated with confidence 0.13355584
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i470]" time="0.094"><properties><property name="score" value="0.1723891318602548" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255476 (text: photo: Shin Fujimoto  Rugby Magazine  Full credit ) is an LLM-generated sample, misclassified as human-generated with confidence 0.17238913&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255476 (text: photo: Shin Fujimoto  Rugby Magazine  Full credit ) is an LLM-generated sample, misclassified as human-generated with confidence 0.17238913
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i471]" time="0.075"><properties><property name="score" value="0.1398490902842092" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255477 (text: Mr Groom Key To Successful First Mission?  By Staf) is an LLM-generated sample, misclassified as human-generated with confidence 0.13984909&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255477 (text: Mr Groom Key To Successful First Mission?  By Staf) is an LLM-generated sample, misclassified as human-generated with confidence 0.13984909
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i472]" time="0.063"><properties><property name="score" value="0.17396520051718295" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255478 (text: I know what you're thinking – you've already finis) is an LLM-generated sample, misclassified as human-generated with confidence 0.1739652&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255478 (text: I know what you're thinking – you've already finis) is an LLM-generated sample, misclassified as human-generated with confidence 0.1739652
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i473]" time="0.028"><properties><property name="score" value="0.08228876917352532" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255479 (text: Giant Tax Credit, These Of You Think You're Small?) is an LLM-generated sample, misclassified as human-generated with confidence 0.08228877&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255479 (text: Giant Tax Credit, These Of You Think You're Small?) is an LLM-generated sample, misclassified as human-generated with confidence 0.08228877
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i474]" time="0.067"><properties><property name="score" value="0.10169636526907398" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255480 (text: (NaturalNews) Here's a boneheaded misconception th) is an LLM-generated sample, misclassified as human-generated with confidence 0.10169637&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255480 (text: (NaturalNews) Here's a boneheaded misconception th) is an LLM-generated sample, misclassified as human-generated with confidence 0.10169637
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i475]" time="0.014"><properties><property name="score" value="0.17577706196341691" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255481 (text: President Donald Trump may have been joking about ) is an LLM-generated sample, misclassified as human-generated with confidence 0.17577706&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255481 (text: President Donald Trump may have been joking about ) is an LLM-generated sample, misclassified as human-generated with confidence 0.17577706
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i476]" time="0.015"><properties><property name="score" value="0.25752627403360373" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i477]" time="0.015"><properties><property name="score" value="0.06158303506241247" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255483 (text: Description  The first track of the Deadshot EFX G) is an LLM-generated sample, misclassified as human-generated with confidence 0.06158304&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255483 (text: Description  The first track of the Deadshot EFX G) is an LLM-generated sample, misclassified as human-generated with confidence 0.06158304
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i478]" time="0.033"><properties><property name="score" value="0.0737723970080706" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255484 (text: Mas Themis Scholar Elemental School Public School ) is an LLM-generated sample, misclassified as human-generated with confidence 0.0737724&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255484 (text: Mas Themis Scholar Elemental School Public School ) is an LLM-generated sample, misclassified as human-generated with confidence 0.0737724
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i479]" time="0.084"><properties><property name="score" value="0.07379619942263989" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255485 (text: If you go looking for healthy foods at the grocery) is an LLM-generated sample, misclassified as human-generated with confidence 0.0737962&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255485 (text: If you go looking for healthy foods at the grocery) is an LLM-generated sample, misclassified as human-generated with confidence 0.0737962
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i480]" time="0.075"><properties><property name="score" value="0.10323482873920274" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255486 (text: New Mexico City (CNN) Rodríguez Arkuren acribed a ) is an LLM-generated sample, misclassified as human-generated with confidence 0.10323483&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255486 (text: New Mexico City (CNN) Rodríguez Arkuren acribed a ) is an LLM-generated sample, misclassified as human-generated with confidence 0.10323483
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i481]" time="0.075"><properties><property name="score" value="0.10945664822976442" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255487 (text: Celebrating their 25th anniversary in 2000, on Jul) is an LLM-generated sample, misclassified as human-generated with confidence 0.10945665&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255487 (text: Celebrating their 25th anniversary in 2000, on Jul) is an LLM-generated sample, misclassified as human-generated with confidence 0.10945665
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i482]" time="0.084"><properties><property name="score" value="0.170198398827165" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255488 (text: In the running for 2018 PROBLEM? Make a terrific m) is an LLM-generated sample, misclassified as human-generated with confidence 0.1701984&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255488 (text: In the running for 2018 PROBLEM? Make a terrific m) is an LLM-generated sample, misclassified as human-generated with confidence 0.1701984
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i483]" time="0.078"><properties><property name="score" value="0.1342040109932985" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255489 (text: Scoreboard CBS ABC FOX NBC UNI CW Adults 18-49: Ra) is an LLM-generated sample, misclassified as human-generated with confidence 0.13420401&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255489 (text: Scoreboard CBS ABC FOX NBC UNI CW Adults 18-49: Ra) is an LLM-generated sample, misclassified as human-generated with confidence 0.13420401
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i484]" time="0.015"><properties><property name="score" value="0.06534017665808523" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255490 (text: TRANSITIONS NOTICE  - April 22, 2018.   Link Trans) is an LLM-generated sample, misclassified as human-generated with confidence 0.06534018&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255490 (text: TRANSITIONS NOTICE  - April 22, 2018.   Link Trans) is an LLM-generated sample, misclassified as human-generated with confidence 0.06534018
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i485]" time="0.026"><properties><property name="score" value="0.1725242299325173" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255491 (text: TRACKER REPLAY: Vaymanu Ericsson's 1st goal vs. th) is an LLM-generated sample, misclassified as human-generated with confidence 0.17252423&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255491 (text: TRACKER REPLAY: Vaymanu Ericsson's 1st goal vs. th) is an LLM-generated sample, misclassified as human-generated with confidence 0.17252423
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i486]" time="0.023"><properties><property name="score" value="0.05863078130673438" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255492 (text: Worth noting: the lead up to the European election) is an LLM-generated sample, misclassified as human-generated with confidence 0.05863078&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255492 (text: Worth noting: the lead up to the European election) is an LLM-generated sample, misclassified as human-generated with confidence 0.05863078
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i487]" time="0.083"><properties><property name="score" value="0.15261483887461433" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255493 (text: After a few days of rumors going around, Lenovo ha) is an LLM-generated sample, misclassified as human-generated with confidence 0.15261484&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255493 (text: After a few days of rumors going around, Lenovo ha) is an LLM-generated sample, misclassified as human-generated with confidence 0.15261484
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i488]" time="0.072"><properties><property name="score" value="0.16949618078195366" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255494 (text: Capture The Flag is still one of my favorite strug) is an LLM-generated sample, misclassified as human-generated with confidence 0.16949618&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255494 (text: Capture The Flag is still one of my favorite strug) is an LLM-generated sample, misclassified as human-generated with confidence 0.16949618
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i489]" time="0.080"><properties><property name="score" value="0.08895916218769787" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255495 (text: So I was hoping to get some reactions from you guy) is an LLM-generated sample, misclassified as human-generated with confidence 0.08895916&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255495 (text: So I was hoping to get some reactions from you guy) is an LLM-generated sample, misclassified as human-generated with confidence 0.08895916
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i490]" time="0.089"><properties><property name="score" value="0.09732619366995113" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255496 (text: Since Republicans are not doing quite as well with) is an LLM-generated sample, misclassified as human-generated with confidence 0.09732619&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255496 (text: Since Republicans are not doing quite as well with) is an LLM-generated sample, misclassified as human-generated with confidence 0.09732619
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i491]" time="0.081"><properties><property name="score" value="0.13123064938330192" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255497 (text: ORLANDO, Fla. -- In just his fifth game with the C) is an LLM-generated sample, misclassified as human-generated with confidence 0.13123065&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255497 (text: ORLANDO, Fla. -- In just his fifth game with the C) is an LLM-generated sample, misclassified as human-generated with confidence 0.13123065
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i492]" time="0.020"><properties><property name="score" value="0.025364540987515838" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255498 (text: An AFA program for the prevention and treatment of) is an LLM-generated sample, misclassified as human-generated with confidence 0.02536454&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255498 (text: An AFA program for the prevention and treatment of) is an LLM-generated sample, misclassified as human-generated with confidence 0.02536454
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i493]" time="0.107"><properties><property name="score" value="0.13097199968860412" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255499 (text: I've got a bit of trouble keeping track at the mom) is an LLM-generated sample, misclassified as human-generated with confidence 0.130972&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255499 (text: I've got a bit of trouble keeping track at the mom) is an LLM-generated sample, misclassified as human-generated with confidence 0.130972
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i494]" time="0.098"><properties><property name="score" value="0.10110679839116742" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255500 (text: Link to this page: https://www.socialistparty.org.) is an LLM-generated sample, misclassified as human-generated with confidence 0.1011068&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255500 (text: Link to this page: https://www.socialistparty.org.) is an LLM-generated sample, misclassified as human-generated with confidence 0.1011068
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i495]" time="0.105"><properties><property name="score" value="0.07842747378045728" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255501 (text: Disney Infinity 3.0 UPDATES: &quot;We've made it so tha) is an LLM-generated sample, misclassified as human-generated with confidence 0.07842747&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255501 (text: Disney Infinity 3.0 UPDATES: "We've made it so tha) is an LLM-generated sample, misclassified as human-generated with confidence 0.07842747
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i496]" time="0.023"><properties><property name="score" value="0.0853403942583153" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255502 (text: At the border, a workman was knocked flat on his b) is an LLM-generated sample, misclassified as human-generated with confidence 0.08534039&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255502 (text: At the border, a workman was knocked flat on his b) is an LLM-generated sample, misclassified as human-generated with confidence 0.08534039
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i497]" time="0.103"><properties><property name="score" value="0.21339239909367264" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255503 (text: I join the countless others who have been surprise) is an LLM-generated sample, misclassified as human-generated with confidence 0.2133924&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255503 (text: I join the countless others who have been surprise) is an LLM-generated sample, misclassified as human-generated with confidence 0.2133924
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i498]" time="0.038"><properties><property name="score" value="0.08111963082108282" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255504 (text: The Washington Post  Monday, March 6, 2008; Page A) is an LLM-generated sample, misclassified as human-generated with confidence 0.08111963&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255504 (text: The Washington Post  Monday, March 6, 2008; Page A) is an LLM-generated sample, misclassified as human-generated with confidence 0.08111963
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt2_jsonl[i499]" time="0.117"><properties><property name="score" value="0.22566374811301457" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255505 (text: CoverStory  Everyone can do help and to help  Can ) is an LLM-generated sample, misclassified as human-generated with confidence 0.22566375&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255505 (text: CoverStory  Everyone can do help and to help  Can ) is an LLM-generated sample, misclassified as human-generated with confidence 0.22566375
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Glacier Ridge Christian School\n\nGlacier Ridge Christian School is a private Christian school in Johnstown, Ohio. It was founded in the fall of 1999 by Gary and Tammy Smith.\n\nThe school started with just 13 students in grades 5-8. By the end of the first year, the school had grown to 65 students in those same grades. As the school has continued to grow, so has the number of teachers.\n\nBy the year 2009, Glacier Ridge had expanded to offer a preschool through 12th grade education, and the school had grown to almost 200 students. In the fall of 2013, the enrollment was over 300 students.\n\nThe school started with the first graduating class of 2002. They are now on their 11th graduating class, and the school's graduates continue to receive acceptance to universities across the United States. The school offers several Advanced Placement classes.\n\nGlacier Ridge has had its students participate in various competitions and competitions. Most notable are the sports teams, especially their basketball teams, which have qualified for state competition multiple times. The school has also had various music groups participate in state and regional competitions, and in the 2010-2011 school year, the school sent students to participate in an International Children's Games competition in Romania.\n\nIn the 2010-2011 school year, a group of the school's teachers were sent to Haiti to help establish a new orphanage, &quot;Eli's Children's House&quot;, which was dedicated to Eli Pierre, who had died in a motorcycle accident. They were instrumental in bringing the orphanage into the community by helping establish new programs for the children, including English as a Second Language, Computer Lab, and Physical Education classes. The teachers assisted in the education of the orphans as well as working with the staff on the property.\n\nThe school has also taken various trips over the years, including trips to Virginia and New York City. The school also hosted an educational trip to Washington D.C. for one of their graduating classes.\n\nThe school has participated in various community events and service projects. One of the most notable is their annual &quot;Twilight Prom&quot; that is held each spring. Each year, the school hosts a prom for local high school students who have special needs. The students' dates are typically their own siblings.\n\nOn March 21, 2011, a fire destroyed the school's auditorium, causing over $1 million in damages. After the school received an outpouring of support, the school was able to begin construction on a new auditorium in time for the next school year.\n\nIn 2015, Glacier Ridge Christian School became part of the Christian School Association of Northeast Ohio.\n\nThe School is affiliated with the Association of Christian Schools International, the Ohio Christian Education Association, and the Christian School Association of Northeast Ohio.\n\n]" time="0.039"><properties><property name="score" value="0.014773298343345693" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[2003 NCAA Division I-AA football season\n\nThe 2003 NCAA Division I-AA football season, part of college football in the United States organized by the National Collegiate Athletic Association at the Division I Football Championship Subdivision level, began on September 6, 2003, and concluded with the National Championship Game of the NCAA Division I Football Championship on December 21, 2003, at Finley Stadium in Chattanooga, Tennessee. The Georgia Southern Eagles won their first I-AA championship, defeating the Furman Paladins by a final score of 27\u221210. \n\n&lt;nowiki&gt;*&lt;/nowiki&gt; &quot;Denotes host institution&quot;\n\n&lt;nowiki&gt;*&lt;/nowiki&gt; &quot;Denotes host institution&quot;\n\n&lt;nowiki&gt;*&lt;/nowiki&gt; &quot;Denotes host institution&quot;\n]" time="0.018"><properties><property name="score" value="0.0074247071560895606" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00742471&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00742471
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Adiadochokinesia\n\nAdiadochokinesia, or more fully dysdiadochokinesia, is a neurologic symptom, described as the inability to perform rapid alternating movements, as seen in ataxia.\n\nExamples of this are \n\nA classic test is &quot;the finger to nose test&quot;. The examiner repeatedly taps the patient's finger with the examiner's finger and asks the patient to reciprocate in the same rhythm.\n\nThis should be distinguished from dysdiadochokinesia, a progressive condition involving more or less the same symptoms.\n]" time="0.022"><properties><property name="score" value="0.017053084251544837" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01705308&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01705308
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Akyem Kotoku\n\nAkyem Kotoku is one of the five traditional districts of the Ashanti Empire. The people are collectively known as Akyem Kotokus.\n\nAkyem Kotoku lies in the Eastern region of Ashanti and is bounded by the East Mamprusi to the East, Nkwantanan to the North, Ashanti to the West and Ahanta West to the South. It is part of Ashanti Region. The area is mainly mountainous with some rich valleys and hills. Akyem Kotoku is known as the &quot;Area of Plenty&quot;.\n\nThe people are mainly farmers and fishermen. Its capital is Akyem Kotoku. It has two paramountcies, Buokrom and Kumasi. Buokrom, the newest paramountcy, was established in 1975. The Akyem Kotoku Traditional Council has nine chiefs.\n\nThe major towns are Abokobi, Domenase, Akropong, and Nsuaem. Other settlements include Aduanful, Boankraful, Denkyemerem, Anhwiaso, Gyankroma, Kwabenya, Nananom, Obofram, Okwensu, Ofoase, Nkoranza, Nsutam, Aduaso, Adom, Ofram, Odomankoma, Asafo, Anyigbe, Obofram, Kwabenya, Denkyemerem, Odumase, Nsutam, Aduaso, Nsuaem, Nsuaso, Tafo-Nsuaem, Manhyia, Boaso, Adom, Ofram, Anhwiaso, Adoase, Anyigbe, Boankraful, Okwensu, Ofoase, Nkoranza, Akropong, Gyankroma, Domenase, Egyir, Afram, Denkyerem, Nananom, Akuapem South, Asafo, Aboadze and Koma.\n\nThe Akyem Kotoku Traditional Council has nine paramountcies.\n\nThe Akyem Kotoku people have two coats of arms, one to use at home and the other for international use.\n\nThe arms used at home is a gold shield with a black anchor in the center. On the shield is a white dove with a golden olive branch in its beak. The dove is supported by two black wild dogs facing each other.\n\nThe shield is surmounted by a gold mural crown with five points and black stripes.\n\nThe shield is supported by two black lions standing on a white scroll. The scroll bears the motto: &quot;MUSTER FAITHFULNESS&quot; (Cape Coast Latin: &quot;Fidelitas Summa Est&quot;).\n\nThe arms used outside the country is a gold shield with a black ram's head in the center. On the shield is a green laurel wreath.\n\nThe shield is surmounted by a gold mural crown with five points and black stripes.\n\nThe shield is supported by two black and white wild dogs standing on a white scroll. The scroll bears the motto: &quot;MUSTER FAITHFULNESS&quot; (Cape Coast Latin: &quot;Fidelitas Summa Est&quot;).\n\nThe two dogs represent vigilance. The two lions in the home arms represent fortitude and strength.\n\n]" time="0.044"><properties><property name="score" value="0.005635020690857334" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00563502&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00563502
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Aulus Atilius Calatinus\n\nAulus Atilius Calatinus was consul of the Roman Republic in 221 BC with Gaius Claudius Nero. The primary action of the consulship was the siege and capture of the Volscian city of Suessa Pometia.\n\nIn the same year he was sent to Tarentum with a fleet of 10,000 infantry and 800 cavalry to aid the Tarentines in the Second War with the Bruttians. Calatinus was captured in a naval action off the coast near Tarentum, and the Bruttian admiral Lucerius was rewarded with his command for this victory. The Roman troops under his command, however, escaped to Thurii.\n\nThe earlier Aulus Atilius Calatinus was elected consul in 222 BC, and killed in a battle with the Samnites.\n\n]" time="0.021"><properties><property name="score" value="0.04568572379304858" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04568572&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04568572
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Frederic Myers\n\nFrederic William Henry Myers (15 January 1843 \u2013 18 January 1901) was a classical scholar, poet, playwright, novelist, and the founder of the Society for Psychical Research, which he led until his death.\n\nFrederic Myers was born at Kew, Surrey, the eldest son of Frederic Myers, a surgeon. He was educated at King's College School, and at St Paul's School in London. After a year at Trinity College, Cambridge (1861\u20132), he was a schoolmaster at Rugby from 1862 to 1871. In 1872 he became classical lecturer at University College, London. He became a member of the Council of the College (now part of UCL), and was appointed professor of Latin there in 1880. In 1876 Myers joined the Metaphysical Society, along with Alfred Tennyson, J. M. Barrie and others. His friendship with Barrie lasted for the remainder of his life, and Myers even accompanied Barrie and his wife on their honeymoon. Myers and Tennyson became close friends and met frequently in later years, as seen in their correspondence.\n\nMyers wrote prose and verse for periodicals, including a parody of Gerard Manley Hopkins, and some detective stories, and contributed to books of poetry and belles lettres, such as &quot;Whistler and His Art&quot; and the &quot;Yellow Book&quot;. His poetry was praised by Henry James, who]" time="0.028"><properties><property name="score" value="0.10303968461366497" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.10303968&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.10303968
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[His father was a professor and his mother was an economist. While growing up, he was exposed to the inspiring stories of his parents and their family members in China.\n\n\n\n\n\nKoh attended the University of Michigan where he received a bachelor of arts degree in economics, then went to the University of California, Berkeley where he earned a master of arts and doctor of philosophy degrees in economics.\n\n\n\n\n\nHe returned to Singapore in 1998 and joined the Singapore Civil Service.\n\n\n\n\n\nAs Deputy Secretary of the Ministry of Finance (2001-2007), he was involved in the design and implementation of the Singapore Savings Bond and the Central Provident Fund (CPF) investments.\n\n\n\n\n\nIn 2008, Koh was appointed as the Managing Director of the Monetary Authority of Singapore (MAS), where he oversees the development and implementation of macroeconomic policies for Singapore.\n\n\n\n\n\nHe is also the Chairman of the Government of Singapore Investment Corporation (GIC) which is responsible for the management of Singapore\u2019s foreign reserves and investment portfolio.]" time="0.024"><properties><property name="score" value="0.020143986940784897" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02014399&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02014399
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Little Tikes LTX Battery Operated Tricycle\n\nHe\u2019s been dreaming about his first bike for a while, and now he can hop on and ride around with friends and siblings alike on this bright red Little Tikes LTX Battery Operated Tricycle. It features an adjustable seat, a footrest and a chain guard so he\u2019ll stay safe while riding. Battery-operated, it requires 2 AA batteries (not included). This is a toy for children 3 years and older.]" time="0.023"><properties><property name="score" value="0.033790463213334565" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03379046&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03379046
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Trixie Textured - My cat licked my thumb.\n\n\n\nJoin Date: Nov 2006 Posts: 4,592\n\nI have not been to China, though I have been to Japan and love it.\n\n\n\nI have friends in China who have also been to Japan, and they both love it, and neither of them think of it as a &quot;poor&quot; country. In fact, a lot of people who have been there say that China is much nicer.\n\n\n\nChina is not a poor country. They have tons of money and resources. Their problem is that they do not spend enough on infrastructure and don't value education as highly as they could. So their people do not make as much money as they could. And they don't have nice things. It is much nicer than India, but not as good as Japan. I would like to visit it, because I do not think I would be disappointed, but I don't expect it to be better than Japan.\n\n\n\nThey have a different mindset than us, and do things differently. Their government is not a democracy. It is a communist government. I think that is not a bad thing, and has it's benefits. They have much fewer guns and gun crime than we do. They are, for the most part, more honest than we are. Their infrastructure, while not as nice as ours, is very good. They spend more on infrastructure than we do. They have much more public transportation, and better public transportation. They are working on their environmental problems, and they have made more progress than we have. They have fewer social problems, and less crime than we do. They have a very good education system, especially compared to India.\n\n\n\nThey spend more money on schools and education, and invest more in their future, than we do.\n\n\n\nChina is a country that is far more well developed than many people in the US would think.\n\n\n\nThe fact that they have fewer guns, and better education and fewer social problems is probably why they have fewer school shootings than we do. I think we have way too many guns in this country, and that is a big part of the problem.\n\n\n\nJust my thoughts.]" time="0.041"><properties><property name="score" value="0.018741751286313772" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[&quot;Say to them: 'As surely as I live, declares the Sovereign Lord, I take no pleasure in the death of the wicked, but rather that they turn from their ways and live. Turn! Turn from your evil ways! Why will you die, people of Israel?' declares the Sovereign Lord.&quot;\n\nEzekiel 33:11-11\n\nGod wants every sinner to be saved. He does not want people to perish but be saved. God hates sin, and sin destroys a person. But He loves people. He created them to have fellowship with Him and one day live in His Kingdom, and He does not want that for one person to be lost.\n\n\n\n\n\nGod's command to Ezekiel was that He was telling him to go and tell the people of Israel that they need to turn to God, because they were going to die if they did not. We know that Ezekiel told them the same thing, because God's word is truth. God's Word does not fail.\n\n\n\n\n\nWhat God is saying is that people are going to die if they do not turn to Him. People who reject God will die. The Bible says in Romans 1:32, &quot;Though they know God's righteous decree that those who do such things deserve death, they not only continue to do these very things but also approve of those who practice them.&quot; God says that those people who do these things deserve death, and yet people continue to do them and approve of them.\n\n\n\n\n\nGod will not tolerate the sinful acts of humanity, but He wants every person to be saved. There is only one way to be saved. The Bible says in John 14:6, &quot;Jesus answered, 'I am the way and the truth and the life. No one comes to the Father except through me.'\u201d\n\n\n\n\n\nNo one can be saved except through Jesus Christ, but when people put Jesus as their savior, they are saved. It's that simple.\n\n\n\n\n\nThere are many people who call themselves Christians who say that God wants everyone to be saved, and they quote Ezekiel 33:11 and John 3:16, but they don't realize that God's word is truth. God says in Revelation 21:8, &quot;But the cowardly, the unbelieving, the vile, the murderers, the sexually immoral, those who practice magic arts, the idolaters and all liars\u2014their place will be in the fiery lake of burning sulfur. This is the second death.&quot;\n\n\n\n\n\nThere is only one way to be saved, and it is not through anyone or anything but Jesus Christ. Jesus is the only way to heaven. God says that all the people who are going to perish are going to go to a place called the second death. That's hell.\n\n\n\n\n\nWhen people quote Ezekiel 33:11 and John 3:16, they don't realize that God's word is truth. They don't realize that when they say that God wants everyone to be saved, they are not telling the truth. When they say that God wants everyone to be saved, they are putting God in a box, because God's word is truth. God says that those who do not accept Jesus as their Savior are going to perish. God's word is truth, and people who do not put their faith in Jesus Christ are going to die.\n\n\n\n\n\nThe Bible says in John 14:6, &quot;Jesus answered, 'I am the way and the truth and the life. No one comes to the Father except through me.'\u201d\n\n\n\n\n\nJesus is the only way to the Father, and no one can be saved except through Him. He died on the cross for our sins, and He rose from the dead on the third day, and He sits at the right hand of God, interceding for us. It's that simple.\n\n\n\n\n\nLet's take a look at Ezekiel 33:11-11, and read it carefully.\n\n\n\n\n\n&quot;Say to them: 'As surely as I live, declares the Sovereign Lord, I take no pleasure in the death of the wicked, but rather that they turn from their ways and live. Turn! Turn from your evil ways! Why will you die, people of Israel?' declares the Sovereign Lord.&quot;\n\n\n\n\n\nGod says that He takes no pleasure in the death of the wicked. God does not like it when people die. But God hates sin, and sin leads to death. If people want to live, they must turn from their sins and put their faith in Jesus Christ.\n\n\n\n\n\nIf people don't put their faith in Jesus Christ, they are going to perish. God doesn't want that. He doesn't want people to perish, but to live.\n\n\n\n\n\nGod wants people to put their faith in Jesus Christ. If people put their faith in Jesus Christ, they are saved. If they don't put their faith in Jesus Christ, they are going to perish. That's the truth. That's the message of Ezekiel 33:11-11.\n\n\n\n\n\nPeople are going to die if they don't put their faith in Jesus Christ. People will perish. People will be lost. People will be doomed. People will go to hell.\n\n\n\n\n\nLet's read Ezekiel 33:11-12.\n\n\n\n\n\n&quot;For I take no pleasure in the death of anyone,&quot; declares the Sovereign Lord. &quot;Repent and live!&quot;\n\n\n\n\n\nGod takes no pleasure in the death of anyone. God does not want anyone to perish. He doesn't want anyone to die. He wants people to repent and live. God is saying that He is not going to stand back and watch as people perish. He's going to do something about it.\n\n\n\n\n\nGod says in Ezekiel 33:13-16, &quot;Turn! Turn from your evil ways! Why will you die, people of Israel? I have no pleasure in the death of the wicked, but rather that they turn from their ways and live. Do you think I take pleasure in the death of the wicked? declares the Sovereign Lord. Rather, am I not pleased when they turn from their ways and live? Don't presume that you can say to yourselves]" time="0.104"><properties><property name="score" value="0.04147289350145867" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Image caption Bailiffs removed protesters from Grangemouth plant in October 2013\n\nFive people have been charged in connection with an anti-fracking protest at an oil refinery in Grangemouth.\n\nThe protesters climbed on to the chimney at Ineos petrochemicals plant on 9 October.\n\nPolice Scotland said two men, aged 21 and 24, had been charged in connection with the protest and would appear at Falkirk Sheriff Court on Monday.\n\nThree women, aged 40, 44 and 48, were charged with breach of the peace and will also appear at court.\n\nThe 40-year-old was also charged with breach of the peace.\n\nThe other three were arrested in connection with a protest outside a hotel in Glasgow on 10 October.]" time="0.025"><properties><property name="score" value="0.03672980211585353" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0367298&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0367298
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Working with an example - rotating through an array\n\nFor a number of years, I've been fascinated by the idea of writing code that can rotate through an array. Let's say I have the following array of numbers:\n\nvar myNumbers = [ 1, 2, 3, 4, 5 ];\n\nThe following snippet of code would display the same numbers in reverse order:\n\nfor ( var i = myNumbers.length - 1; i &gt;= 0; i-- ) { console.log( myNumbers[i] ); }\n\nWhat's going on here?\n\nWe know that the index of an array can be used as a &quot;pointer&quot; to the location in memory that contains a particular item in an array. If the array myNumbers has five items, then myNumbers[0] , myNumbers[1] , myNumbers[2] , myNumbers[3] and myNumbers[4] will point to the values 1 , 2 , 3 , 4 and 5 respectively.\n\nIn the previous example, the code starts with the length of the array (5) and subtracts one from it. As we know that index 0 in an array contains the first item in the array, the previous code will execute the following steps:\n\nTake the length of the array (5) Subtract 1 (so we now have 4) Start at the index that contains the first item in the array (myNumbers[4] = 3) and run down the array until we reach the last item in the array (myNumbers[3] = 5)\n\nSo the above code will end up displaying the items in the array in reverse order, starting at index 4 and working backwards to index 0.\n\nAt this point, we have a good idea of how we can access the values in an array in a specific order, but how do we go the other way?\n\nIt would be great if we could simply use the same code as above, but add a &quot;--&quot; operator before the last number. That way, we could add a simple parameter to the code to control the range of numbers it will use.\n\nUnfortunately, we can't do that. While JavaScript will happily work with -- , it will also happily take -- in code and convert it to -1 , or the last number in an array. That's not going to do us any good, as we want the code to be flexible enough that we can work with a number that we specify as the range of numbers.\n\nSo let's look at a simple function that can give us the results we need.\n\nfunction rotate(numbers, direction, number) { numbers.reverse(); numbers.push(number); numbers.unshift(number); };\n\nThe rotate function above has three parameters, and will rotate the values in the numbers array.\n\nnumbers is the array we want to use.\n\ndirection is the parameter that allows us to choose whether we want the values to be rotated in the direction we specify. In this case, we've chosen direction to be either &quot;right&quot; or &quot;left&quot;\n\nnumber is the number we're using as the upper bound of the range that we want to use.\n\nLet's take a look at a simple example:\n\nvar numbers = [ 1, 2, 3, 4, 5 ]; rotate( numbers, &quot;right&quot; ); console.log( numbers ); //output: [ 1, 2, 3, 4, 5 ]\n\nAs we can see, the numbers array is simply rotated right by the number we specified. That's fine, but let's add in some code that we can use to control which numbers are displayed:\n\nvar numbers = [ 1, 2, 3, 4, 5 ]; rotate( numbers, &quot;right&quot;, 1 ); console.log( numbers ); //output: [ 1, 4, 3, 5, 2 ]\n\nWe can see that we've rotated the numbers so that the first number in the array is now number 1 , and the next number in the array is 4 .\n\nWe've also specified that we want to rotate in a right-handed manner, by passing &quot;right&quot; as the second parameter. This means that we have access to the following rotation values:\n\nrotate( numbers, &quot;left&quot; ); rotate( numbers, &quot;right&quot; ); rotate( numbers, &quot;right&quot; ); rotate( numbers, &quot;right&quot; ); rotate( numbers, &quot;left&quot; );\n\nIf we try to access the rotate function with &quot;left&quot; as the parameter, we'll see the following output:\n\nvar numbers = [ 1, 2, 3, 4, 5 ]; rotate( numbers, &quot;left&quot; ); console.log( numbers ); //output: [ 4, 3, 5, 2, 1 ]\n\nNot quite what we wanted, right?\n\nWhat we want is to be able to take an array, specify the number of numbers we want to display, and be able to choose the direction of rotation.\n\nWe also want to ensure that the values we are displaying are always within the array's bounds, and we're using them correctly.\n\nThe good news is that it's not too difficult to do this. We simply need to know what we want the code to look like, and write the code to do what we want it to do.\n\nLet's take a look at some simple code to display the numbers in an array in a given direction:\n\nfunction rotateArray( numbers, direction, startAt ) { numbers.reverse(); numbers.push( startAt ); numbers.unshift( startAt ); };\n\nrotateArray function\n\nThis is the code we'll use to display the numbers in the numbers array in the direction we specify. We'll start with a simple parameter:\n\nstartAt is the number we're displaying. It can either be the first number in the array, or any number within the array.\n\nLet's take a look at the code again:\n\nfunction rotateArray( numbers, direction, startAt ) { numbers.reverse(); numbers.push( startAt ); numbers.unshift( startAt ); };\n\nUsing this simple function, we can take any number of numbers, and display them in the direction we want. Let's look at a quick example:\n\nvar numbers = [ 1, 2, 3, 4, 5 ]; rotateArray( numbers, &quot;left&quot;, 3 ); console.log( numbers ); //output: [ 4, 3, 5, 2, 1 ]\n\nThis time, we've used &quot;left&quot; as the direction, and we've also used the parameter &quot;3&quot; as the startAt value. The result is that we now have the numbers displayed in the array starting at the number 3 .\n\nIf we use a number that isn't within the bounds of the array, we'll get the following result:\n\nvar numbers = [ 1, 2, 3, 4, 5 ]; rotateArray( numbers, &quot;right&quot;, 3 ); console.log( numbers ); //output: [ 5, 2, 4, 1, 3 ]\n\nAs you can see, the direction we've specified is &quot;right&quot;, and the value we've specified for startAt is 3 . Unfortunately, this value isn't within the bounds of the array, so the values in the array have been displayed starting from the last number in the array.\n\nLet's take a look at how we can use this code to work with a number that's outside the bounds of the array:\n\nvar numbers = [ 1, 2, 3, 4, 5 ]; rotateArray( numbers, &quot;right&quot;, 6 ); console.log( numbers ); //output: [ 5, 2, 4, 1, 3 ]\n\nThis time, we've specified &quot;right&quot; as the direction, and a value of 6 as the startAt parameter. The result is that we now have the values displayed in the array starting at the number 6 .\n\nThis time, the result we want is displayed in the array, but there's a problem. We don't want the number 6 to be the start of the array. Instead, we want it to be the last number in the array.\n\nThe good news is that we can easily handle this by modifying the code slightly:\n\nfunction rotateArray( numbers, direction, startAt ) { numbers.reverse(); numbers.push( startAt ); numbers.unshift( startAt ); };\n\nUsing the simple rotation code above, we can now take any array, specify the number we want to display, and choose the direction we want to use. This means we can write some simple code that can work with any number, regardless of its position in the array.\n\nThe next step is to write some code that can ensure the number we specify as the startAt is in the range that we expect.\n\nLet's take a look at how we can do that:\n\nfunction rotateArray( numbers, direction, startAt ) { if( startAt &gt;= numbers.length ) { throw new RangeError(&quot;Start at is outside of the]" time="0.112"><properties><property name="score" value="0.05699481052956177" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Contents\n\n1. The Meaning of Love\n\n2. Our Attitude Toward Others\n\n3. The Two Great Commandments\n\n4. Developing Love\n\n5. How to Help Others\n\n6. The Golden Rule\n\n7. The Commandments in the New Testament\n\n8. Tolerance\n\n\nCHAPTER 1\n\nThe Meaning of Love\n\nBefore we can love others, we must know the meaning of love. It is important to love God and obey Him and to have a right relationship with Him. But to have a right relationship with Him we must love Him. That means that we must have a good relationship with all His laws.\n\nOf course we can't keep all the commandments all the time. Yet we must have a good relationship with them. Otherwise we will not know what God is like.\n\nLove is based on God's law.\n\nRomans 13:10. The commandments, You shall not commit adultery, You shall not steal, You shall not bear false witness, You shall not covet; and any other commandment, are summed up in this word, You shall love your neighbor as yourself.\n\nGalatians 5:14. For all the law is fulfilled in one word, even in this: You shall love your neighbor as yourself.\n\nLove is based on God's law. We should study God's commandments and know the laws He has given us. Then we will know how to have a good relationship with Him.\n\nOne reason we are to love God and obey Him is that it is important to have a good relationship with God. We need to know how much God loves us. Without knowing this we cannot have a right relationship with God. We need to have a good relationship with Him so we can be happy and serve Him better.\n\nLet us understand what love is.\n\n1 John 4:8,9. He who does not love does not know God, for God is love. In this the love of God was manifested toward us, that God has sent His only begotten Son into the world, that we might live through Him.\n\nIn these verses John explained that love is the opposite of lawlessness. Love is the opposite of murder, hatred, malice, evil, adultery, stealing, lying, and a lot of other bad things.\n\nRomans 13:8-10. Owe no man anything, but to love one another: for he that loves another has fulfilled the law. For this, You shall not commit adultery, You shall not kill, You shall not steal, You shall not covet, and if there be any other commandment, it is briefly comprehended in this saying, namely, You shall love your neighbor as yourself. Love does no harm to a neighbor; therefore love is the fulfillment of the law.\n\nWe learn from this that the opposite of love is lawlessness. The opposite of love is murder, hatred, envy, malice, evil, and all the other things mentioned in these verses.\n\nVerse 10 says, &quot;Love does no harm to a neighbor.&quot; In other words, love is a good thing. It is a good thing to have a good relationship with God.\n\nRomans 13:8 says, &quot;Love does no harm to a neighbor.&quot; Love helps and blesses others.\n\nLove is a good thing. The reason love is a good thing is because God is love.\n\n1 John 4:16. And we have known and believed the love that God has for us. God is love; and he who abides in love abides in God, and God in him.\n\nGod is love.\n\n1 John 4:7-11. Beloved, let us love one another: for love is of God; and every one that loves is born of God, and knows God. He that loves not knows not God; for God is love. In this was manifested the love of God toward us, because that God sent his only begotten Son into the world, that we might live through him. Herein is love, not that we loved God, but that he loved us, and sent his Son to be the propitiation for our sins. Beloved, if God so loved us, we ought also to love one another. No man has seen God at any time. The only begotten Son, which is in the bosom of the Father, he has declared him.\n\nVerse 8 says, &quot;God is love.&quot; God's name is Love.\n\nGod's name is Love.\n\nJohn 4:8. \u2026 the well was deep; and I said, &quot;How shall I get down?&quot;\n\nAnd he said, &quot;Come, and I will show you.&quot;\n\n9. And he said, &quot;Draw you [a] little water out of the well, and \u2026 drink.&quot;\n\n10. And I said, &quot;Sir, give me [a] little to drink.&quot;\n\n11. And he said, &quot;Drink, my [b] son, and \u2026 your eyes shall be opened.&quot;\n\n12. And I drank, and it was as if I had had new vision given to me; for I could see [c] through those walls of water.\n\n13. And he said to me, &quot;Go [a] down, for the water is [b] yet a little way below the earth. \u2026\n\n15. \u2026 This water is [c] that which I said \u2026 will make you come up hither.&quot;\n\n16. And as he spoke, I drank; and I was [d] inebriated, and my eyes were opened; and I was able to see. \u2026\n\n19. \u2026 &quot;Come,&quot; he said, &quot;follow me \u2026 and I will give you \u2026 rest.&quot;\n\nGod is the God of rest.\n\n1 Thessalonians 4:9. \u2026 God has not appointed us to wrath, but to obtain salvation by our Lord Jesus Christ.\n\nGod is our salvation. God is our rest. God is our peace.\n\nGod is our rest.\n\n1 Peter 2:6. \u2026 you may be \u2026 a kind of first-fruits of his creatures.\n\nGod is our first fruits.\n\n2 Timothy 2:13. If we are faithless, he remains faithful; for he cannot deny himself.\n\nGod is faithful.\n\nGod is the God of rest.\n\nGod is our salvation.\n\nGod is our peace.\n\nGod is our rest.\n\nHebrews 4:10-13. For he who has entered His rest has himself also ceased from his own works, as God did from His. Let us therefore be diligent to enter that rest \u2026\n\n12. \u2026 For the word of God is living and active, and sharper than any two-edged sword, and piercing even to the division of soul and spirit, and of joints and marrow, and is a discerner of the thoughts and intents of the heart.\n\n13. \u2026 neither is there any creature that is not manifest in his sight; but all things are naked and opened to the eyes of him with whom we have to do.\n\nGod is a discerner of the thoughts and intents of the heart.\n\nGod sees everything.\n\nGod sees everything.\n\n1 John 4:8,9. He who does not love does not know God, for God is love. In this the love of God was manifested toward us, that God has sent His only begotten Son into the world, that we might live through Him.\n\nGod is love.\n\n1 John 3:16,18. Hereby perceive we the love of God, because he laid down his life for us: and we ought to lay down our lives for the brethren.\n\n19. But whoso has this world's goods, and sees his brother in need, and shuts up his heart from him, how does the love of God abide in him?\n\nGod is love.\n\nGod is our rest.\n\nGod is our salvation.\n\nGod is our peace.\n\nGod is a discerner of the thoughts and intents of the heart.\n\nGod is love.\n\nGod is our rest.\n\nGod is our salvation.\n\nGod is our peace.\n\nGod is a discerner of the thoughts and intents of the heart.\n\nThe Bible says that God is love.\n\nWe can study love.\n\n1 John 4:7-11. Beloved, let us love one another: for love is of God; and every one that loves is born of God, and knows God. He that loves not knows not God; for God is love. In this was manifested the love of God toward us, because that God sent his only begotten Son into the world, that we might live through him. Herein is love, not that we loved God, but that he loved us, and sent his Son to be the propitiation for our sins. Beloved, if God so loved us, we ought also to love one another. No man has seen God at any time. The only begotten Son, which is in the bosom of the Father, he has declared him.\n\nVerse 11 says, &quot;Herein is love, not that we loved God, but that he loved us.&quot; The opposite of love is lawlessness. But the opposite of love is not lawlessness. The opposite of love is law.\n\nLoving God and obeying Him is loving Him.\n\nThe Bible says that God is love.\n\nLoving God and obeying Him is loving Him.\n\nLoving God and obeying Him is loving God.\n\nJohn 14:21. He that has my commandments,]" time="0.104"><properties><property name="score" value="0.0034145100392555874" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Interviews\n\nInterview with Ariel Efraim (Soilwork)\n\nInterview with Steven Rosen (August Burns Red)\n\nInterview with Brent Hinds (Mastodon)\n\nInterview with Paul Kuhr (I See Stars)\n\nInterview with Christian Antidormi (Hatebreed)\n\nInterview with Jack Syperek (Structures)\n\nInterview with Ben Gordon (The Treatment)\n\nInterview with Doc Coyle (God Forbid)\n\nInterview with Chris Barnes (Six Feet Under)\n\nInterview with Mike D (The Sword)\n\nInterview with Russ Rankin (The Slackers)\n\nInterview with Claudio Sanchez (Coheed and Cambria)\n\nInterview with Udo Dirkschneider (Accept)\n\nInterview with Ben Weinman (Dillinger Escape Plan)\n\nInterview with Matt Pike (High On Fire)\n\nInterview with Joe Duplantier (Gojira)\n\nInterview with Johannes Eckerstr\xf6m (Evergrey)\n\nInterview with Pete Loose (Reel Big Fish)\n\nInterview with Pete Wentz (Fall Out Boy)\n\nInterview with Daniel Svensson (Raubtier)\n\nInterview with Chino Moreno (Deftones)\n\nInterview with Pierre Bouvier (Simple Plan)\n\nInterview with Jon Egan (The Ghost Of A Thousand)\n\nInterview with Mike D (Lack Of Afro)\n\nInterview with Michael Crafter (Sevendust)\n\nInterview with Charlie Benante (Anthrax)\n\nInterview with Rob Caggiano (Volbeat)\n\nInterview with Ronnie Atkins (Pretty Maids)\n\nInterview with Gee Anzalone (Exhorder)\n\nInterview with Micky Shirley (Five Finger Death Punch)\n\nInterview with Oderus Urungus (GWAR)\n\nInterview with Ryan Downey (Texas Hippie Coalition)\n\nInterview with Chris Chasse (Unearth)\n\nInterview with Lee Altus (Suffocation)\n\nInterview with Roy Mayorga (Soulfly)\n\nInterview with Gus G (Firewind)\n\nInterview with Paul Allender (In Flames)\n\nInterview with Francesco Artusato (All Shall Perish)\n\nInterview with Rody Walker (Pyogenesis)\n\nInterview with Mike Dean (Slipknot)\n\nInterview with Angela Gossow (Arch Enemy)\n\nInterview with Trevor Strnad (The Black Dahlia Murder)\n\nInterview with Mike Peters (The Alarm)\n\nInterview with Jonas Ekdahl (Freak Kitchen)\n\nInterview with Dave Ellefson (Megadeth)\n\nInterview with Corey Taylor (Slipknot)\n\nInterview with Keith Buckley (Every Time I Die)\n\nInterview with Jack Frost (Kill Devil Hill)\n\nInterview with Josh Rand (Temple of Brutality)\n\nInterview with Michael Amott (Arch Enemy)\n\nInterview with Scott Gorham (Thin Lizzy)\n\nInterview with Matt Heafy (Trivium)\n\nInterview with Joe Kresge (Unearth)\n\nInterview with Randy Blythe (Lamb of God)\n\nInterview with Benji Webbe (Dimmu Borgir)\n\nInterview with Max Cavalera (Soulfly)\n\nInterview with John Bush (Armored Saint)\n\nInterview with Davey Muise (Exhorder)\n\nInterview with Stephen Carpenter (Deftones)\n\nInterview with Dani Filth (Cradle Of Filth)\n\nInterview with Leif Cuzner (Darkane)\n\nInterview with George \u201cCorspegrinder\u201d Fisher (Cannibal Corpse)\n\nInterview with Steve Asheim (Hatebreed)\n\nInterview with Bobby Blitz Ellsworth (Overkill)\n\nInterview with Adam Dutkiewicz (Killswitch Engage)\n\nInterview with Mick Murphy (Exhumed)\n\nInterview with Doug Pinnick (King\u2019s X)\n\nInterview with Alexi Laiho (Children Of Bodom)\n\nInterview with M. Shadows (Avenged Sevenfold)\n\nInterview with Alex Holzwarth (Rhapsody Of Fire)\n\nInterview with Carlos Cruz (Soulfly)\n\nInterview with Glenn Hughes (Deep Purple/Black Country Communion)\n\nInterview with Oliver Palotai (Kamelot)\n\nInterview with Meegs Rascon (The Mars Volta)\n\nInterview with Joakim Brod\xe9n (Sabaton)\n\nInterview with Tim \u201cRipper\u201d Owens (Iced Earth/Yngwie Malmsteen)\n\nInterview with Matt Heafy (Trivium)\n\nInterview with Warrel Dane (Nevermore)\n\nInterview with Zach Myers (Shinedown)\n\nInterview with Michael Poulsen (Volbeat)\n\nInterview with Michael Amott (Arch Enemy)\n\nInterview with Mat Sinner (Primal Fear)\n\nInterview with Bj\xf6rn \u201cSpeed\u201d Strid (Soilwork)\n\nInterview with Matt Heafy (Trivium)\n\nInterview with Rob Flynn (Machine Head)\n\nInterview with Serj Tankian (System Of A Down)\n\nInterview with Ville Valo (HIM)\n\nInterview with Jimmy DeGrasso (Megadeth)\n\nInterview with Joey Jordison (Slipknot)\n\nInterview with Chris Jericho (Fozzy/Fozzy Osbourne)\n\nInterview with Jon Donais (Shadows Fall)\n\nInterview with Jason Hook (Five Finger Death Punch)\n\nInterview with Steve Mazur (Grave Digger)\n\nInterview with Bj\xf6rn \u201cSpeed\u201d Strid (Soilwork)\n\nInterview with Sonny Sandoval (P.O.D.)\n\nInterview with Serj Tankian (System Of A Down)\n\nInterview with Fredrik Thordendal (Meshuggah)\n\nInterview with Wes Borland (Limp Bizkit)\n\nInterview with Alexi Laiho (Children Of Bodom)\n\nInterview with Blaze Bayley (Iron Maiden)\n\nInterview with Mark Hunter (Chimaira)\n\nInterview with Max Cavalera (Soulfly/Sepultura)\n\nInterview with Dave Lombardo (Slayer)\n\nInterview with Bruce Dickinson (Iron Maiden)\n\nInterview with Morgan Lander (Fear Factory)\n\nInterview with Oliver Palotai (Kamelot)\n\nInterview with James Hetfield (Metallica)\n\nInterview with Al Jourgensen (Ministry)\n\nInterview with Corey Taylor (Stone Sour/Slipknot)\n\nInterview with Alex Skolnick (Testament)\n\nInterview with Charlie Benante (Anthrax)\n\nInterview with Chris Reifert (Autopsy)\n\nInterview with Lajon Witherspoon (Sevendust)\n\nInterview with Jason McMaster (Grave Robber)\n\nInterview with Jon Howard (Light The Torch)\n\nInterview with Neil Fallon (Clutch)\n\nInterview with Dino Cazares (Fear Factory)\n\nInterview with Oderus Urungus (GWAR)\n\nInterview with Rex Brown (Pantera)\n\nInterview with Dave Lombardo (Slayer)\n\nInterview with Serj Tankian (System Of A Down)\n\nInterview with John Moyer (Disturbed)\n\nInterview with Mike Portnoy (Adrenaline Mob)\n\nInterview with King Diamond\n\nInterview with Trevor Strnad (The Black Dahlia Murder)\n\nInterview with Mikael \xc5kerfeldt (Opeth)\n\nInterview with Scott Ian (Anthrax)\n\nInterview with Paul Gray (Slipknot)\n\nInterview with Casey Chaos (Amen)\n\nInterview with Opeth\u2019s Mikael \xc5kerfeldt\n\nInterview with Mark Morton (Lamb Of God)\n\nInterview with Shawn Drover (Megadeth)\n\nInterview with Chino Moreno (Deftones)\n\nInterview with Serj Tankian (System Of A Down)\n\nInterview with Bruce Corbitt (Corrosion Of Conformity)\n\nInterview with Ray Alder (Fates Warning)\n\nInterview with Rob \u201cBlasko\u201d Nicholson (Ozzy Osbourne)\n\nInterview with Tom Araya (Slayer)\n\nInterview with Max Cavalera (Soulfly)\n\nInterview with Ryan Martinie (Cynic)\n\nInterview with Max Cavalera (Soulfly)\n\nInterview with Jason Newsted (Voivod)\n\nInterview with Mick Murphy (Exhumed)\n\nInterview with Lars Frederiksen (Rancid)\n\nInterview with Mike Muir (Suicidal Tendencies)\n\nInterview with Dave Mustaine (Megadeth)\n\nInterview with Neil Fallon (Clutch)\n\nInterview with Bruce Dickinson (Iron Maiden)\n\nInterview with Maynard James Keenan (Tool)\n\nInterview with Mike \u201cRat\u201d Ratner (S.O.D.)\n\nInterview with Max Cavalera (Soulfly)\n\nInterview with Tom Araya (Slayer)\n\nInterview with Chuck Billy (Testament)\n\nInterview with Tom Araya (Slayer)\n\nInterview with Phil Anselmo (Down)\n\nInterview with King Diamond\n\nInterview with Corey Taylor (Stone Sour/Slipknot)\n\nInterview with Dave Mustaine (Megadeth)\n\nInterview with Lars Ulrich (Metallica)\n\nInterview with Jerry Cantrell]" time="0.188"><properties><property name="score" value="0.06871329760346359" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0687133&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0687133
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Apple\u2019s ResearchKit is off to a good start, but there\u2019s plenty of work to be done.\n\n\n\nResearchKit was launched in March as a way for Apple\u2019s developers to easily tap into medical data from the iPhone to help researchers track and collect information about a variety of health conditions. Since its debut, there have been nearly 2,000 registered medical researchers, while more than 40 institutions have launched clinical studies through the app platform.\n\n\n\nBut there are several hurdles for Apple and its partners. On Wednesday, Dr. Harvey Fineberg, president of the Institute of Medicine, will join at a panel discussion at Mount Sinai Hospital in New York to examine what works and what\u2019s missing in the fast-evolving field of mobile health.\n\n\n\nThe session, titled \u201cWhat Are the Future Directions for ResearchKit?\u201d comes after a first wave of successful projects on ResearchKit, including apps that help people manage diabetes, heart disease and asthma.\n\n\n\nThere are also some examples of the problems Apple is facing as it tries to get more partners to adopt the tool. Earlier this month, a major player in the diabetes community said it would not be taking part in the app\u2019s data collection. The Juvenile Diabetes Research Foundation said in a statement that its members were not interested in the iPhone platform and \u201cwould prefer to conduct research studies using our own patient data that is gathered on an Android platform.\u201d\n\n\n\nThat\u2019s likely an expensive prospect for the organization, given the popularity of iPhones in the United States. As of last year, Android devices accounted for just over half of all U.S. smartphone subscribers, while iOS devices made up around 41 percent. That\u2019s according to the Pew Research Center, which estimated last year that there were 121 million iPhone users in the United States.\n\n\n\nThe success of Apple\u2019s platform will hinge in part on how it handles data privacy and security concerns. Apple had to clear some hurdles in the early days of ResearchKit because many people were concerned about privacy when they heard it would rely on data from Apple\u2019s servers. Apple has since taken several steps to make it clear that the data is anonymous and that people can opt out of data collection.\n\n\n\nThe app program could also be a major part of the iPhone\u2019s future if it takes off. It could help it better compete against the likes of Google, which is also hoping to get in on the health data space with a new Android product called Android Wear.\n\n\n\nThe first public version of ResearchKit will be unveiled on Wednesday, the panel will also include representatives from Mount Sinai and Pfizer.\n\n\n\nThat event will be at 11:30 a.m. EDT in New York City, and it will be livestreamed on Fortune\u2019s website.]" time="0.034"><properties><property name="score" value="0.03745691979306596" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03745692&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03745692
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[\n\nSend this thread via email by filling out the fields below\n\n\n\nThread: Lost 10lbs in the last 3 weeks Send To:\n\nUse semi-colon to separate multiple addresses Your Name: Your Email: Add a personal message: Send this thread via email by filling out the fields below\n\n\n\n]" time="0.017"><properties><property name="score" value="0.0030622203062435904" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00306222&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00306222
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[I don't think it is your attitude. I think it is just the way you are.\n\nIt isn't a criticism, just a statement of fact. It's like saying I don't think you are stupid.\n\nJust that you are naive, sweet, guileless, open. You have so much going for you, you don't have to be cynical.\n\nBut it isn't fair to the rest of us, not fair to expect us to feel your innocence and not to respond to it.\n\nYou need to toughen up, otherwise you will just be hurt. But don't ever be cynical.]" time="0.017"><properties><property name="score" value="0.04762038779169697" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04762039&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04762039
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Get the biggest daily news stories by email Subscribe Thank you for subscribing We have more newsletters Show me See our privacy notice Could not subscribe, try again later Invalid Email\n\nAs the Queen's cousin, Lady Amelia Windsor is more used to socialising with the royal family than you or I.\n\nBut even for her, it seems the sight of seeing Kate Middleton and Prince William in attendance at Pippa Middleton's wedding - as well as the newlyweds' own big day - was enough to make her feel &quot;uncomfortable&quot;.\n\nSpeaking at the wedding of one of her friend's brothers in Kent, the 28-year-old spoke about how it made her feel to have them there, reports the Daily Mail .\n\nShe said: &quot;Well I think it is a beautiful wedding, obviously, because it's Pippa and James.\n\n&quot;I've never been to a wedding before, so I don't know really what it should be like. But I do find it quite uncomfortable seeing them.&quot;\n\nThe Royal family came together to celebrate the wedding of Kate's sister to hedge fund manager James Matthews in the Berkshire village of Englefield.\n\n(Image: Getty Images Europe)\n\n(Image: PA)\n\n(Image: PA)\n\nAnd while Pippa and James' wedding was a private affair with just 200 guests, they were still surrounded by some royals as well as other notable names.\n\nLady Amelia's father is the Queen's first cousin, so it was no surprise to see the Queen herself attend the wedding, alongside Prince Harry and Duchess Meghan.\n\nThe Princesses Beatrice and Eugenie were also there, as well as Kate's sister, Pippa's brother James Middleton, David and Victoria Beckham and Sir Elton John.\n\nLady Amelia revealed that she even had a few words with Harry and Meghan, with the former asking her how she felt about her cousin's big day.\n\nShe said: &quot;I had a little conversation with Harry and Meghan, and they said how lovely the wedding was. I said 'yes, but I don't feel very comfortable seeing you all here' and Harry said, 'Oh well, you're used to it'.\n\n(Image: PA)\n\n(Image: PA)\n\n&quot;I said, 'Yes, but it is still very strange to see you here'. I suppose you can't blame me, I am a bit naive. But the whole thing was really lovely.&quot;\n\nShe added: &quot;I mean, James is my cousin, I grew up with him and I love him dearly, so it was really lovely to see him. It was a lovely wedding, they are such a lovely couple. I am so pleased they are so happy.&quot;\n\n(Image: Getty Images Europe)\n\n(Image: Getty Images Europe)\n\nAnd as for what she thought of her father's outfit - which caused much comment among fans - she admitted she wasn't sure.\n\nShe said: &quot;Well, he always wears an odd choice of clothes, but that is what he likes. It wasn't bad, I mean he didn't look like a big prawn.&quot;\n\nShe added: &quot;It was lovely to see him happy.&quot;]" time="0.034"><properties><property name="score" value="0.08175610893541" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08175611&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08175611
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Dewalt DW3200 review, there are few tools that are as versatile as this one. You may have a thousand bucks of cordless tools, but there\u2019s nothing that will do as much as this one will do. As you might have noticed, the construction of this unit is incredibly sturdy, so that you don\u2019t have to worry about it breaking when you use it. Also, it is a bit heavy, and it will take a while to get used to using it, especially if you have not used any other heavy drills.\n\nPower\n\nWe also tested the power of this Dewalt DW3200 and we must say that it is very impressive. It delivers 3,300 RPM, which is more than enough to bore holes and drill in all kinds of materials. It can easily penetrate into wood, brick, or concrete and it can bore holes up to two inches. The trigger lock mechanism also works well, which makes this drill safer to use.\n\nCordless or corded?\n\nIt is important to remember that this drill can work either as a corded or a cordless unit, and this is very important for you to know. This Dewalt DW3200 uses the 20 volt lithium-ion batteries and they can last longer than you might expect. They are also very light and durable, so you will be able to use them for a long time. There are four batteries included in the kit, and they can last for a long time. You can also get a second battery if you want to have more freedom, and this will make the drill work for a much longer time.\n\nExtra Features\n\nYou may also like the extra features included in this Dewalt DW3200. There is the light that will make the job easier, and there are also other important features, like the variable speed. This is also a very durable drill, so it will last for a very long time. Also, the build quality is fantastic, so you don\u2019t have to worry about it breaking easily.\n\nOverall\n\nIf you have a lot of different drilling needs, this is the drill for you. This drill is perfect for the casual user, but it can also be used by professionals. It is also very durable, so you can use it for a very long time, and you will get the most for your money. You should also remember that this drill is very powerful, and it can work for a long time, without breaking down.\n\nIf you want to see what more we have, you can visit Amazon.com and check all of the available drills. You can also check our article on the best reciprocating saw, to see what the best product in this category is. Also, don\u2019t forget to like and share this article, if you liked it.]" time="0.034"><properties><property name="score" value="0.028007214728601425" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02800721&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02800721
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Source: Food\n\nAs I\u2019m sure you\u2019re well aware by now, it\u2019s officially National Grilled Cheese Sandwich Month. With that in mind, this week\u2019s Food Week: Hors D\u2019Oeuvre features a variety of hors d\u2019oeuvres you can make in your own home. After all, a good grilled cheese is just like a fine wine: it gets better with age.\n\nContinue reading]" time="0.019"><properties><property name="score" value="0.046697083768793" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04669708&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04669708
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Jack and Justin are two of the hottest and most popular guys on the block. They make the perfect couple and know exactly what each other likes. But when they meet Tyler, things get complicated. This sizzling trio enjoys their time together and both Tyler and Jack crave more. But Justin is not so easily convinced. How will this all work out?\n\nWarning: This story contains three scorching hot guys and lots of hot and hard action.\n\n\n\n\xa92015 Scarlet Rose (P)2017 Scarlet Rose]" time="0.019"><properties><property name="score" value="0.05529173267715959" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05529173&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05529173
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[LAS VEGAS, NV - JANUARY 07: Recording artist Demi Lovato performs onstage during the 2017 DIRECTV NOW Super Saturday Night Concert at Park Theater at Park MGM on January 7, 2017 in Las Vegas, Nevada. (Photo by Kevin Winter/Getty Images for DIRECTV) (Photo: Kevin Winter, 2017 Getty Images)\n\nLAS VEGAS - Former Disney star Demi Lovato will release a new single from her sixth studio album on June 15th.\n\nThat's according to a video of Lovato, 26, posted on the YouTube channel of Las Vegas radio station KVEG.\n\nThe new album is the follow-up to 2015's &quot;Confident,&quot; which features the top 10 hits &quot;Cool for the Summer&quot; and &quot;Confident.&quot;\n\nMore: Everything you need to know about this year's 2017 Route 91 Harvest festival\n\nMore: Demi Lovato set for national tour with DJ Khaled\n\nMore: Demi Lovato re-engages with ex-boyfriend Wilmer Valderrama, so could there be more?\n\nThe pop singer also revealed she is launching a partnership with American Airlines to celebrate Hispanic Heritage Month, and she plans to go on a national tour with DJ Khaled.\n\nDemi Lovato to Release New Single, Announce Album Title on June 15th https://t.co/OBkKbXWySZ \u2014 Pop Crave (@PopCrave) June 1, 2017\n\nRead or Share this story: https://usat.ly/2rVY4mA]" time="0.018"><properties><property name="score" value="0.17867831682152335" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.17867832&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.17867832
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[RACISM IN BRITAIN HAS BECOME SO entrenched that it is becoming a bigger problem than it was at the time of the Birmingham pub bombings, when the threat of violent nationalism in the late 1970s prompted the first official inquiry into racial attacks, according to a report published today by the Equality and Human Rights Commission (EHRC).\n\nThe report, We\u2019re still struggling: the legacy of the Stephen Lawrence inquiry, says that despite years of measures to increase the representation of ethnic minorities in public life and protect them from discrimination, they are still \u201con the receiving end of racism\u201d, with black people twice as likely to suffer police stop and search tactics as whites.\n\nFigures from the Crown Prosecution Service (CPS) show that more than 40 per cent of cases involving charges of racially aggravated harassment are dropped, with charges reduced to more minor ones, or not pursued. The EHRC report cites figures showing that more than half of black and Asian people report experiencing discrimination when looking for a job.\n\nIn October, the Telegraph reported how]" time="0.015"><properties><property name="score" value="0.0647860861509475" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06478609&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06478609
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[# This Source Code Form is subject to the terms of the Mozilla Public\n\n# License, v. 2.0. If a copy of the MPL was not distributed with this\n\n# file, You can obtain one at http://mozilla.org/MPL/2.0/.\n\n# The following comment must be at the start of a new line.\n\n#\n\n#\n\n# The contents of this file are subject to the Mozilla Public\n\n# License Version 1.1 (the &quot;License&quot;); you may not use this file\n\n# except in compliance with the License. You may obtain a copy of\n\n# the License at http://www.mozilla.org/MPL/\n\n# Software distributed under the License is distributed on an &quot;AS\n\n# IS&quot; basis, WITHOUT WARRANTY OF ANY KIND, either express or implied.\n\n# See the License for the specific language governing rights and\n\n# limitations under the License.\n\n# The Original Code is Mozilla Communicator client code, released\n\n# March 31, 1998.\n\n# The Initial Developer of the Original Code is\n\n# Netscape Communications Corporation.\n\n# Portions created by the Initial Developer are Copyright (C) 1998\n\n# the Initial Developer. All Rights Reserved.\n\n# Contributor(s):\n\n#\n\n# Alternatively, the contents of this file may be used under the terms of\n\n# either of the GNU General Public License Version 2 or later (the &quot;GPL&quot;),\n\n# or the GNU Lesser General Public License Version 2.1 or later (the &quot;LGPL&quot;),\n\n# in which case the provisions of the GPL or the LGPL are applicable instead\n\n# of those above. If you wish to allow use of your version of this file only\n\n# under the terms of either the GPL or the LGPL, and not to allow others to\n\n# use your version of this file under the terms of the MPL, indicate your\n\n# decision by deleting the provisions above and replace them with the notice\n\n# and other provisions required by the GPL or the LGPL. If you do not delete\n\n# the provisions above, a recipient may use your version of this file under\n\n# the terms of any one of the MPL, the GPL or the LGPL.\n\n#\n\n# ***** END LICENSE BLOCK *****\n\nMOCHITEST_MANIFESTS += [\n\n' xpcshell.ini ' ,\n\n' mochitest-glue.ini ' ,\n\n' mochitest-plain.ini ' ,\n\n]\n\nDIST_DIRS += [\n\n' external/ ' ,\n\n' ipc/ ' ,\n\n' modules/ ' ,\n\n]\n\nEXTRA_COMPONENTS += [\n\n' mozLoop.js ' ,\n\n' webextensions.js ' ,\n\n]\n\ndeps = [\n\n'../base ' ,\n\n'../browser/base ' ,\n\n'../browser/components/migration ' ,\n\n'../browser/components/places ' ,\n\n'../browser/themes/shared ' ,\n\n'../content/base ' ,\n\n'../content/html ' ,\n\n'../content/xul ' ,\n\n'../embedding/browser ' ,\n\n'../embedding/components/browser ' ,\n\n'../embedding/components/toolkit ' ,\n\n'../extensions/browser/api ' ,\n\n'../extensions/common/moz.build ' ,\n\n'../gfx/2d ' ,\n\n'../gfx/layers ' ,\n\n'../intl ' ,\n\n'../netwerk ' ,\n\n'../printing ' ,\n\n'../services/sync ' ,\n\n'../toolkit/components/global ' ,\n\n'../toolkit/components/printing ' ,\n\n'../toolkit/components/satchel ' ,\n\n'../toolkit/crashreporter ' ,\n\n'../toolkit/library ' ,\n\n'../toolkit/mozapps/extensions ' ,\n\n'../toolkit/mozapps/extensions/extensions-types.js ' ,\n\n'../xpcom/base ' ,\n\n'../xpcom/build ' ,\n\n'../xpcom/build/shared-nspr ' ,\n\n'../xpcom/build/shared-nspr/src ' ,\n\n'../xpcom/ds ' ,\n\n'../xpcom/ds/io ' ,\n\n'../xpcom/ds/xpcom ' ,\n\n'../xpcom/glue ' ,\n\n'../xpcom/io ' ,\n\n'../xpcom/threads ' ,\n\n'../xpcom/threads/src ' ,\n\n'../xpcom/threads/xpcom ' ,\n\n' ../xpfe/appshell/nsAppShell.cpp ' ,\n\n' ../xpfe/appshell/nsAppShell.h ' ,\n\n' ../xpfe/appshell/nsAppsShellService.cpp ' ,\n\n' ../xpfe/appshell/nsAppsShellService.h ' ,\n\n' ../xpfe/appshell/nsNativeAppSupportWin.cpp ' ,\n\n' ../xpfe/appshell/nsNativeAppSupportWin.h ' ,\n\n' ../xpfe/appshell/nsNativeAppSupportXPConnect.cpp ' ,\n\n' ../xpfe/appshell/nsNativeAppSupportXPConnect.h ' ,\n\n' ../xpfe/appshell/nsWindowRoot.cpp ' ,\n\n' ../xpfe/appshell/nsWindowRoot.h ' ,\n\n' ../xpfe/appshell/nsXULAppAPI.h ' ,]" time="0.079"><properties><property name="score" value="0.017332589536445474" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The world of professional wrestling is always changing, but the superstars themselves can sometimes get lost in the shuffle. Sometimes, the changes within the industry are so drastic that a superstar's popularity can get left in the past. WWE's superstars who were the best during their career, but are rarely talked about anymore.\n\nWWE is still the leader of the wrestling world. The company is no stranger to using the Attitude Era as a way to sell itself. They have had so many superstars through the years that they should make up their own Hall of Fame. Some of those superstars are the ones who are mentioned the most.\n\nThe casual fan knows about Hulk Hogan, Stone Cold, and The Rock. That is what makes this so fun. Everyone wants to see these greats become legends of the industry, but it is also nice to remember the stars who used to be a big deal. Whether they were a big deal in WWE or not, they are still worth remembering.\n\nContinue scrolling to keep reading Click the button below to start this article in quick view Share Tweet Email Copy Link Copied\n\n10 Sin Cara\n\nvia trbimg.com\n\nSin Cara has become a joke of a superstar. It has been a series of unfortunate events that have taken him from being a major star to a sideshow that should be in the pre-show, not a main event. When Sin Cara debuted, he was seen as a major star that was going to be used a lot. It took some time, but it finally happened.\n\nThe problem was that he never delivered.\n\nHe was so unimpressive that he was booted from WWE in 2016. He is making an appearance at WrestleMania 33. For his sake, it would be best if he kept quiet and didn't interact with the crowd.\n\n9 Kelly Kelly\n\nvia tumblr.com\n\nKelly Kelly was a fan favorite during her time in WWE. She made her debut as a popular model who was wrestling on ECW. From there, she rose to the top as a valuable part of the roster. Kelly became a popular figure who had her own dedicated fans. She had the right amount of talent to be a decent worker, but it was her good looks that gave her the extra push.\n\nAfter leaving WWE in 2012, Kelly kept wrestling. She is back on the roster with TNA, but she has not appeared on a television episode of Impact Wrestling since December of 2015.\n\n8 Alberto Del Rio\n\nvia wrestlingrumors.net\n\nIt is a sad truth that Alberto Del Rio has fallen from his former position in the industry. He was a very big part of WWE. He was not only the United States Champion, but also a World Champion. Alberto had a unique presence as a superstar that could do things no one else could do. He also had the microphone skills to back up his moves.\n\nHis tenure in WWE was marred with incidents that made it clear he was not going to be there much longer. His time in Lucha Underground was more memorable, but it was a step down from what he used to be.\n\n7 The Great Khali\n\nvia cagesideseats.com\n\nThe Great Khali is not really a superstar that anyone really misses. His presence was laughable at times. The thing about Khali is that he had a lot of potential. He was very big and very strong, but he never seemed to have any clue about how to use those traits to his advantage.\n\nKhali is now an actor and a former wrestler. He is part of the reality show Khali Kisses, which involves him doing a lot of kisses on people. It is a far cry from his WWE days.\n\n6 Kelly Kelly's Boyfriend\n\nvia youtube.com\n\nWrestlers who are lucky enough to find a wife have a much easier time in the wrestling industry. The same could not be said for Mike Knox. He was only a superstar for a very short time, and he was a part of one of the most bizarre moments in wrestling history when he got the Attitude Adjustment from Kane. Knox is not talked about much these days, but there was a time when people were curious about him.\n\nHe married the popular WWE superstar Kelly Kelly. The couple got married in 2014. It was a marriage that was not meant to last, as the couple divorced in 2016.\n\n5 Santino Marella\n\nvia ringsidenews.com\n\nSantino Marella will forever be a part of WWE. The WWE Hall of Fame is home to some great wrestlers and some very bad wrestlers. Santino is among those who have made an impression for one reason or another. His career with the company was not as successful as he had hoped, but he made a huge impact. He was the guy who was hilarious because he was so bad.\n\nSantino's last match was in 2014. He has done some backstage work, but the star has pretty much disappeared from WWE.\n\n4 Tyson Kidd\n\nvia heavy.com\n\nTyson Kidd was a true performer in WWE. He had a unique look, and he could really do it all. The man was skilled on the mic, in the ring, and he had a good personality. Everything came together for Tyson, but that would all change during an unfortunate injury. Tyson was a mid-carder who could have risen to a top level, but that was not meant to be.\n\nKidd is now recovering from his neck injury. He is still hoping to get into the ring again, but it is looking unlikely.\n\n3 Christian\n\nvia bodybuilding.com\n\nChristian is not a name that is being mentioned much these days. It is a shame, as he is one of the best performers of his time. Christian has made a career of being known for his impressive in-ring work and his ability to entertain fans. Christian was a popular superstar during his time in WWE and he did some good things for TNA.\n\nChristian has not been in the ring since December of 2015. He is trying to get back into the ring, but there is a possibility that his neck injury may have ended his career.\n\n2 Dolph Ziggler\n\nvia wrestlingnewssource.com\n\nDolph Ziggler had the chance to become a very big star in WWE. Unfortunately, things have not been going well for Ziggler. He was once a main eventer, but he has fallen to the bottom of the card. He is one of the many guys who will be pushed as the underdog who is going to have the moment of their life.\n\nZiggler has not been the focus of WWE. It is likely that he will be entering a new chapter of his career in the near future. The most likely place would be in Ring of Honor, but WWE has shown a lot of interest in Ziggler.\n\n1 Wade Barrett\n\nvia wrestlingrumors.net\n\nIt is going to be very hard for the casual fan to remember Wade Barrett. He had a short run in WWE, but he did his part as a mid-carder. Barrett is a former Intercontinental Champion. He did well in WWE as a member of The Nexus, but his star power never quite took off.\n\nBarrett has been out of WWE since 2014. He has had some ups and downs in the independent circuit, but his career has largely gone downhill. It is possible that he could get another chance with WWE. That would probably be the only way that fans would get to see him again.\n\nBarrett has been trying to make a name for himself in the business world. He launched a website last year. It is not likely that he will become a success, but it is nice to see him try to make something of his life.]" time="0.101"><properties><property name="score" value="0.08685589632241927" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0868559&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0868559
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[University: The primary mission of the University is to provide education. For that purpose, the University has established different faculties and departments, of which the oldest and most important is the faculty of theology. The latter is a privileged entity, by virtue of its academic reputation, both in the local and the global sphere. The mission of the University is to educate the professors and the students in the science of theology, which means, it also needs to equip them with the specific professional skills, which make them able to exercise the profession of the ministry.]" time="0.015"><properties><property name="score" value="0.031019691841016606" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The brother-in-law of the Palestinian-American teenager who was filmed being brutally beaten by Israeli police at a West Bank checkpoint says the incident \u201chas put a wedge\u201d between the Palestinians and Israelis.\n\nYoussef Almogy said he has had more than 50 phone calls from news media outlets following the brutal arrest of 16-year-old Tariq Abu Khdeir, the Associated Press reported.\n\n\u201cWhen I saw it, I cried,\u201d Almogy said. \u201cI was so shocked, but after that, I started thinking about my brother, who was jailed for 11 years, and then released for nothing. And I said, \u2018If they can do that to him, then imagine what they can do to us.\u2019\u201d\n\n\u201cAnd this has put a wedge between Palestinians and Israelis,\u201d he added.\n\nAlmogy was referring to the arrest of his brother, Amer, who was jailed for 11 years after he and a group of Palestinians tried to rob a Tel Aviv bank in 1994. Almogy said Amer was never convicted of a crime and that he believes the former prisoner is being held because of his brother\u2019s connection to the teen, the AP said.\n\n&quot;We're used to seeing this sort of thing,&quot; he said.\n\nAbu Khdeir, who was born in Florida but living with his family in East Jerusalem, was arrested Thursday after he was identified as being involved in a clash between]" time="0.016"><properties><property name="score" value="0.05845482952228642" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05845483&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05845483
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[\n\nToday we\u2019re excited to announce that Bethesda Game Studios has joined Oculus as a developer for the Oculus Rift. The partnership will see the creation of two exclusive, made-for-VR games from Bethesda Game Studios; the first is currently being designed for virtual reality and the second is being built for traditional monitors.\n\n\n\n\u201cAt Oculus, we\u2019re dedicated to helping developers and partners to grow, and the support and dedication we\u2019ve received from Bethesda Game Studios has been outstanding,\u201d said Brendan Iribe, CEO of Oculus. \u201cJohn Carmack, founder of id Software and now CTO at Oculus, believes in our vision and has created a custom version of Doom 3 BFG Edition for Oculus. We look forward to continuing to work with Bethesda as we develop our first generation of virtual reality games and experiences.\u201d\n\n\n\n\u201cI have a lot of personal interest in making sure that the VR gaming space is successful because I really believe in the technology and its potential to change gaming,\u201d said Carmack. \u201cAs a developer, the Rift is something that I\u2019ve wanted to work with for a long time, but the existing technology wasn\u2019t quite there yet. I\u2019m happy to be able to help to make the Rift a success and contribute to the rebirth of VR gaming.\u201d\n\n\n\nAt Bethesda Game Studios, developers are working on a number of unannounced projects including new entries in major franchises. Those who have seen this year\u2019s E3 show floor, Gamescom or Quakecon presentations have seen teases of what the studio is up to and a few of the titles we have in development are also being designed for virtual reality. The two games announced today are entirely new IP and we can\u2019t wait to show them to you when they\u2019re ready for primetime.\n\n\n\nFor those of you unfamiliar with Bethesda Game Studios, the acclaimed developer is responsible for creating many of the industry\u2019s most popular games including the \u201cElder Scrolls\u201d series, the single-player add-on to the original \u201cDoom\u201d and the \u201cFallout\u201d series. In fact, \u201cFallout 3,\u201d \u201cFallout: New Vegas\u201d and \u201cThe Elder Scrolls V: Skyrim\u201d were each awarded game of the year at the 2008, 2010 and 2011 BAFTA awards, respectively.\n\n\n\n\u201cWe have made some long-term technology investments with VR in mind,\u201d said Todd Howard, Game Director and Executive Producer at Bethesda Game Studios. \u201cHardware is finally catching up to the visions developers have had for virtual reality. Now when you put on a VR headset, you\u2019re transported to new and amazing worlds.\u201d\n\n\n\nThis year, Bethesda Game Studios showcased \u201cDoom 3 BFG Edition\u201d for Oculus. \u201cDoom 3 BFG Edition\u201d is a complete overhaul of \u201cDoom 3\u201d that brings modern technology and a classic, storied franchise into the world of virtual reality. This is the first title designed for virtual reality.\n\n\n\n\u201cJohn Carmack\u2019s work in the early 90s helped usher in the era of modern 3D gaming, and his genius continues to inspire developers to innovate,\u201d said Brendan Iribe, CEO of Oculus. \u201cThe team at Bethesda Game Studios is made up of the most talented developers in the industry and we can\u2019t wait to bring their newest games, made for virtual reality, to life.\u201d\n\n\n\nOver the past 20 years, Bethesda Game Studios has earned its reputation as one of the industry\u2019s most respected and accomplished game development studios. Creators of the 2006 and 2008 Game of the Year, \u201cThe Elder Scrolls IV: Oblivion\u201d and 2010 Game of the Year, \u201cThe Elder Scrolls V: Skyrim,\u201d Bethesda Game Studios has earned a reputation for excellence and quality with each of its five releases.\n\n\n\nThis partnership is another step toward realizing the future of virtual reality. It\u2019s an exciting time to be a gamer with the release of these groundbreaking VR titles, the arrival of next generation consoles and the continued evolution of the PC.\n\n\n\nStay tuned for more news on Bethesda Game Studios\u2019 titles for Oculus.\n\n\n\n\u2013 The Oculus Team]" time="0.101"><properties><property name="score" value="0.04395603902457275" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04395604&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04395604
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[\n\nHi~ B een a long time no see. Thank you for the continued support on the blog. So, how have you all been doing?\n\nToday, I have a special post. I know most of you do not know but it is the 'World Animal Day'. The day is celebrated worldwide on October 4 to create awareness of animal rights and welfare. The day is a very special one for me because I am a vegetarian, animal lover and a dog-person.\n\n\n\n\n\nSo for the last one year, I have been trying to spread the awareness among the people. I have been trying to encourage people to stop using animals as fur coats, cosmetics and for food. I want people to know the feelings of these animals, the pains they suffer and what they go through while they are alive. Also, the environmental damage they cause to the planet. It is very important to spread this message because animals are treated as if they are some sort of toys that we can play with.\n\n\n\n\n\nToday, I would like to share with you some tips on how we can help the animals and make the planet a better place to live. I also have a small personal pledge to make. Also, if you have not signed the pledge yet, you can also make your own pledge to protect the animals and the planet.\n\n\n\n\n\n\n\n\n\nTips and pledges to help the animals\n\n(made by my friend!)\n\nOne way we can make a change is by making an effort to not use cosmetics that are tested on animals. We can choose to use all cruelty free products and make a conscious effort not to use anything that has been tested on animals.\n\nWays to reduce our carbon footprints and help protect the environment:\n\nAvoid eating meat, or even better, go vegetarian (dairy, eggs are not vegan but I still encourage you to stop using dairy products as well)\n\nUse eco-friendly products - Not just cosmetics but also in our everyday lives.\n\nGo cruelty free - Choose cruelty free products and wear clothes that have not been made out of leather, fur or wool.\n\nStop buying those animals that we do not need. Also, choose to buy second hand or recycled clothes. Avoid buying pets if you cannot provide the proper care for them.\n\nAdopt, don't buy! You can adopt a dog or cat from an animal shelter. Do not buy animals from pet stores or breeders. They are sold to you as &quot;breed-free&quot; animals but most of them have been bred in horrible conditions. Also, make sure you know the sex of the animal you are getting, because they usually are sold as 'un-neutered' or 'un-spayed' animals.\n\nMake a conscious effort to not use plastic bags. Use reusable bags or better still, try to avoid using bags at all.\n\nTry to use products that do not harm the environment and are biodegradable.\n\nWays to help the animals:\n\nSign the pledge - To show that you support the cause to protect the animals.\n\nShare the post - Let others know of the cruelty that the animals go through. Let them know how they can help the animals too.\n\nJoin or start a protest or campaign to help the animals.\n\nThe small pledge I am making is to help the animals. Every month, I will be donating some money to an animal shelter in my town to help feed the dogs and cats there.\n\nThese are just some of the ways we can help the animals and the planet. We can all make a change, and make this planet a better place to live.\n\n\n\n\n\nIf you want to know how to become a vegan, and which cosmetics are tested on animals, you can read this post: How to be a vegan?\n\nRemember, we are what we eat, so be what you want to be. Be a good person, who loves animals and helps to save the planet.\n\nWith lots of love,\n\nPriya]" time="0.083"><properties><property name="score" value="0.03752992206107605" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03752992&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03752992
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Top Customer Reviews\n\nWhere reviews refer to foods or cosmetic products, results may vary from person to person. Customer reviews are independent and do not represent the views of The Hut Group.\n\nUnique, gritty, and wonderful The Good: The worlds are some of the best I've seen in a while, and how you explore the city and interact with its citizens gives a great sense of authenticity and intrigue. As a huge comic fan, the references are abound. The story and narrative is one of the best I've ever experienced. The Bad: The game is very linear and the combat while satisfying can get old at times. Overall: This is one of the best games I've ever played, period. Everything about it is gritty, dark, atmospheric, and enticing. While you have a very linear path to follow in the main story, the side quests and interactions you can have with the citizens is one of the best I've seen. You are truly the villain in this game, so if you want to be good and helpful to everyone, this is not for you. But for the fans of dark, gritty stories, this is one of the best. The story and narrative are top notch. The villains are some of the most compelling I've ever seen in a video game, with even the main villains being so much more compelling than the good guys. I loved this game from start to finish, and can't wait to see where the series goes from here. The ending while great is somewhat confusing, so if you aren't]" time="0.017"><properties><property name="score" value="0.01790114614976246" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01790115&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01790115
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Mrs. Mahilani Konczey\n\n&quot;The tutors at HW have made a huge difference in my child's academic performance, as well as in her confidence. The tutors are friendly, professional and punctual. My child enjoys her sessions and she has learned a lot from them. HW has a good location and easy parking. I highly recommend them.&quot;]" time="0.014"><properties><property name="score" value="0.01995624976446897" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01995625&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01995625
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Flavor Pairs\n\n(Home)\n\nPickled Meat\n\nBacon, Flank Steak, Pork Butt, Bacon, Hams, Sausage, Canned Ham\n\nFor Cooking/Grilling\n\nPeppers, Onions, and other Vegetables\n\nPotatoes, Onion, Green Peppers, Jalapenos, Garlic, and other Vegetables\n\nPastas\n\nPasta, Macaroni, Shells, Ramen Noodles, Ziti, Gnocchi, Fusilli, and other Pastas\n\nSoups and Stews\n\nSoups, Stews, Chili, and other One-Pot Meals\n\nCorn, Peas, and other Grains\n\nCorn, Peas, Risotto, Polenta, Grits, Quinoa, Rice, and other Grains\n\nBeverages\n\nBeer, Alcohol, Soda, Juices, Tea, Coffee, Soda, and other Beverages]" time="0.015"><properties><property name="score" value="0.08689211956261489" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08689212&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08689212
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[And now, ladies and gentlemen, I give you the breathless high-speed ranting I did when I saw the (potentially) wackadoo conclusion to this storyline.\n\nOh, my God, you guys, this episode has a freaking atomic bomb on the side of it.\n\nOkay, cool. It\u2019s the mid-season finale. I know there\u2019s going to be some serious plotlines.\n\nAnd seriously, what was up with the weird science in this episode?\n\nNo. No, I did not just watch an episode with the headline: \u201cOuch, My Butt!\u201d\n\nThat plotline was weird as hell.\n\nBut I can understand why they did it. He\u2019s trying to be a \u201csophisticated ladies man\u201d and not the boy they\u2019ve been dealing with all these years.\n\nOh, now this is hilarious. But it\u2019s also dangerous, if you\u2019ve watched any superhero movie ever.\n\nIn the brief moment that I was allowed to breathe, I actually enjoyed this part.\n\nNot only is this scientifically improbable, but also\u2026 how does a security camera shoot a laser?\n\nI know this is only in fiction, but I do believe that they\u2019re committing suicide.\n\nThis is also part of the reason why I wanted to talk about this episode. I know it\u2019s not a good look. It\u2019s not exactly smart of him to do this.\n\nI don\u2019t even know what to say here.\n\nWHY?\n\nHe\u2019s not afraid of bugs, but he\u2019s scared of that snake?\n\nI mean, yeah, I get that he\u2019s doing it for a different reason, but\u2026\n\nIf it weren\u2019t for that, then I would not have laughed so hard at this.\n\nOkay, look, I was done with this season for a while. I thought that it was becoming a boring clich\xe9d plot that they were sticking to, so I was ready to say goodbye to it. But then I watched this, and I just felt like I was watching a new show. There was so much going on, and they were throwing so much at us at once, and it was just such a good use of time. I\u2019m really happy that they had this show, because it was seriously pretty funny and it was also action-packed. There was a lot going on, but it was very clear, and it was fun to watch. I just loved how much was going on. I loved it, and I really did enjoy it. It was a bit much, and they took a lot of crazy chances, but it was just so fun to watch. That was some crazy stuff, and it was nice to see something so good after so much of the regular garbage we were getting. Overall, I loved it.]" time="0.026"><properties><property name="score" value="0.09325548480920283" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.09325548&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.09325548
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Specification:\n\nProduct Name Bags Model SA01A Material Canvas + Zipper Size 29*15.5*10.5cm Color Red, Green, Black, Blue, Gray, Purple, Brown, Khaki, Yellow, Orange Weight 10g Quantity 1 Piece(s) Pattern Solid Inner Pocket Yes Closure Drawstring\n\nPackage include:\n\n1 X Canvas Bags\n\n\n\n\n\nDescription:\n\nDimensions of the storage bag, cool way to store the supplies for travel and daily life.\n\nLightweight, durable and sturdy canvas construction.\n\nA simple pouch, plenty of storage space to put many small items.\n\nThere is a handle for easy carrying and a good choice for daily life.\n\nUsed for carrying, practical, suitable for outdoor, camping, traveling, adventure, daily life, perfect for storing jewelry, makeup, keys, etc.\n\n\n\n\n\nSize: 29*15.5*10.5cm\n\n\n\nWeight: 10g\n\nMaterial: Canvas + Zipper\n\n\n\nColor: Red, Green, Black, Blue, Gray, Purple, Brown, Khaki, Yellow, Orange\n\nPattern: Solid\n\nInner Pocket: Yes\n\n\n\nClosure: Drawstring\n\n\n\n\n\n\n\nPackage include:\n\n1 X Canvas Bags]" time="0.014"><properties><property name="score" value="0.04924566354524984" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04924566&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04924566
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Author:\n\nFollowing a stint as a park ranger in Wyoming, Dan Beal relocated to Seattle and received a degree in science writing from the University of Washington.\n\nFollowing a stint as a park ranger in Wyoming, Dan Beal relocated to Seattle and received a degree in science writing from the University of Washington.]" time="0.013"><properties><property name="score" value="0.04858133161535494" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[(Adapted from the 2014 Urban Tree Canopy Conference)\n\nPotholes are problematic for all of us, and for our city\u2019s trees they represent a real danger. Roots can become entangled in the small cracks between the asphalt and the cement to create the impression that the tree is growing out of the road! And when the roots do become entwined, the pressure can be so great that the roots are destroyed. We all know the danger this creates for the trees.\n\nThe solution to this problem is obvious: pave over the whole road surface. Not so fast! In the City of Chicago, for example, there is a practice that preserves the utility of the tree\u2019s roots and of the road surface itself. The practice is called \u201cmilling\u201d. Milling involves cutting the asphalt out in a shape that will not disturb the roots and adding back in a new surface. While the tree is protected, the roots are not disturbed and the road surface is left intact. When a tree is scheduled for removal, this is often the method of choice. Milling also involves improving the root zone around the tree. The roots of many trees can\u2019t penetrate the compacted soil that is found beneath pavement, so it is necessary to mill the area to promote root growth.\n\nMilling is less expensive than paving over a road surface and can also be less expensive than removal of a tree. Milling should be done whenever possible and certainly before a tree is scheduled for removal. There is no need to sacrifice a mature, healthy tree.\n\nAdvertisements]" time="0.013"><properties><property name="score" value="0.036745342370636536" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03674534&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03674534
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[How do i search for files with a specific suffix using Unix?\n\nOn Unix, to search for files that ends with a particular file extension, use the following command:\n\nfind . -name '*.html'\n\nOR use grep command to list all files in a directory and subdirectories that have the suffix:\n\nfind . -type f | xargs grep -l &lt;suffix&gt;\n\nSample outputs:\n\n./Ajax/contact-us/index.html ./test/index.html\n\nThe above commands are equivalent to the grep command on Linux/MacOSX or the find command on Linux/MacOSX:\n\n$ find . -name &quot;*.html&quot; ./test/index.html ./Ajax/contact-us/index.html $ grep -l -Ri -f -E &quot;.*\\.html&quot; ./ test/index.html Ajax/contact-us/index.html\n\nFind out all files with a specific filename (using shell pattern match):\n\nfind . -name '*.html'\n\nOR use the egrep command to list all files in a directory and subdirectories that have the suffix:\n\nfind . -type f | xargs egrep -l &lt;suffix&gt;\n\nSample outputs:\n\n./Ajax/contact-us/index.html ./test/index.html\n\nOR\n\n$ find . -name &quot;*.html&quot; ./test/index.html ./Ajax/contact-us/index.html $ egrep -i &quot;.*\\.html&quot; . Ajax/contact-us/index.html test/index.html\n\nFind files in a directory and subdirectories:\n\nfind . -name &quot;*.html&quot;\n\nOR\n\nfind . -type f | xargs grep -l &lt;suffix&gt;\n\nSample outputs:\n\n./Ajax/contact-us/index.html ./test/index.html\n\nOR\n\n$ find . -name &quot;*.html&quot; ./test/index.html ./Ajax/contact-us/index.html $ egrep -i &quot;.*\\.html&quot; . Ajax/contact-us/index.html test/index.html\n\nFind files in a directory and all]" time="0.013"><properties><property name="score" value="0.22947472662214752" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Russia has expanded its ban on U.S. pork to include all American meat imports as the diplomatic row over alleged election meddling and the situation in Syria rumbles on.\n\nLast week, the Russian government added five new countries to the blacklist: South Korea, Montenegro, Iceland, Liechtenstein and Ukraine. While the list included major agricultural producers such as Brazil and Canada, U.S. meat producers were noticeably absent. However, on Wednesday, the government added the U.S. to the list of banned meat importers.\n\nThe ban was announced by the state consumer watchdog Rospotrebnadzor. A government official told Russian media that American meat imports were now being kept out of Russia on a \u201ctemporary basis\u201d.\n\nRussia is the fourth largest importer of U.S. meat, buying $51 million worth of poultry and $22 million of pork in 2016. According to the National Pork Producers Council, the U.S. shipped 27,600 metric tons of pork to Russia in 2016, which made up 0.3 percent of total U.S. pork exports.\n\nRead more: Polls Say Few Americans Think Trump is Putin's Puppet\n\nThe ban comes at a particularly difficult time for the U.S. meat industry, which is still struggling to recover from a disastrous salmonella outbreak that began in 2015. China, Vietnam and Mexico have all banned U.S. pork imports, while the EU and South Korea are still testing shipments.\n\nThe bans come as U.S. meat exporters have had difficulty convincing importers to trust American meat following the 2015 crisis. The outbreak, which resulted in more than 60 percent of pork being recalled, was linked to at least 184 illnesses in 20 states. While U.S. authorities traced the outbreak back to a farm in Missouri, they were unable to identify the exact source of contamination.\n\n\u201cThis has been a major setback for us and it will take a while for the industry to recover,\u201d said National Pork Producers Council spokesman Dave Warner.\n\nIt's not clear when Russia will lift its ban on U.S. meat, but Russia's state consumer protection watchdog has said it would allow Ukrainian meat imports.]" time="0.025"><properties><property name="score" value="0.08315763953610984" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08315764&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08315764
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Is there anything you would change about this book?\n\nIt's too long. I got bored. It's the wrong length for a book about the Crusades. I know you want to make it as comprehensive as possible, but let's face it, the Muslim perspective is the only one that matters. You can understand why people fought for this or that minor piece of territory, but you can't understand why they fought in the first place.\n\nHow would you have changed the story to make it more enjoyable?\n\nI wish they had done a better job of establishing the atmosphere of the time, instead of just going for sensationalism.\n\nHow could the performance have been better?\n\nThe narrator had a good voice for this.\n\nWas Crusade the right choice for you? Why or why not?\n\nNo, I'd never read about the Crusades before.]" time="0.014"><properties><property name="score" value="0.029908921044535797" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02990892&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02990892
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Most Helpful Positive Review\n\nI would purchase this product again\n\nI have used a baby wrap before for a previous baby and loved it, but this one is my favorite. The material feels great and is breathable, the design is cute and easy to use, and I love...Read complete review\n\nI have used a baby wrap before for a previous baby and loved it, but this one is my favorite. The material feels great and is breathable, the design is cute and easy to use, and I love how it fits under my shirt to show a little but isn't quite as obvious as a sling. I use this for my 1 month old and we have never had any problems with the carrier not staying tight enough. I would definitely recommend this product!]" time="0.012"><properties><property name="score" value="0.036983757214403834" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[A northwest Houston woman is under arrest, accused of child abuse.\n\nPolice say they were called to the woman's home on Almeda School Road around 10:45 a.m. Thursday. They say her husband had to take their child to the emergency room after the child suffered a burn.\n\nPolice say the child suffered burns to the back of his neck, shoulder and upper back area.\n\nPolice say the wife, 31-year-old Mercedes Victoria Gonzalez, was arrested.]" time="0.016"><properties><property name="score" value="0.04875156435037997" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04875156&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04875156
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[(NC)\u2014Despite the odds, there are some consumers who decide to head to the stores this weekend to start Christmas shopping. For some, the reason is because they still have some of their shopping list to complete. Others, however, have an entirely different reason.\n\nAccording to Statistics Canada, the average Canadian adult will spend $473.12 on gifts this holiday season. That means the average consumer will spend almost $35 more this year than last year.\n\n\u201cA lot of consumers have gone overboard,\u201d says Ted Michalos, founder of CMG Financial. \u201cIt\u2019s not necessarily the spending that\u2019s the problem, it\u2019s what they\u2019re spending the money on.\u201d\n\nThat\u2019s because many shoppers are buying gifts for themselves. That doesn\u2019t necessarily mean they\u2019re spending on the latest fashions, but rather toys and electronics. And the majority of these gifts, at 41.8 per cent, are going to be purchased online.\n\nHowever, the reason why so many people are spending online is because they\u2019re trying to avoid the crowds and waiting in lines.\n\n\u201cSome consumers just get too stressed and they go online to get their stuff,\u201d says Michalos. \u201cAnd if they do that, they\u2019re likely going to be paying more for the gifts.\u201d\n\nThat\u2019s because while there are fewer people at the store, they\u2019re paying more per person. For example, if someone is paying $50 for a gift in the store, they\u2019re likely to pay more than $50 online.\n\n\u201cThis is really the time where you\u2019re going to have to do your best to have everything done by the 15th,\u201d says Michalos. \u201cAnd it\u2019s definitely a good idea to sit down with your spouse or significant other and look at your budget. And when you have a child, the best thing to do is to write down what you\u2019re going to spend. That\u2019s a real good thing to do, because if you don\u2019t, you\u2019ll overspend.\u201d\n\nAnd that could come with consequences, because if you\u2019re overspending, you could be limiting your ability to spend on your children or other family members in the future.\n\n\u201cYou can go overboard and give away everything you have to your kids and that could be a problem,\u201d says Michalos. \u201cThis could be an opportunity to teach your kids about moderation, and also the difference between \u2018I want\u2019 and \u2018I need.\u2019\u201d\n\nBecause with Christmas, most people are shopping with the mindset of \u2018I want,\u2019 not \u2018I need.\u2019 And that can be a problem, because it will prevent you from saving or investing.\n\n\u201cIf you\u2019re going to be shopping on a credit card, you\u2019re going to be in big trouble,\u201d says Michalos. \u201cIf you\u2019re not able to pay off that debt by the end of January, you\u2019re going to be in big trouble.\u201d\n\nWant to know more about debt? Try these articles:\n\nNew Christmas ad showcases true meaning of the holidays\n\nTake advantage of these holiday deals\n\nHow to get your kids to be thankful for what they have\n\nWhat do you think? Join the conversation in the comments below or contact us using the \u201cHave Your Say\u201d form on this page.]" time="0.028"><properties><property name="score" value="0.03021594934011624" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03021595&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03021595
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Developing in a Virtual Machine on your computer or laptop\n\nHello , Here we are going to learn how to create virtual machine on your computer for testing purpose. Before we start , it is better to know what is virtual machine. A virtual machine is a simulated hardware that allows you to run other operating systems (or other versions of the same OS) at the same time as your current OS, inside a window on your current OS. It can also be thought of as a computer inside a computer. With a virtual machine, you can: Run another operating system (or another version of the same operating system) at the same time as your current operating system. Try out a new operating system without having to install it on your hard disk. Run applications for one operating system on another, without having to reboot your computer.\n\nStep 1: Create a New Virtual Machine\n\nVirtualBox provides a wizard that will assist you in creating a virtual machine. Launch the program and select \u201cNew\u201d.\n\nStep 2: Create a Virtual Machine\n\nThe first window you see is the New Virtual Machine Wizard. Select \u201cCreate a virtual machine\u201d.\n\nStep 3: Create a Virtual Machine\n\nIn the next window, enter the name for your virtual machine. This name will appear in the main VirtualBox window next to the icon for the virtual machine.\n\nStep 4: Create a Virtual Machine\n\nThe next window is where you select the operating system you want to install.\n\nStep 5: Create a Virtual Machine\n\nIn the next window, select the amount of memory you want to allocate to your virtual machine. The more memory you have, the more applications you can run simultaneously. However, keep in mind that the more memory you allocate, the slower your computer will be.\n\nStep 6: Create a Virtual Machine\n\nSelect \u201cCreate a virtual hard disk\ufffd]" time="0.025"><properties><property name="score" value="0.09069610318827134" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The internet is a funny place, where people can get away with saying and doing the weirdest stuff that would otherwise be totally unacceptable.\n\nBut it turns out not even the internet is a free pass to sexually harass people.\n\nThat's what one woman learned when she was sent a totally inappropriate, NSFW photo from a stranger, that happened to be someone she knew in real life.\n\nTwitter user @kendtee, who describes herself as &quot;an arm candy&quot;, tweeted about her encounter with the man she knew.\n\nAccording to @kendtee, the incident occurred after she agreed to take a picture with a guy. She then deleted the picture from her camera roll when it was over, and the man asked if she had. She said yes, and then he said this:\n\n&quot;It's a shame we can't see the photo anymore because it was pretty good&quot;\n\nShe didn't know how to respond. He then tweeted her back, asking her to &quot;leave [the photo] up for a little while&quot;. She declined.\n\nApparently, he then texted her back, asking her to send him a copy of the picture.\n\nShocked by this behaviour, she shared the exchange on Twitter, and said the worst part was how &quot;unbothered&quot; he seemed about it.\n\nEventually she shared a screenshot of the exchange, and people were not impressed.\n\n&quot;Oh wow. 'I'd do ya if I could' and then the photo request. You should probably just block him and be done with it. 'Doing ya' or not,&quot; one person said.\n\nAnother person had an idea on how to deal with it, tweeting, &quot;You know what? Be a bigger person, send it to him. For posterity.&quot;\n\nMany others were in complete agreement with this sentiment.\n\n&quot;I don't know this man and I probably never will. But I 100% understand why he's going to be blocked. A man who says that is definitely not a good person. All women should know their boundaries with men and this seems to me to be clear cut,&quot; one person said.\n\nIn the end, @kendtee went back to him and asked him to explain why he'd sent her that photo.\n\nAccording to @kendtee, he sent her this message in response:\n\n&quot;How do you think I know that you're gonna see this tweet and block me or delete it.\n\n&quot;I'm not trying to be a creep you just said that you'd do me so I figured we had a shot lol\n\n&quot;I like your style and want to get to know you better\n\n&quot;If you send the picture I promise not to be creepy and stop bothering you\n\n&quot;I'm being serious here and I don't think I deserve the reaction I got\n\n&quot;All I wanted to do was take a picture with a nice girl and see if she'd let me\n\n&quot;I don't want to be a bad person or make you feel uncomfortable.\n\n&quot;If you like I'll just go away\n\n&quot;You're gorgeous and I'm sorry I don't deserve to meet you but I thought I'd try my luck\n\n&quot;I won't message you again if you don't want me to\n\n&quot;I'll delete the tweet and block you\n\n&quot;I just wanted to try\n\n&quot;Please&quot;\n\nThe rest of Twitter wasn't impressed, calling him out for still not getting it, and for trying to shift the blame.\n\n&quot;Like you have to explain what 'likes your style' means, like you're trying to shift the blame. But it's not ok. It's not ok for you to message someone for a photo. It's not ok for you to say 'I'd do you' and it's not ok for you to try and justify your actions,&quot; one person said.\n\nAnother pointed out that he was trying to pressure her into sending him the photo, and wasn't taking responsibility for his actions.\n\n&quot;He's a creep, and still wants to force you to send him a picture of your body. He'll delete the tweet if you do, otherwise you can block him. This is on him, not you. You have no responsibility to 'meet' him,&quot; another person said.\n\n&quot;Ok let's stop there with the 'I would've taken a photo with you anyways' thing. That's not what this is about. This is about someone sending unsolicited nudes,&quot; another person pointed out.\n\nIn response, @kendtee wrote that she wasn't bothered, because at the end of the day she was a &quot;pretty woman who didn't need to be respected&quot;.\n\n&quot;I didn't get mad or freak out or feel the need to drag his name through the mud. I just calmly informed him that I have a man in my life and that I'm flattered but I don't share photos like that and I'm taken. He didn't like it so he blocked me and I'm okay with that,&quot; she wrote.\n\nALSO ON HUFFPOST AUSTRALIA]" time="0.072"><properties><property name="score" value="0.08317489011280768" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08317489&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08317489
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Perhaps one of the most beautiful areas of Southern California is the Pacific Palisades, a neighborhood filled with high rises, luxury estates, and tons of culture. From the obvious to the under the radar, Pacific Palisades has some of the most fascinating museums, gardens, and activities you can visit. When you need to get out of the house and explore a new part of Los Angeles, the Pacific Palisades is a fantastic choice.\n\nThe Getty Villa\n\nThe Getty Villa is a sight to behold, and is a must see when you are in the Pacific Palisades. You are welcome to explore the museum for free on the first Thursday of every month. They offer a free trolley that will take you to the Getty Villa, so you don\u2019t have to worry about navigating through the Pacific Palisades on a crowded day. This is the perfect time to visit the Pacific Palisades.\n\nThere are two events going on at the Getty Villa on Thursdays: Story Time and Pacific Voices. Story time will give you a better understanding of ancient Greek and Roman culture. It is a good idea to purchase your tickets in advance, so you can get a discount. The Pacific Voices event is a singing performance, and a great way to enjoy your time at the Getty Villa.\n\nSanta Monica Mountains National Recreation Area\n\nThe Santa Monica Mountains are a beautiful way to spend a day in the Pacific Palisades. The National Recreation Area has a fantastic number of trails you can explore, and is a great way to get out of the house and spend time with your family. Take the trail to Leo Carillo State Beach, or go for a hike along Rustic Canyon. Whichever trail you take, you are sure to enjoy your time in the Pacific Palisades.\n\nSanta Monica Baykeeper\n\nThe Santa Monica Baykeeper is a fantastic organization that works hard to clean up the Santa Monica Bay. This is a great way to spend time in the Pacific Palisades and make a real difference. Check their website for upcoming volunteer opportunities. You will learn more about the Santa Monica Bay, and get a chance to spend time with some interesting people.\n\nPacific Palisades Center for the Arts\n\nThe Pacific Palisades Center for the Arts is an excellent way to spend an afternoon. This isn\u2019t a museum, so you don\u2019t need to book in advance. It is free to walk around the grounds, and to watch a performance. They have everything from jazz concerts to visual arts. If you want to learn more about the Pacific Palisades, then this is a great place to start.\n\nPacific Palisades Farmers Market\n\nThe Pacific Palisades Farmers Market is an excellent way to spend your Saturday in the Pacific Palisades. From fruits and vegetables, to homemade items and baked goods, there are plenty of delicious items to choose from. This is a great place to bring the kids, as they have tons of activities.\n\nPacific Palisades Art Walk\n\nThe Pacific Palisades Art Walk is a fantastic opportunity to view some of the beautiful artwork created by locals. It is a fun evening filled with laughter and good food. You are welcome to bring your own picnic basket and enjoy it in the beautiful Pacific Palisades.]" time="0.071"><properties><property name="score" value="0.036461209093048556" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[We have been very pleased with the support, workmanship and professionalism of the team at WeFix. They have been able to provide quick and effective solutions for any problems that have arisen. The attention to detail, willingness to understand and understand the importance of high quality work have been important to us. The flexibility of being able to accommodate and adapt to any situations has been very helpful in terms of our team and their working styles.\n\n- Rick Phillips, Site Manager \u2013 Maersk Oil]" time="0.013"><properties><property name="score" value="0.012447011134097963" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Mastering financial knowledge and planning for a good life is a lifelong process. But when it comes to money, how many people have the knowledge to plan for a good life? And what if you had this information, would you use it? How would you live? Would your life change for the better? Would your relationships improve? Would your life be happier?\n\nIn most cases, we don\u2019t have enough financial knowledge to make the right decisions. We don\u2019t have the knowledge and therefore, we end up making bad decisions. The result of these bad decisions is that we lose more than we gain. We lose money and we also lose our peace of mind.\n\nI have experienced both sides of the coin and I know the emotional, mental and physical impact of being financially unstable. Money makes and breaks people. It has broken me in the past and it took me years to recover. But I am better now.\n\nIf you too want to make peace with your money]" time="0.014"><properties><property name="score" value="0.0007165836627798239" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Tom Shanks (writer)\n\nTom Shanks (born 1961) is a Scottish writer, journalist, and musician. He is the author of eight novels, including &quot;The Hidden&quot;, the &quot;Vampyrrhic Times&quot; trilogy, and &quot;Eve&quot;, an &quot;Eve&quot; (comics) spin-off novel.\n\nShanks has been living in France since 1983 and is married to the French writer Marie H\xe9l\xe8ne Poitras.\n\nShanks was born in Edinburgh, Scotland in 1961.\n\nShanks has been writing since the mid-1990s and has published several novels.\n\nShanks has also been writing for the British comic book magazine &quot;2000 AD&quot;. His first script for them was the short story &quot;Trevor's Tale&quot;, drawn by Simon Harrison, in &quot;2000 AD&quot; #1704, published in March 2008.\n\nThe comic book &quot;Eve&quot; was written by Shanks. It was released by Black House Comics in 2011 and the second volume was published in 2012. In 2012, IDW Publishing published a comic book prequel to &quot;Eve&quot;, &quot;Eve: The Awakening&quot;. The book was written by Shanks and Marie H\xe9l\xe8ne Poitras and illustrated by Enrica Angiolini.\n\nShanks has also written the script for &quot;Apocalypse Cow&quot;, which was originally published in French and German as &quot;La Vache Apocalypse&quot;. It was adapted as a comic book by Tony Lee.\n\nIn March 2010, the novel &quot;Vampyrrhic Times&quot; was published by Gollancz in the UK and by Roc in the US. It was illustrated by Tony Harwood. The second volume of the trilogy, &quot;The Wisdom of Dead Men&quot;, was published in the UK in 2011 and the US in 2012. The third volume, &quot;The Trade of Queens&quot;, was published in the UK in 2013 and in the US in 2014.\n\n\n\n\n\n\n\nShanks and Poitras run the Black House Comics publishing house.\n\nShanks is a member of the band The Bassholes, along with co-writer of Eve, Marie H\xe9l\xe8ne Poitras. The band is composed of Shanks and Poitras on vocals and bass, Poitras' husband, French artist St\xe9phane Paitreau, on guitar, and drummer &quot;Chino&quot; Gonzales. They released their first album, &quot;No More Mr Nice Guy&quot;, on May 1, 2010, on Stuck on a Name Records. The album includes covers of classic rock songs, such as &quot;Jailhouse Rock&quot; and &quot;Born to be Wild&quot;. In 2013, the band released a second album, &quot;Sex Lies &amp; Basslines&quot;.\n]" time="0.024"><properties><property name="score" value="0.05692888205376786" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05692888&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05692888
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Rick Lake\n\nRichard G. Lake II (born April 7, 1979) is an American mixed martial artist who most recently competed in the Welterweight division of the Ultimate Fighting Championship. A professional competitor since 2004, he has also competed for the Quad Cities Silverbacks of the IFL, and was a competitor on Spike TV's &quot;&quot;.\n\nBorn and raised in the Quad Cities area, Lake attended East Moline High School.\n\nAfter graduating from high school, Lake worked for two years as a full-time welder, before deciding to pursue a career in mixed martial arts.\n\nLake made his professional mixed martial arts debut on October 30, 2004 for the Quad Cities Silverbacks of the International Fight League. He won his first three fights for the promotion, before suffering his first loss to Ian Loveland on April 8, 2005.\n\nAfter picking up three wins in the IFL, Lake was signed by the UFC. He was expected to make his debut at UFC Fight Night 7 against Jason Gilliam. However, Gilliam was forced from the card with an injury, and was replaced by Jorge Rivera. Lake lost to Rivera via first-round KO on August 23, 2007.\n\nFor his second fight in the promotion, Lake dropped down to Lightweight and faced Kyle Bradley at UFC Fight Night 11 on November 17, 2007. He won the fight via unanimous decision (30-27, 30-27, 29-28).\n\nIn his third fight for the promotion, Lake dropped to Featherweight to take on Mike Brown at UFC Fight Night 13 on June 16, 2008. He lost the fight via unanimous decision (29-28, 29-28, 29-28).\n\nFor his fourth fight in the promotion, Lake dropped to Bantamweight and faced Brad Pickett at UFC 95 on February 7, 2009. He lost the fight via TKO in the first round.\n\nFor his fifth fight in the promotion, Lake faced Tommy Hayden at UFC 100 on July 11, 2009. He won the fight via submission in the first round, earning &quot;Submission of the Night&quot; honors.\n\nFor his sixth fight in the promotion, Lake faced George Roop on March 21, 2010 at . He won the fight via unanimous decision.\n\nFor his seventh fight in the promotion, Lake fought Roland Delorme on June 19, 2010 at . He lost the fight via submission in the first round.\n\nFor his eighth fight in the promotion, Lake faced John Gunderson on November 20, 2010 at UFC 122. He won the fight via split decision.\n\nFor his ninth fight in the promotion, Lake fought Jason Reinhardt on February 5, 2011 at UFC 126. He won the fight via TKO in the first round.\n\nFor his tenth fight in the promotion, Lake faced newcomer Darren Uyenoyama on June 11, 2011 at . He won the fight via unanimous decision (30-27, 30-27, 30-27).\n\nFor his eleventh fight in the promotion, Lake faced former WEC Featherweight Champion Mike Brown on October 29, 2011 at UFC on Versus 6. He lost the fight via unanimous decision (29-28, 29-28, 29-28).\n\nFor his twelfth fight in the promotion, Lake faced Nam Phan on May 15, 2012 at UFC on FX 4. He lost the fight via unanimous decision.\n\nFor his thirteenth fight in the promotion, Lake faced Sam Sicilia on July 11, 2012 at UFC 149. He won the fight via unanimous decision.\n\nFor his fourteenth fight in the promotion, Lake faced Ivan Menjivar on December 15, 2012 at UFC 155. He won the fight via split decision.\n\nFor his fifteenth fight in the promotion, Lake faced Jeremy Larsen on June 15, 2013 at UFC 161. He won the fight via TKO in the second round.\n\nFor his sixteenth fight in the promotion, Lake faced promotional newcomer Justin Salas on November 16, 2013 at UFC 167. He lost the fight via split decision.\n\nFor his seventeenth fight in the promotion, Lake faced fellow newcomer Max Holloway on February 15, 2014 at UFC 169. He lost the fight via TKO in the first round.\n\nAfter his release from the UFC, Lake dropped down to Bantamweight and made his debut on April 11, 2015 at IFC: International Fighting Championship 2 in Des Moines, Iowa against Rafael &quot;Morcego&quot; da Silva. He lost the fight via knockout in the first round.\n\nLake faced Gabriel Checco on May 28, 2015 at CXF 5. He lost the fight via split decision.\n\nLake faced Daniel Weichel on August 14, 2015 at Superior Challenge 12. He lost the fight via submission in the first round.\n\nIn his fourth fight since returning from the UFC, Lake faced Ryan Quinn on October 23, 2015 at CXF 8. He won the fight via knockout in the first round.\n\nLake faced Jim Alers on January 8, 2016 at CXF 10. He won the fight via TKO in the second round.\n\nLake faced Martin Stapleton on February 19, 2016 at MFW: A Night of Champions. He lost the fight via split decision.\n\nLake next faced Dylan Logan on September 2, 2016 at Cage Warriors 82. He won the fight via TKO in the second round.\n\nOn February 28, 2017, it was announced that Lake had signed with the UFC and was expected to face Alessio Di Chirico on March 18, 2017 at UFC Fight Night 107. However, on March 7, Di Chirico pulled out from the event, citing injury, and was replaced by promotional newcomer Marcin Prachnio. Lake won the fight via TKO in the second round.\n\nLake faced Gregor Gillespie on June 25, 2017 at UFC Fight Night 112. He lost the fight via split decision.\n\nLake faced Jarred Brooks on December 9, 2017 at UFC Fight Night 123. He lost the fight via submission in the first round.\n\nLake faced Drakkar Klose on February 24, 2018 at UFC on Fox 28. He lost the fight via TKO in the second round.\n\nLake faced Andre Fili on August 25, 2018 at UFC Fight Night 135. He lost the fight via TKO in the first round.\n\nLake faced Desmond Green on February 17, 2019 at UFC on ESPN 2. He lost the fight via technical knock out.\n\nIn July 2007, Lake was diagnosed with Crohn's disease, an inflammatory bowel disease that causes ulcers in the digestive tract. He began experiencing symptoms shortly after moving to California to train full-time. His Crohn's was brought under control through medication and diet, and it is currently in remission.\n\n\n\n]" time="0.073"><properties><property name="score" value="0.038204507508995254" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Fenix Funk\n\nFenix Funk is the debut studio album by German electronic music group Telepopmusik. It was released on 9 October 2000 by Columbia Records. It was a success in Japan and peaked at #50 on the Oricon Albums Chart.\n]" time="0.017"><properties><property name="score" value="0.04325450089121552" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0432545&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0432545
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[If you\u2019re working on a SharePoint app, you\u2019ll eventually need to get hold of a SharePoint context. At the time of this writing, there are two options: SharePointContext and SPUtility. Both work fine, and I think it\u2019s really up to your personal preference and comfort level to choose either. But for my purposes, I\u2019m a fan of using SharePointContext because it will use the most current context available, which is what I like to do.\n\nLet\u2019s take a quick look at the two SharePoint context providers that we have at our disposal: SharePointContext and SPUtility. We\u2019ll use these in conjunction with the new SharePoint REST API to get a SharePoint context.\n\nSharePointContext\n\nThe SharePointContext class represents a SharePoint context. The context is the central part of the SharePoint client object model, and it\u2019s available from any client, including SharePoint-hosted, remote, or SharePoint-hosted-remote applications.\n\nTo create a SharePointContext object, you call the SharePointContext.GetContextAsync() method and pass in a SharePointContextOptions object. The SharePointContextOptions object provides the user information and authentication settings. We\u2019ll dive into the details in the \u201cAuthorization\u201d section of this article.\n\nOnce you have your context, you can then call methods on it. The context has a number of methods on it to retrieve different information.\n\nLet\u2019s take a quick look at the available methods:\n\nGetContextAsync() - The method that is used to create the context.\n\n\n\n- The method that is used to create the context. SharePointContextPermission - The property that contains the permission token used to retrieve the current user.\n\n\n\n- The property that contains the permission token used to retrieve the current user. SharePointPermissions - The property that contains the current permission level of the user.\n\n\n\n- The property that contains the current permission level of the user. ListPermissions - The property that contains the permission level for the user to access lists.\n\n\n\n- The property that contains the permission level for the user to access lists. SPHostUrl - The property that contains the URL of the current host application.\n\n\n\n- The property that contains the URL of the current host application. Site - The property that contains the current site of the user.\n\n\n\n- The property that contains the current site of the user. UserPrincipalName - The property that contains the username of the current user.\n\n\n\n- The property that contains the username of the current user. Web - The property that contains the current web of the user.\n\n\n\nSPUtility\n\nThe SPUtility class provides static methods to retrieve a SharePoint context and return it. To get the context, you call the SPUtility.GetSharePointContext() method.\n\nThe methods you have available on the SPUtility.SharePointContext object are:\n\nClientContext\n\n\n\nWeb\n\n\n\nGetClientContext()\n\n\n\nGetWeb()\n\n\n\nSharePointContext Permission\n\nBefore we dive into the SharePoint context, let\u2019s talk about permission. The SharePoint context is going to be pretty useless without a permission token. We\u2019re going to talk about this permission token a lot in this article, so let\u2019s just take a quick look at how it\u2019s generated.\n\nWe\u2019ll need to use the SharePoint permissions provider to generate the token. To do so, you\u2019ll need to call the SharePointPermissionManager.GetPermissionsAsync() method. You\u2019ll need to pass in an instance of the site where you want the permissions for.\n\nIn this example, we\u2019re going to call GetPermissionsAsync() and get the current user\u2019s permissions.\n\nGetPermissionsAsync() returns a SharePointPermissions object, which you can use to set your permissions on lists, folders, and documents. The SharePoint permissions are applied in a single pass. This means that you can\u2019t apply different permissions to different entities. For example, if you want to apply permissions on the web, the site, and the lists and folders, you\u2019ll need to pass in the permission on the web, then on the site, then on the lists and folders. You can, however, pass in different permissions to different lists and folders.\n\nOne other thing to note about SharePoint permissions is that they only grant access to the web, not to the list or to the folder itself.\n\nFor example, if a user doesn\u2019t have the Edit permission to a list, he will still be able to view it. But if the user has the Edit permission to the list, he will be able to edit it.\n\nIf you want to deny access to the whole list, you can set the permissions for the list and all its subfolders to Deny.\n\nIn order to create the context, we\u2019ll use the context, the SharePoint permissions provider, and a helper class, which we\u2019ll talk about later in the article. The full example is shown in Listing 1.\n\nThe context object will now be passed to the constructor of a class called FileRootsProvider. The FileRootsProvider class is part of a helper class that I wrote and will be available to you in this article.\n\nThe provider will be created with the context and then called to generate the access token. Once the access token is generated, it will be returned to us, and we can start to retrieve information about the site and its lists and folders.\n\nAccess Token\n\nWhen you call the SharePoint permissions provider, you\u2019ll receive a token that represents the permissions for the current user. The token is returned as a SharePointPermissions object, which is what we\u2019ll use to apply the permissions.\n\nTo create a SharePointPermissions object, you need to pass in the site URL where the permissions should apply. For our example, we\u2019ll pass in the site URL to the SharePointPermissions constructor.\n\nAfter you\u2019ve created the permissions, you can apply them to the lists and folders. You can do this by passing in an instance of a list or folder to the SharePointPermissions constructor. If you want to apply permissions to all lists and folders, you can simply pass in an empty list or folder.\n\nNow that we\u2019ve created the permissions, we can use the FileRootsProvider class to generate the access token. The access token is returned as a SharePointContextPermissions object. The SharePointContextPermissions object contains the permission token and other details that we\u2019ll use to get information from the SharePoint site.\n\nAs you can see, the code to get the access token is pretty straightforward. The example is shown in Listing 2.\n\nList\n\nGetting the access token isn\u2019t all you need to do. You also need to be able to get a list and get a list item. To do this, you\u2019ll need to call the SharePointContext.GetListAsync() method. The method is easy to use and returns a list object.\n\nIf you\u2019re looking for a specific list, you can pass in the list GUID or the URL of the list. Listing 3 shows an example of both the URL and GUID to get the lists from.\n\nItems\n\nThe code to get an item from a list is just as easy as getting the list. You can pass in the list item GUID to the GetListItemAsync() method. Listing 4 shows an example of how to get the list item with the GUID.\n\nSummary\n\nThis article was a quick look at how to get a SharePoint context. The examples used a helper class that I created to make it a little easier to get access to lists and items from SharePoint.]" time="0.071"><properties><property name="score" value="0.11422689706748015" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[A few months back, I wrote an article about the journey from yearning to prayer. Since I have been praying on and off for many years, I was finally blessed with the chance to journey from yearning to prayer. It took a few weeks but it did happen. In the past, my experience of praying was somewhat a miserable one. It was not that I was not able to pray but the truth was I never really enjoyed the praying process. As a matter of fact, I used to hate praying. When I was younger, I was more interested in action than in words. I just want to know what I have to do and get it done. Thus, I never really understood praying. I just felt I was doing nothing but looking at the ceiling. It was only after I found my faith in Jesus that I have understood praying.\n\nMy first experience of prayer was on the cross. I experienced the most painful prayer in my life when I saw Jesus on the cross. He was in pain and so am I. He was looking at the ceiling of the cross and so did I. In the midst of my pain, I opened my eyes and saw Him. I realized that my prayers were not on my behalf but for Him. I looked at Him and realized that He was the only one who can pray for me. I looked at Him and asked him to pray for me. I have never felt anything like it. I have never felt more connected to my God than that moment. I have never felt more at peace than that moment. I realized that praying is actually for Jesus. I realized that Jesus came to this world for a purpose. The whole point of his coming was to save mankind. I realized that He came to the]" time="0.024"><properties><property name="score" value="0.02775482711659486" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The Outrageous Princess Costume is a fun, plus size costume for Halloween. The costume is made of plush, purple velvet. The bodice is purple with a black bow at the top. Black ruffles hang from the bodice. A gold foil belt is attached to the skirt with gold studs. The skirt has a black tulle underskirt with a gold and silver trim. Two purple tulle bows are attached at the hem of the skirt. The wig is a fun purple, curly hairstyle with a tiara and gold tiara veil. Get some props and have fun with this plus size costume.\n\nKey Features:]" time="0.014"><properties><property name="score" value="0.04260629750676426" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0426063&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0426063
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[See What We\u2019ve Been Up to!\n\nDo you want to know what we\u2019ve been doing with the movies? Well, here\u2019s a quick rundown.\n\nCOLD WAR (with Tim Roth) \u2013 Dan is currently in development on a feature-length version of the short with Tin Star TV.\n\nLIFE SUPPORT (with Julie Walters) \u2013 The full feature has just finished post-production and is looking for a distributor. It has had a private screening at this year\u2019s BFI London Film Festival.\n\nDREAMWORLD (with Douglas Henshall) \u2013 Life and times of Alfred Wallace, who came up with the theory of natural selection, but had his ideas pinched by the infamous Charles Darwin. The film is in pre-production with Electric Pictures.\n\nSHELTER (with Adam Rayner and Bernard Hill) \u2013 UK feature about a homeless man and the estranged daughter who cares for him. Dan is currently in pre-production with Lionsgate.\n\nCHEERIOS (with Harry Lloyd) \u2013 Another one of Dan\u2019s shorts, this is a comedy drama about a family with a dying son. The short has been selected for film festivals around the world and won numerous awards.]" time="0.013"><properties><property name="score" value="0.1331539191289033" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.13315392&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.13315392
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[To find the best restaurants near you, just enter your location into the search field and the restaurants will show up in a few seconds.\n\nWhat to eat in Italy?\n\nThe selection of Italian restaurants near me is the best on a restaurant and cafe app. Select your region or type in a town, village, postcode, a street name or a square and you will see a list of restaurants near me. Filter by price, type of cuisine or what's popular.\n\nItalian cuisine is known for it's fresh ingredients and regional specialties. To get a good feel for Italian cooking, you need to try each region's specialities.\n\nItalian Cuisine and Where to Eat it\n\nItalian food is all about local and fresh ingredients. Expect pasta, meat, cheese, fish, wine, and bread to feature heavily on a typical Italian menu. Eating out in Italy is an experience in itself, with food served on real plates and in restaurants with atmosphere to match. You can get a taste for Italian food in Britain, but it won't match the real thing.\n\nItalian food in Britain\n\nCrispy, fried, or grilled? With so many different types of pizza, the Italian restaurant menu can be hard to navigate. Luckily, we're here to help you out. Get ready to tuck into some of the best food on the planet. What are you waiting for? Order now and enjoy the best pizza!\n\nItalian Food Delivery\n\nDo you want to make the perfect meal but you are not sure what to cook? Don\u2019t worry, in the times that you are not sure of what to cook, the Italian restaurants have the solution. Just order what you want and our restaurant is the solution for you. Now you can order online and everything you need is in the palm of your hand.]" time="0.024"><properties><property name="score" value="0.05034869831845954" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0503487&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0503487
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Unleash the Power of Your Language\n\n\u201cLearn How To Speak French Fluently In Just a Few Weeks\u201d\n\nFrench For Beginners: Step-By-Step French (5 CD set) \u2013 Learn to speak and understand everyday French words and phrases!\n\nDiscover The Easy Way To Learn How To Speak French\n\nAchieve Your Language Learning Goals \u2013 And Discover What French Is All About!\n\nSee a sample of this course\n\nTOTAL PAGES: 90\n\nGet Started In French Today\n\nMastering the French language may seem challenging, but with this easy-to-follow course, you will quickly become more confident in the French language. This course is designed to help you achieve your goals in speaking, reading, and understanding the language.\n\nLanguage is a fascinating topic, and knowing how to communicate effectively with other people is a key to success in a globalized society. This course is an effective means for you to learn the language of French, and more specifically, the local French spoken in France.\n\nIf you have never learned a foreign language before, this course can get you started right.\n\nIt will give you a clear insight into the basics of the language, and its essential parts. It is also a complete guide to getting to know the local French.\n\nYou will learn how to pronounce French words with clarity, and have a basic understanding of how to converse in the language.\n\nLearn how to hold a conversation in French with confidence!\n\nThis course has been written for people who would like to start learning the French language, as well as those who want to increase their vocabulary, expand their cultural understanding, or need French for their work.\n\nIf you have been studying French for some time and have gotten to a plateau, then this course will help you go beyond that plateau. The course can also be used as a refresher course to give you more confidence in the language.\n\nThere is a no-risk guarantee with this course. You are protected by a 60 day, money-back guarantee, and are fully covered by our no-risk, no-nonsense guarantee.\n\nSome of the topics covered in this course include:\n\nFacts About The French Language\n\nGreetings In French\n\nIdioms And Phrases In French\n\nA Guide To French Pronunciation\n\nAs well as grammar rules, grammar explanations, and hundreds of vocabulary terms.\n\nHere\u2019s What People Are Saying About This Course\n\nGaby Hernandez \u2013 \u201cGreat course! My instructor is patient and very thorough. Thanks for providing such a great resource!\u201d\n\nChloe Evans \u2013 \u201cThis is an excellent course. I learned so much! I would definitely recommend this course to anyone who wants to start learning French.\u201d\n\nPeter Johnson \u2013 \u201cI bought the course to learn a bit more French and for the advanced grammar. The instructor is excellent. The course is concise and well organized.\u201d\n\nSandra Hernandez \u2013 \u201cI am an advanced French student who has taken 4 years of French in high school and 6 years of French in college. I have a French instructor who is native French. I bought the course to brush up on grammar and to work on my listening skills. I thought the course was really good and the instructor is really good. I really enjoyed the course and feel that it is good for anyone, whether they are beginners or advanced. I would recommend this course to anyone who wants to learn French.\u201d\n\nChristina B \u2013 \u201cI bought this course because I am planning a trip to France, and I wanted to brush up on my French. I was pleasantly surprised. I\u2019ve learned so much. The course covers a lot of material in 5 hours. The instruction is very clear. I am very happy with my purchase.\u201d\n\nWhat People Are Saying About LanguageForBeginners.com\n\n\u201cThe lessons are very well laid out and explained.\u201d\n\n\u201cVery thorough, well presented and written.\u201d\n\n\u201cAn excellent course and great teacher.\u201d\n\n\u201cVery clear and to the point.\u201d\n\n\u201cThis course is perfect for beginners.\u201d\n\n\u201cThe course has helped me a lot.\u201d\n\n\u201cExcellent course for beginners.\u201d\n\nThe audio is very good. The teacher is excellent. The course is very well done.\n\nSee what you\u2019ll learn inside this course:\n\nWhat The French Language Is All About\n\nFrench Grammar In Detail\n\nBasic Sentence Structures\n\nGreetings In French\n\nWhat It Is Like To Speak French\n\nA Few Examples Of Conversations In French\n\nHow To Pronounce Words In French\n\nA Few Simple Phrases In French\n\nA Few French Idioms\n\nHow To Learn The Language Of French\n\nAnd much more!\n\nPlus, with this no-risk guarantee, you can\u2019t lose!\n\nIf you want to learn how to speak French, then you are in the right place. Take the first step towards improving your communication skills, confidence, and experience with a new language.\n\nClick the \u201cBuy Now\u201d button now to get started on your journey!\n\nRelated\n\nLanguage For Beginners: How To Speak French]" time="0.073"><properties><property name="score" value="0.04505366056536925" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04505366&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04505366
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[BOOM!! 10 years of Brikk Fast! Wow, time sure flies when you are having fun, making things, and working hard. Brikk Fast has changed a lot over the years, and has become much more than we ever thought it would. The staff at Brikk Fast has grown, we've brought on some awesome employees that have become just as much a part of the Brikk Fast family as we have. In 2010, we opened up a storefront, and that has allowed us to expand on all the wonderful things we do, and the way we do them. This year we've grown a lot as a company, in ways both great and small. As a company we've become much more involved in the local community, with events and fundraisers]" time="0.014"><properties><property name="score" value="0.030129845427390167" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03012985&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03012985
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Self-Selection and Pupils\u2019 Labour Market Outcomes\n\nYoung adults from low socioeconomic status backgrounds are significantly less likely to go to university than their better-off peers. There is a significant self-selection in the probability of attending university and of high A-level achievement, which are key drivers of university participation. This paper exploits the introduction of minimum attainment targets in the National Curriculum for schools to explore the impact of early education policies on students\u2019 achievement and on their likelihood to study at university. We find that when minimum attainment targets are introduced, pupils from low socioeconomic status backgrounds are more likely to reach higher grade boundaries in the A-level exams, are more likely to go to university, and are more likely to study science or maths degrees.\n\nContents\n\nIntroduction\n\nData and methods\n\nAnalysis\n\nDiscussion\n\nConclusion\n\nIntroduction\n\nUniversity participation in the UK has grown substantially since the early 1990s. There are now 1.4 million undergraduates, which is almost double the number in 1992. However, participation rates have remained relatively low among those from lower socioeconomic groups (Furlong and Millar, 2008). For example, 40 per cent of young adults from high-income backgrounds were participating in higher education in 2011-12, compared to 21 per cent from the poorest fifth of families.\n\nTo some extent, low participation among young people from less advantaged backgrounds is to be expected given the historical association of university education with privileged groups. Participation rates increased significantly in the 1960s, after state support for university was extended to less affluent groups. Participation rose further during the 1960s, 1970s and 1980s, but then stagnated in the 1990s and declined slightly after 2002. Only in the past decade has participation among the less well-off again risen to the levels recorded in the early 1990s (see Figure 1).\n\nFigure 1. University participation rates by parental income, 1962-2012\n\nNote: Figures are cohort entry rates for those who went to university in each year. Sources: Department for Business, Innovation and Skills (2014); Sutton Trust/Ipsos Mori (2012); Education Policy Institute (2013).\n\nDespite increased participation, the gap in participation rates between young people from more and less advantaged backgrounds is still large, even after controlling for qualifications. On average, young people from the most advantaged fifth of families are about twice as likely to go to university as those from the least advantaged fifth (see Figure 2).\n\nFigure 2. Proportion of young people who go to university by family income, 1996-2012\n\nSource: Participation rates (log scale) are from the UCAS End of Cycle Reports, 1996-2012. The relative income quintiles are from the Family Resources Survey, 2008-2012. The quintiles have been adjusted to allow for comparison across years by using a fixed poverty line.\n\nAt a recent seminar organised by the ESRC Centre for the Microeconomic Analysis of Public Policy, Professor Martin Brown (2012) suggested that it is \u2018a great mystery why this gap is not widening even more\u2019. He also pointed out that there is a significant self-selection in the probability of attending university and of high A-level achievement, which are key drivers of university participation. It is well-known that, in addition to their socio-economic background, non-cognitive skills, such as motivation, are important predictors of academic achievement. Since the increase in university participation has been driven by the growth in participation of less advantaged groups, it is not surprising that it has also been associated with a fall in the average grades of university entrants.\n\nThis paper is motivated by this observation, that early education policies have played an important role in the increase in participation among less advantaged groups. We focus on two such policies: the requirement to meet minimum attainment targets in English, mathematics and science, and the national literacy and numeracy strategies. We explore the impact of these policies on students\u2019 achievement and on their likelihood to study at university. We find that when minimum attainment targets are introduced, pupils from low socioeconomic status backgrounds are more likely to reach higher grade boundaries in the A-level exams, are more likely to go to university, and are more likely to study science or maths degrees. We also find that the increase in the A-level scores and the higher participation rate associated with the introduction of minimum attainment targets cannot be explained by the improvement in the overall level of academic performance in schools or by the lower socioeconomic status of those schools that introduced minimum attainment targets.\n\nData and methods\n\nTo analyse the impact of minimum attainment targets, we use the school-level dataset that the Department for Education makes available annually for the National Pupil Database (NPD). The NPD contains rich information about individual pupils and their schools. We first use the NPD to examine the evolution of achievement at the end of compulsory education by using cohort-level information on final examinations. We then explore the impact of minimum attainment targets using changes in the composition of achievement. The results we present in this paper are robust to the alternative methods of analysis of achievement levels and changes in the composition of achievement.\n\nIn the first stage of our analysis, we explore the evolution of the A-level examination results by using the cohort-level information on final examinations. To study the evolution of achievement at the end of compulsory education, we use the data on A-level grades in GCSE and A-level examinations from 1996 to 2012. In this period, the examinations have changed significantly, from the General Certificate of Secondary Education (GCSE) to the General Certificate of Secondary Education (GCSE) and the A-level examinations. The A-level examination was first introduced in the early 1950s as an academically rigorous alternative to the less selective B-level examinations (Wynne, 2010). Until 1988, A-level grades (i.e. O/A levels) were assigned by teachers, not by the students. From 1988, the responsibility for grading moved to schools. The examination was revised in the early 1990s, after which it was separated into two components: A-level and AS-level. The A-level examinations were re-structured in 2000. Most importantly, the exams have become linear rather than modular. The exams in each subject are now taken at the end of the two-year course. The exams are now externally assessed, and the grade is assigned according to the nationally defined A-level grade boundaries. The percentage of the cohort that sat the A-level examinations increased from about two-thirds in the early 1990s to more than 90 per cent in the mid-2000s. This reflects a policy shift in the assessment of vocational qualifications, which reduced the proportion of vocational qualifications that counted towards a student\u2019s A-level grade.\n\nWe use a slightly modified version of the grading system developed by Hayward, Parry and Stewart (2009) and Bias et al. (2011). We use a scale of 5 grades to assess the performance of students in different subjects: 5 is an \u2018A*\u2019 grade, 4 is an \u2018A\u2019 grade, 3 is a \u2018B\u2019 grade, 2 is a \u2018C\u2019 grade and 1 is a \u2018D\u2019 grade. The best possible grade in each subject is an \u2018A*\u2019. This grading system differs from the one used in most of the existing analyses of A-level results, which use an A-E grading system. The difference between the two grading systems is illustrated in Figure 3, which shows that about 60 per cent of the students in the upper grade boundary of \u2018A*\u2019 under the old system would be below the upper boundary of \u2018A*\u2019 under the new system. Similarly, about 45 per cent of the students in the lower boundary of \u2018C\u2019 would be in the upper boundary of \u2018C\u2019 under the old system, and about 25 per cent of students in the lower boundary of \u2018C\u2019 would be in the upper boundary of \u2018C\u2019 under the new system. We expect that the increase in the A-level scores associated with the change in the grading system will be smaller for subjects that were not directly affected by the reform, such as English, maths and science. We do not analyse students\u2019 performance in the core and non-core subjects, because data on the subject of A-level examination was not available for every pupil.\n\nFigure 3. Grading scales of GCSE and A-level examinations\n\nIn Figure 4, we present the cumulative distribution of A-level scores for each cohort, by the subject studied. Figure 5 shows that the average A-level score increased from a little less than \u2018C\u2019 in 1996 to more than \u2018C\u2019 in 2012, which is consistent with the improvement in the average GCSE results. The gap in the average grades between different subjects remained stable. The increase in the average A-level grades was particularly large in science subjects and in humanities subjects. We do not find evidence of a widening gap in A-level scores between the richest and the poorest students. In the humanities subjects, there is some evidence of an increase in the A-level scores of the less well-off. However, the increase in the average A-level scores was so large that the difference in the average scores between the richest and the poorest students decreased even though it was not eliminated.\n\nFigure 4. Cumulative distribution of A-level scores in different subjects, 1996-2012\n\nSource: National Pupil Database, Department for Education.\n\nFigure 5. Average A-level scores, by subject, 1996-2012\n\nSource: National Pupil Database, Department for Education.\n\nA key issue in the analyses of trends in A-level achievement is the impact of the increase in the proportion of students taking A-level examinations on the average score]" time="0.075"><properties><property name="score" value="0.0524276731476981" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[East Face\n\n\nSkye\n\nThe Complete Guide\n\n\nSummit view from the top of the Cuillin Ridge\n\nTrevor Nace\n\n\nFirst published in Great Britain in 2015 by\n\nThe Crowood Press Ltd\n\nRamsbury, Marlborough\n\nWiltshire SN8 2HR\n\nwww.crowood.com\n\nThis e-book first published in 2015\n\n\xa9 Trevor Nace 2015\n\nAll rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publishers.\n\nBritish Library Cataloguing-in-Publication Data\n\nA catalogue record for this book is available from the British Library.\n\nISBN 978 1 84797 713 2\n\nFrontispiece: Climber on the summit of the Cuillin Ridge in front of\n\nMuckross Ridge and Beinn Eighe\n\n\nTo my partner Sally, whose support and love made it possible for me to fulfil my dream of a lifetime on Skye.\n\n\nContents\n\n\nTitle Page\n\nDedication\n\nCopyright\n\nAcknowledgements\n\nList of Plates\n\n1 Introduction\n\n2 An Introduction to the Cuillin\n\n3 Access\n\n4 West of Glen Brittle\n\n5 The Coire Gabhail / North-West Ridge\n\n6 The Black Cuillin Ridge\n\n7 The Central Cuillin Ridge\n\n8 The South-East Ridge\n\n9 The Ar\xeate of the Great Tower and the Clach Glas Ar\xeate\n\n10 An Dubh Sgeir\n\n11 The North-East Ridge\n\n12 The Minnoch Burn\n\n13 Selected Bibliography\n\n\nList of Plates\n\n\nPlate 1: Spade Level, Coire an Lochain, Glen Brittle, in winter\n\nPlate 2: View of the Beinn Eighe massif from the east\n\nPlate 3: Snow-filled Coire na Banachdich, Coire an Lochain\n\nPlate 4: Black Cuillin ridge from the Spade Level\n\nPlate 5: Near the end of the Black Cuillin ridge\n\nPlate 6: Climbers on the summit of Tower Ridge\n\nPlate 7: Skye seen from the summit of the Cuillin ridge\n\nPlate 8: Climbers near the summit of Clach Glas Ar\xeate\n\nPlate 9: Snow on the summit of An Dubh Sgeir\n\nPlate 10: Landscape of the Minnoch Burn\n\n\nAcknowledgements\n\n\nFirst of all, I would like to thank John Brailsford, who accompanied me on my very first walk in the Cuillin in 1996 and whose generous and knowledgeable company has been a constant source of inspiration to me. A particular thanks also to Donnie Hall for all his practical support and guidance during my exploration of the Cuillin. I would also like to thank the helpful staff of the Skye Library in Kyle of Lochalsh, and Alison Connell for the excellent maps. I am grateful to the Climbers\u2019 Club for publishing my article on the Cuillin.\n\nFinally, a big thank you to Sally, who has been a constant source of encouragement, inspiration and love, and to my family and friends who have given me endless support.\n\n\nPlate 1: Spade Level, Coire an Lochain, Glen Brittle, in winter\n\n\nPlate 2: View of the Beinn Eighe massif from the east\n\n\nPlate 3: Snow-filled Coire na Banachdich, Coire an Lochain\n\n\nPlate 4: Black Cuillin ridge from the Spade Level\n\n\nPlate 5: Near the end of the Black Cuillin ridge\n\n\nPlate 6: Climbers on the summit of Tower Ridge\n\n\nPlate 7: Skye seen from the summit of the Cuillin ridge\n\n\nPlate 8: Climbers near the summit of Clach Glas Ar\xeate\n\n\nPlate 9: Snow on the summit of An Dubh Sgeir\n\n\nPlate 10: Landscape of the Minnoch Burn\n\n\n1\n\n\nIntroduction\n\n\nThe spectacular granite peaks of the Cuillin on the Isle of Skye have captured the imagination of hillwalkers and climbers for over a hundred years. When George Henry Alcock (1864\u20131941), a climber and mountaineer, visited the Cuillin in 1893, he wrote in his book Through the Highlands and Islands of Scotland:\n\nNo one who had not seen the Black Cuillin and the peculiar forms which the rocks assume would believe that nature could assume such strange shapes as are seen here. The first view of the Cuillin from the south makes them look like a huge fortress of rock, with fantastic turrets and chimneys, and battlemented walls.\n\n\nToday these classic features, viewed from the slopes of the nearby mountains of the mainland, such as the Beinn Eighe and M\xf2ine Mh\xf2r, have hardly changed. But the Cuillin are not just aesthetically pleasing. To their first British explorers in the mid-nineteenth century, the Black Cuillin, which form a massive ridge running from Coire an Lochain in Glen Brittle to An Dubh Sgeir, seemed a forbidding mountain barrier, \u2018a fortress of rock\u2019. Their modern reputation, and that of the smaller range of the Red Cuillin to the south-east, as a major challenge to both rock climbers and mountaineers, was established during the second half of the twentieth century. However, despite the presence of the two main ridges on the Black Cuillin (North-West Ridge and Central Ridge), many individual peaks, and the climbing routes up them, were relatively unexplored. This was to change in the 1970s and 1980s when many of the hard, mixed and winter routes, most of which were first climbed in the Cuillin, were recorded by enthusiastic pioneers. Today, climbing and scrambling in the Cuillin can be combined with hillwalking, and the varied routes, combined with the varied weather of Skye, offer an extremely satisfying and testing experience for a mountaineer.\n\nIn writing this guidebook, I have had to make many decisions regarding what routes to describe, and which to omit. Many climbers may not agree with my choices, and it is inevitable that a guidebook such as this will not be comprehensive. As well as describing my chosen routes, I have tried to include a range of other interesting aspects of the Cuillin, in order to make the book relevant to anyone interested in a mountaineering trip to Skye.\n\nGeology\n\nThe geology of Skye, and especially the Cuillin, is fascinating, and as well as a number of peaks and ridges, it has given rise to a series of dramatic glens, the upper parts of which are in the Inaccessible Pinnacle (Coire nam Bian) to the north of the main ridge. The Black Cuillin consist of metamorphic rocks which were formed as a result of a major mountain-building event in the Caledonian Orogeny, the mountain-building event which took place between 500 and 400 million years ago. At this time, a number of folds (anticlines and synclines) were formed, some of which can be seen today. The Blue Stack of the Old Man of Storr, for example, was the result of the raising up of a syncline (Fig. 1). This event also resulted in the granite being exposed to the surface, but this did not happen everywhere. To the east of the island, most of Skye is underlain by metamorphic rocks, but to the west, particularly on the western coast of the Trotternish peninsula, granite is exposed. The Blue Stack is also part of a major fold which has resulted in the formation of the Trotternish Ridge.\n\nFigure 1. The Blue Stack of the Old Man of Storr.\n\n\nThese major events have also resulted in a relatively complex geological map, with three main rock types (Fig. 2). A general layer of granite, called the Inaccessible Pinnacle Quartz Diorite, extends from the Blue Stack north-eastwards to Beinn Eighe. This granite has been intruded by the Red Cuillin Granite, which has a different composition and thus has different characteristics to the underlying rock. In the central and eastern parts of the main ridge of the Black Cuillin, the rocks are composed of the Papcastle Gneiss, which is a granitic metamorphic rock. However, to the west, around Sgurr nan Gillean and Glas Bheinn, the Papcastle Gneiss has been intruded by the Kintail Granite, giving rise to a more mottled appearance in this part of the Cuillin. A further complex of rocks is exposed on the north side of Glen Brittle, which includes the Torridonian Sandstone, Lewisian Gneiss, and Dalradian schists.\n\nFigure 2. Geology of Skye showing the main rock types.\n\n\nPeaks\n\nAs well as providing interest to geologists, the geomorphology of the Cuillin is also fascinating for climbers and walkers. With so many ridges and peaks, there is much to explore, and each summit offers its own rewards. Some of the individual peaks are not very high, but offer serious scrambles and rock climbs, while others, such as the highest summit of Sgurr nan Gillean (1,018 m) and the first major peak on the Central Cuillin Ridge, Sgurr Dearg (974 m), offer impressive views. Sgurr nan Gillean]" time="0.078"><properties><property name="score" value="0.05450877521319854" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05450878&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05450878
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The Orangeville Banner\n\nOrangeville, Ontario, Canada\n\nApril 9, 1905\n\nPAPER HUNT CONTINUES\n\nINSTEAD OF CASHING $5.00 BILL\n\nA message has been received at the Herald office from Orangeville to the effect that a lady, who has had $5.00 stolen from her purse, did not give the money to an expressman, as stated in last issue, but to another man whom she has not been able to identify, and that the messenger from St. Catherine will not get the money unless he turns it over to her. We have been searching for this lady for several days, and when the Herald office was last searched, we overlooked this message.\n\nAnother story of the stolen money says the woman, whose name is not given, gave the money to a young man who said he was a messenger from St. Catherine, and promised to deliver the money to the party it was addressed to at Kincardine.\n\nTHE MAYOR DISQUALIFIED\n\nA meeting of the citizens of Orangeville was held last Saturday evening at the Orangeville Town Hall, with Dr. A. S. Gibson, M.P., in the chair, to take into consideration the statement made by W. A. MacLean that Mayor D. M. McKenzie was disqualified from holding office under the Municipal Act of Ontario. It was resolved that the matter be placed before the proper authorities for them to decide. The mayor then withdrew from the hall.\n\nMayor McKenzie\u2019s statement in relation to his withdrawal from office is as follows:\n\n\u201cIn this connection I have no desire to act in any unbecoming manner, and I shall resign from the position of mayor. However, I do not desire to leave town. I will reside here and attend to my business as usual, and my many friends who know me and who are not residents of the municipality will be glad to meet me. I want to get at the bottom of this business and find out who it was that drew up the notice of my being disqualified. I was the first one to speak in this town against the big sewer, but I am satisfied now that my opinions are shared by the majority. I do not think that the matter will go much further than the city hall, and as soon as the legal steps are taken the city will be rid of this question.\u201d\n\nSILVER ORGANIZATION\n\nWe are glad to learn that a Silver organization has been formed in this town, and that a meeting is to be held at the Orangeville Opera House next Saturday evening. This is an organization for the purpose of working for free silver, and they are working under the leadership of W. H. Phillips, M.P.\n\nThe representatives of the Anti-poverty society who are here this week will probably have an engagement to-morrow evening. They are endeavoring to make Orangeville the headquarters of the work in this vicinity, and several of the agents will be here for some time. The meeting will be held at the town hall.\n\nA prominent Orangeville man, who is in the wholesale hardware trade in this city, returned this week from Chicago.]" time="0.030"><properties><property name="score" value="0.09227615140252524" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.09227615&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.09227615
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[&quot;Outstanding! Don't miss this book. No matter what is going on in your life right now, this book will speak to you and help you move beyond whatever has you stuck or hurting. Buy it, read it, and be better for it.&quot;\n\n&quot;I am so very grateful that Kari Byron chose to share her story with the world. As a writer, I love to read memoirs, but this one was different than others I have read. The author shared her story in such a way that it was as if I was hearing it from a dear friend. In this era of time, it is refreshing to read something from someone who has the courage to share the good and the bad from their lives. I laughed, I cried, and I grew to appreciate the value of living. I hope this book will find its way into the hands of many who are struggling with life in a way that brings freedom to their hearts.&quot;\n\n&quot;Kari Byron's book is a refreshing look into a woman's journey to finding joy after hardship. She offers hope and practical tips that can be used by women of all ages. This book is a must read.&quot;\n\n&quot;I couldn't put it down! I'm a mom of 3 young kids and struggle with guilt daily, but Kari Byron's words of encouragement and her willingness to share her personal struggles make me feel like I can face my own obstacles and not feel alone.&quot;\n\n&quot;Kari Byron shares her heart in this honest memoir and gives you the tools to follow your heart, your passion and live your life to the fullest! You are sure to find her journey inspiring!&quot;\n\n&quot;I loved this book. From the moment I picked it up, I couldn't put it down. It was a fun, easy read, but the lessons and life advice Kari shared had me reflecting and taking notes. I laughed and cried, and I'll be reading this book over and over again to encourage myself and remind myself of the lessons it teaches.&quot;]" time="0.028"><properties><property name="score" value="0.03148492059010332" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03148492&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03148492
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Since the beginning of time, humans have been fascinated by the mysterious. The sky has always been full of questions. Our ability to dream big and speculate has enabled us to discover, theorize, and theorize even more. But where do we get the answers? These are the top 10 theories on the unsolved mysteries of the universe.\n\n10. Extraterrestrial Origin\n\nThe human race began to emerge from the Stone Age. The caves and forests were just getting familiar with the new inhabitants. There was a long time of fear and ignorance. Before we started to understand the world, they began to imagine what is outside the world. Some thought of creatures so powerful that they could destroy the world. Other speculated that we are a higher level of evolution and are the result of genetic experimentation of extraterrestrial beings.\n\nLater, these speculations were transformed into mythology and the imagination of the humans. In our modern world, it has turned into science. Many scientists still believe that we are not the only life in the universe and that at some point in history there was a connection with extraterrestrials. Some of them even believe that the first stages of our life were carried out with the help of aliens.\n\n9. Time Travel\n\nSince the beginning of history, people have always been fascinated with time. To travel back in time would allow us to explore the unknown, but would it be worth it? We live in a world where time is constant. Change is one of the most important principles of human life.\n\nNevertheless, many scientists still consider the possibility of time travel. The hypothetical method would be to build a giant machine that would make a hole in time. But how can this happen? It would involve breaking the speed of light and space.\n\nThe machine would have to be so big that it would create a massive black hole. Then the time machine would be moved into the center of the black hole, which would allow it to break the laws of time. So far, scientists have not been able to prove the validity of this theory.\n\n8. Time Travelers\n\nDespite the fact that time travel is still a hypothesis, there are people who believe that it is not. They claim to have traveled back in time to the present or to a previous generation. For this reason, there are also numerous stories about people who claim to have returned from the future to save the world.\n\nMost people who claim to have traveled back in time are just crazy. However, many scientists consider it possible that in the future, when we understand the laws of physics better, we may be able to travel through time. This possibility has been a mystery for a long time. So, this is a list of top 10 theories on the unsolved mysteries of the universe.\n\n7. Deja Vu\n\nThe theory of Deja Vu is so controversial that many scientists do not even dare to think about it. The definition of deja vu is, when a person experiences something, although he has never done it before. Many people think of it as a kind of visual hallucination. The most popular theory about the phenomenon is that it is a past memory that has not been triggered by the usual way.\n\nAnother possible explanation is that the brain creates a fake memory to deceive the person who feels it. The theory is that in reality the person does not experience a deja vu, but it is an internal response to the same scenario. In short, the brain recognizes the environment and recognizes the story that it represents.\n\nIn this case, the illusion is that the person has lived through this scenario. Nevertheless, there are many people who do not think so. The reason is that deja vu is a phenomenon that occurs in any person. So, we have 10 theories on the unsolved mysteries of the universe.\n\n6. Water World\n\nIt is one of the most popular theories in the world of science. It is believed that the moon is a result of a collision between the Earth and a body much larger than it. As a result of the collision, our planet was literally ripped. But what happened to the mass that fell on the surface? According to scientists, this part of the Earth remained liquid.\n\nThe evidence is so compelling that the theory of the water world has become a reality. For example, there is a large number of scientists who believe that the moon was the result of the Earth\u2019s rotation and gravitational collapse.\n\nMoreover, there are evidences that this theory may be true. For example, when scientists tried to drill in the Antarctic, they found water underneath. Also, according to many scientists, the planet Mercury, which has no magnetic field, may be the result of the remains of the water world.\n\n5]" time="0.071"><properties><property name="score" value="0.018501592256069812" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[There are no spoilers in this review\n\nFrom the off, the third episode in the new Star Trek Discovery series has a slow pace. It is the most quiet of the three episodes and is more of a character study in the struggles faced by all of the Discovery crew and how that manifests in different ways. With that in mind, I find it strange that this is not getting the universal praise that the first two episodes were. But that\u2019s the power of spoilers I suppose, no one knew what was coming in the first two episodes.\n\nThe episode itself, despite being more quiet than the first two, still manages to deliver plenty of story. You do have to take it with a grain of salt because some of the \u2018bad guy\u2019 stuff is almost unbelievably naive but that\u2019s the Star Trek universe for you. Despite the low stakes, I was still hooked on every word and scene.\n\nThere are plenty of hints that we can be certain that there is a big bad out there that we have not seen yet. Perhaps that is what made the episode more interesting. There is something looming over them and we don\u2019t know what it is yet.\n\nFor some reason, people seem to have a problem with the fact that the crew spend time getting to know one another, with only a couple of scenes of violence and action. Personally, I found that the social dynamics between the characters to be very interesting and engaging. Yes, this is probably true to life and does have some real life parallels. I am glad they did not waste time with all of that mindless action that I\u2019ve become so accustomed to.\n\nThere are no huge shocks in this episode and the pay offs to the mini story arcs were a bit weak. But when the pay off came, I was still glad that they took the time to introduce us to the characters in this way.\n\nIf you have been put off watching Star Trek Discovery because of the fear of spoilers, then I would suggest you don\u2019t worry too much and go ahead and start the show. If you can manage to resist the temptation of the spoilers for the next two weeks, then I think you will be very happy with the third episode and the direction the show is heading in.\n\nLet us know what you thought of the episode in the comments below and if you have seen the first two episodes.]" time="0.027"><properties><property name="score" value="0.02713400207454207" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.027134&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.027134
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Price includes 15% HST, 10% goes to the developer (Living Balance)\n\nWaterfront construction, 11 homes, classic tudor style, quality &amp; contemporary style homes to suit all tastes. Located in Waterfront Developments development at Schooner Bay, golf and public beaches nearby. Great potential with 3 floor plans offered: MLS # V3643336 Single Family Detached, MLS # V3643329 Single Family Detached, MLS # V3643330 Single Family Detached.\n\nThe trademarks MLS\xae, Multiple Listing Service\xae and the associated logos are owned by The Canadian Real Estate Association (CREA) and identify the quality of services provided by real estate professionals who are members of CREA. Used under license. The trademarks REALTOR\xae, REALTORS\xae, and the REALTOR\xae logo are controlled by The Canadian Real Estate Association (CREA) and identify real estate professionals who are members of CREA. Licensed REALTOR\xae. Used under license. Data last updated: 2019-09-21T20:58:44.137.]" time="0.013"><properties><property name="score" value="0.0009180802308550451" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00091808&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00091808
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[You might know of me. I am one of the project managers for Windows 7. Or Windows 7 team is one of the teams that I work in. As a lot of you know, my favorite and almost only thing to do when I am not working is to read a good book. Well, I have done it again. I just finished the book I am about to recommend to you. This book is called \u201cReady Player One\u201d by Ernest Cline. I can\u2019t recommend this book enough. So what\u2019s the big deal? You are about to find out.\n\nSo what is it about?\n\nThis book is a story about Wade, a teenager, living in the year 2045. The world is in a bad state.\n\nTechnology has progressed very fast, but a lot of humans have been left behind. Wade\u2019s life sucks. He lives with his aunt in the slums of Oklahoma. All he does is play video games.\n\nThen, one day, he hears of a contest to win a fortune by finding the first Easter egg in the virtual reality game OASIS. The creator of OASIS has died and has left behind a secret quest. A quest that has the prize to the first person to complete the quest. But there is a catch: only the creator of the game knows how to find the Easter egg. And he has hidden the clues in the game.\n\nIt is up to Wade to try to find the Easter egg first. But not alone. He quickly assembles a group of fellow OASIS players. This group consists of his best friends and fellow gamers. This quest will soon turn into a journey. And I don\u2019t want to tell you more, because I don\u2019t want to spoil the story for you.\n\nA bit about the Author:\n\nErnest Cline is the author of the book Ready Player One. He is from the state of Texas. He has lived in New York City for a few years, but has now moved back to Texas. He is 40 years old and has been writing books for a long time. His first book was published in 2001.\n\nErnest Cline is very much into old science fiction books, and especially the 1980\u2019s. He loves the music, TV shows, video games and movies of that decade.\n\nMy thoughts:\n\nWhen I read this book, I really couldn\u2019t put it down. I just wanted to know how the story would end. I wanted to know if the good guys would win. I wanted to know how they would beat the bad guys.\n\nI am sure you will love this book, just like I did. It is filled with great references to the 1980\u2019s. If you are born after that time, you will probably have to read the story first to get the references, but I am sure you will enjoy it.\n\nI will also tell you, that this book is a great story, and I can easily recommend this book to any reader. It is not just a story about technology. It is also a story about friendship and courage. And in the end it is a story about hope and dream.\n\nIf you want to read this book, you can find it on Amazon. Or your favorite bookstore. I am not going to tell you how much it costs. I am just going to tell you that it is well worth the price.]" time="0.026"><properties><property name="score" value="0.0376311921866399" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03763119&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03763119
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Additional information\n\nApplicants must be 18 years of age or older and must complete and submit the National Labor Relations Board (NLRB) #6 or #7 form, indicating that they will not engage in secondary boycotts or strikes.\n\nThis position is in the bargaining unit.\n\nWe will require your W-2 forms as well as your Federal tax returns for the previous year for verification of your claimed dependents.\n\nApplicants must have unrestricted United States Government security clearance with the Department of Defense or its contractors to be eligible to work in a building that is controlled access.\n\nThis position is located in a National Defense Center. Individuals must be U.S. Citizens or U.S. Nationals (no exceptions) or eligible to be a U.S. Citizen or U.S. National to be hired in this position.\n\nApplicants who have not yet completed a Security Clearance must be able to obtain one in a reasonable amount of time.\n\nApplicants must have or be able to obtain a high school diploma, GED, or other equivalent qualification, and two years of job related experience to be hired.]" time="0.014"><properties><property name="score" value="0.009138547902226568" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00913855&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00913855
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Exterior painting\n\nBrian's painting was done in a timely manner, kept the site clean and did a nice job with all the details. We will definitely use him again. I would recommend him to all my family and friends.]" time="0.014"><properties><property name="score" value="0.013782227846403927" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01378223&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01378223
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Corporate financial managers are responsible for financial reporting, financial planning and analysis, cash management and budgeting, debt management, financial analysis, and financial system development.\n\nWe can help you on the following course:\n\nCorporate Financial Management (DFM)\n\nFurther information about Corporate Financial Management (DFM)]" time="0.015"><properties><property name="score" value="0.028325946409790692" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[True, she is not the most charismatic of the four ladies, but her choices are very strategic and her lines usually have the ring of truth. If there were ever a reason to cast someone as a drag queen, it would be La Chatelaine. Her mannerisms, her expressions, and the way she carries herself scream Drag Queen. My best guess is that she has a very deep understanding of the difference between a woman and a drag queen and she knew how to use that to her advantage.\n\nHollywood]" time="0.013"><properties><property name="score" value="0.0466308819523642" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04663088&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04663088
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[This is the terrifying moment a deadly great white shark was caught on camera devouring a huge tuna fish near a group of stunned fishermen.\n\nThe fisherman from Esperance, Australia, couldn't believe their eyes when they spotted the enormous shark leap out of the water with a tuna fish in its mouth, as they were fishing for yellowfin tuna at Blue Hole, off the coast of Esperance.\n\nFisherman from Esperance, Australia, spotted a great white shark jumping out of the water with a tuna fish in its mouth\n\nThe footage was captured by amateur fisher Domenic Porcelli and posted to Facebook by fishing website Shark Alerts South Australia and South West Australia.\n\n'Watch in amazement as this huge great white shark leaps out of the water with a massive tuna fish in its mouth,' the]" time="0.013"><properties><property name="score" value="0.004555453061388937" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00455545&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00455545
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[HARRY POTTER and the Order of the Phoenix is a 2007 British-American fantasy adventure film directed by David Yates and distributed by Warner Bros. Pictures. It is based on the novel of the same name by J. K. Rowling. The film, which is the fifth instalment in the Harry Potter film series, was written by Michael Goldenberg and produced by David Heyman and David Barron.\n\nThe film stars Daniel Radcliffe as Harry Potter, alongside Rupert Grint and Emma Watson as Harry\u2019s best friends Ron Weasley and Hermione Granger.\n\nThe story follows Harry\u2019s fifth year at Hogwarts as the Ministry of Magic is in denial of Lord Voldemort\u2019s return. Harry must uncover the truth, with the help of his friends, and confront Voldemort.\n\nThe film began production in early 2007 and was released in theatres in the United Kingdom and United States on 12 July 2007, grossing a total of $292 million at the worldwide box office.\n\nThe film was nominated for two BAFTA Film Awards in 2008.\n\nFollowing a Harry Potter fan\u2019s dream that Harry\u2019s late headmaster Albus Dumbledore is alive, and in a critical condition at the Ministry of Magic, Harry Potter and his friends Ron Weasley and Hermione Granger, decide to rescue him, as the school year comes to a close.\n\nOn the night of their attempt to break into the Ministry, Ministry of Magic employee Delores Umbridge slashes Rubeus Hagrid\u2019s hand with a knife, accusing him of stealing her kitten.\n\nHarry, Ron, and Hermione fight the Death Eaters. Lucius Malfoy and his son Draco both try to attack Harry, but he manages to save himself. Harry, Ron and Hermione get separated, and in the chaos, Ron manages to grab the prophecy from the Ministry\u2019s Hall of Prophecies.\n\nHarry and Hermione escape the Ministry, and Ron awakes to see Harry and Hermione\u2019s signatures on the prophecy, but does not know that Harry had ripped the prophecy out and left it in the hands of his godfather, Sirius Black, before he was killed.\n\nSirius, Harry, and Hermione head to the headquarters of the Order of the Phoenix, where they meet with fellow members Remus Lupin and Nymphadora Tonks, and Sirius\u2019s brother, Regulus.\n\nThe Order is joined by the regular adult wizards and witches in the battle, as Voldemort and his Death Eaters take over the Ministry of Magic and declare martial law on the entire Wizarding community.\n\nAlbus Dumbledore appears to die in battle, but this is revealed to be a ruse, as he and Severus Snape attack Voldemort and Lucius Malfoy, and attempt to take the prophecy from Ron. Lucius disarms Dumbledore, and an enraged Bellatrix kills him.\n\nBellatrix and Narcissa Malfoy stand by their husbands\u2019 sides. Albus Dumbledore\u2019s portrait is replaced with one of Snape. Harry, Ron, and Hermione are rescued by Order members. Voldemort and the Death Eaters leave the Ministry.\n\nHarry, Ron, and Hermione are told by Remus Lupin that all of the other Order members have been killed. Harry attempts to repair his damaged wand. He is visited by Snape, who tells Harry that the boy\u2019s wand is irreparably damaged and needs to be destroyed, as Voldemort has become aware that the wand is of a \u201cunique\u201d nature.\n\nThe two engage in a fierce duel in which Snape calls on his master to save him. Harry is unaffected by the curse due to his ability to cast a shield charm. He manages to shield himself and fight back, and in his distraction, Snape accidentally breaks his neck and dies.\n\nHarry meets with Dumbledore\u2019s portrait, who reveals to Harry that the boy\u2019s mother died to save him, and Harry is filled with his mother\u2019s love. Harry reveals that he feels angry and confused at this revelation.\n\nHarry also learns that the reason that Voldemort has gone to such great lengths to kill him is that Harry is a Horcrux, a piece of Voldemort\u2019s soul that resides in Harry\u2019s body.\n\nHarry then leaves Dumbledore\u2019s portrait and meets with Ron and Hermione, who have just been told that the students and staff are to evacuate the school, as the Death Eaters have declared that Hogwarts is no longer safe.\n\nHarry, Ron, and Hermione are joined by Remus Lupin, Nymphadora Tonks, and several other members of the Order of the Phoenix, and they apparate to 12 Grimmauld Place, the headquarters of the Order.\n\nThe Battle of the Department of Mysteries is followed by the Death Eaters taking over the Ministry of Magic, as they believe the prophecy to be a key to Voldemort\u2019s victory. Harry returns to Hogwarts to retrieve the prophecy from Voldemort.\n\nHe, Ron, and Hermione get ambushed by Death Eaters, and when Harry goes after the prophecies, he ends up dueling with Voldemort.\n\nHarry calls upon the remaining strength of his mother\u2019s love, and as Voldemort\u2019s Killing Curse rebounds upon him and hits Harry, the battle is ended, and the school is cleared of the Death Eaters.\n\nHarry wakes up to find Dumbledore, Sirius, and Remus in the hospital wing, as well as his friends and the rest of the school, and he realizes that he is safe.\n\nDumbledore then proceeds to give Harry his restored invisibility cloak, explaining to Harry that the Invisibility Cloak is one of the Deathly Hallows. Harry then says goodbye to his friends, as he leaves with Dumbledore to destroy the seven horcruxes that contain fragments of Voldemort\u2019s soul, which will make him immortal.\n\nThe film ends with Harry and Dumbledore leaving the school in a car, accompanied by Professor Minerva McGonagall.\n\nI liked this movie as it is full of action and adventure. The plot is great as well as the dialogue. It is a well made movie and it is very entertaining.\n\nThis movie is definitely a must-watch, as it has plenty of action as well as being very humorous.\n\nAdvertisements]" time="0.070"><properties><property name="score" value="0.07572890823646267" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07572891&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07572891
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Hush, little baby, don't you cry\n\nI know a thing or two that will make you happy,\n\nIf you will only come with me.\n\n\n\nHush, little baby, don't you cry\n\nI know a thing or two that will make you happy,\n\nIf you will only come with me.\n\n\n\nIn my bag there's a beautiful surprise,\n\nIf you will only come with me.\n\n\n\nOn my back I'll show you the sights\n\nThere's a rainbow in your eyes,\n\nThere's a sweet, sweet lullaby\n\nIn your mother's smile.\n\n\n\nOn my back I'll show you the sights\n\nThere's a rainbow in your eyes,\n\nThere's a sweet, sweet lullaby\n\nIn your mother's smile.\n\n\n\nHush, little baby, don't you cry\n\nI know a thing or two that will make you happy,\n\nIf you will only come with me.\n\n\n\nIn my bag there's a beautiful surprise,\n\nIf you will only come with me.\n\n\n\nOn my back I'll show you the sights\n\nThere's a rainbow in your eyes,\n\nThere's a sweet, sweet lullaby\n\nIn your mother's smile.\n\n\n\nIn my bag there's a beautiful surprise,\n\nIf you will only come with me.\n\n\n\nIf you will only come with me.\n\n\n\nOn my back I'll show you the sights\n\nThere's a rainbow in your eyes,\n\nThere's a sweet, sweet lullaby\n\nIn your mother's smile.\n\n\n\nIf you will only come with me.]" time="0.013"><properties><property name="score" value="0.29898829805766747" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Mileage : 19,038 (77,223)\n\n: 19,038 (77,223) Transmission : 4-Speed Automatic\n\n: 4-Speed Automatic Exterior Color : Green\n\n: Green Interior Color : Light Blue\n\n: Light Blue Engine : 351\n\n: 351 VIN : #B9QWL214742\n\n: #B9QWL214742 Stock Number : 96988\n\n: 96988 Fuel: Gasoline\n\n1954 Ford Crown Victoria\n\nThis Ford is a very nice car, we are selling it at a fraction of the cost of building it! The motor is a very desirable 351 V8 and is equipped with a nice sounding Holley 650 carburetor. It runs well and has plenty of power. The trans is an automatic C4 and shifts smooth. This Ford is in great shape, the only exterior rust that I could see was on the door bottoms and they are easy to replace. Interior is in very nice shape, new headliner, new seats, new carpet and a nice newer steering wheel. It has power brakes, power steering, air conditioning and power top. These cars are extremely rare to find in this condition and price. If your looking for a car that is cool and nice to drive then this is it.]" time="0.014"><properties><property name="score" value="0.07899833137424705" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07899833&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07899833
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[I had a miserable time, once, when I was hospitalized. In the state of euphoria that was my hospital experience, the only one I could really think of to complain about was that I missed my window of opportunity to go on that uncharacteristic mission to munch fresh snow.\n\n\u201cYou mean\u2026EVERYONE doesn\u2019t get a chance to go on a snow mission,\u201d my mother replied. \u201cYou have to have the proper mindset for a snow mission. You have to be excited about it and you have to want to do it. If you]" time="0.013"><properties><property name="score" value="0.02531699838963508" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.025317&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.025317
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Create our team or join our team.\n\nIf you are an organization, brand, or personality and you want to take your show to the next level then get in touch! Email us at team@geekarmy.com and we will get right back to you.\n\nCreate a new Team\n\nSimply sign in with your username and password. If you do not have an account you can create one here.\n\nAfter signing in you will be able to create a new Team. You will be able to add a short name, a full name, a URL and a description. You will also be able to upload a logo for your team and a picture for the team members.\n\nEdit an existing Team\n\nSimply sign in with your username and password. If you do not have an account you can create one here.\n\nSelect the team you wish to edit from the My Teams page. From there you will be able to edit the team name, upload a new logo, a picture for the team members, and a URL and description. You will also be able to add members to your team.\n\nSearch for a team\n\nTo find a team click on the Teams Tab and you will be able to search by a number of different criteria including team name, brand name, member count, and popularity.\n\nAfter you find a team you are interested in, simply click the button at the top of the page to see more information about the team.\n\nJoin a team\n\nOnce you find a team you are interested in, simply click on the Join button and you will be taken to the join page. Here you will be able to apply to join the team.\n\nAfter you apply you will be sent a confirmation email. The email will contain a link for you to confirm your join request.\n\nAll of our applications are manually reviewed and you will be sent an email to notify you whether you have been approved or not.]" time="0.027"><properties><property name="score" value="0.08826558269938856" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[&quot;While some species are recognized as a single unit, they often consist of several distinct groups that have a common characteristic,&quot; Davis said. &quot;In this case, the characteristic is the presence of this particular steroid in all the species we looked at. If this steroid is a unique identifier for the group, we wanted to know if the species that contain it are closely related. We discovered that some species that produce the steroid are more closely related than others, and this pattern fits with our prediction that the ecological function of the steroid might be to keep different species from competing with one another for food.&quot;]" time="0.014"><properties><property name="score" value="0.002328865027584337" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[SEOUL, Oct. 30 (Korea Bizwire) \u2014 A shopping mall in southern Seoul opened on Wednesday with facial recognition technology installed in the washrooms for toilet paper dispensers.\n\nShoppers at the VOGUE shopping mall on Gwanghwamun in Seoul can use facial recognition to automatically dispense toilet paper, according to the mall\u2019s management company, which boasts that it\u2019s the first time that facial recognition technology has been applied to a toilet paper dispenser.\n\nThe facial recognition software automatically senses the number of sheets dispensed. It also allows the user to set an alarm to sound after a preset number of sheets has been dispensed.\n\nThe software is set to go on trial until November, when the management company will decide whether to expand its application to other toilet paper dispensers.\n\nFacial recognition technology is already being used in other places in Seoul, such as apartment complexes and an upscale golf course.\n\nBy Joseph Shin (jss539@koreabizwire.com)]" time="0.013"><properties><property name="score" value="0.012329021582263566" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01232902&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01232902
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[\n\n15% OFF ANY SINGLE ON EBAY WITH COUPON CODE &quot;PLAYTIME15&quot;\n\n(For US and Canada users only. Offer expires December 13th at 11:59 PM PT)\n\nError: You must have Javascript enabled in order to use this website.\n\nSkyscraper Property Text This card cannot be Special Summoned. Your opponent cannot select a Level 5 or higher monster on their side of the field as an attack target.\n\nTournament Eligibility TCG Advanced TCG Traditional OCG Unlimited Unlimited Unlimited\n\nPurchase Statistics (Last 30 Days) See More Number Purchased Money Spent 0 $0.00\n\nTournament Stats | Yugioh Wiki | Expand/Collapse All\n\nChecking Card Prices Please Wait\n\n(this might take a minute)]" time="0.016"><properties><property name="score" value="0.149568126226457" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.14956813&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.14956813
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[John Mayer Just Made A Fool Of Himself In An Interview With The New York Times\n\nBy James Wilson-Taylor\n\n&quot;We love to focus on my flaws and make them bigger than they are.&quot;\n\nJohn Mayer has come out fighting in a new interview with the New York Times, discussing a range of topics including the ghostwriting controversy, his upcoming tour with the Dead &amp; Company and the end of the line for the John Mayer Trio.\n\nThe article details a frank conversation with the singer, in which he shares his honest thoughts on everything from his troubled past, to his use of social media, to the over-analytical nature of modern day music criticism.\n\nThe interview was conducted as a two-day workshop with NYU's music journalism department, which of course meant that the whole thing was promptly picked over by the internet on Tuesday night.\n\nHere's a selection of the choicest cuts:\n\nHis Vocal Fry Controversy:\n\n&quot;The most controversial thing I\u2019ve done recently is affect a slightly looser, lower register on some songs, which I think makes the songs more interesting.&quot;\n\nOn Taking Himself Less Seriously:\n\n&quot;I remember thinking I have a platform and I have a microphone and I have a position of power, so let me try to change some things, maybe for the better. And people did not like that. I remember having this conversation where somebody said to me, \u2018When you are perceived as yourself, people have a hard time relating to that.\u2019 And I said, \u2018I think you are misperceiving yourself to be someone who has a choice.\u2019 I\u2019m not saying this like, \u2018Hey, world, my message is this, and you\u2019re going to like it.\u2019 That\u2019s not what it\u2019s about. It\u2019s about the moment in which you are caught off guard, and you\u2019re on tour and you\u2019re not doing what you\u2019re supposed to be doing, but you\u2019re still getting a check.&quot;\n\nOn The Benefits Of Ghostwriting:\n\n&quot;I don\u2019t have any beef with the world. I\u2019ve got a bunch of kids and a beautiful girlfriend and a house that I love and that\u2019s not an accident. And I\u2019ve worked for that and I\u2019ve gotten that because I\u2019m an easygoing guy who really enjoys life. And I\u2019ve always been super-vigilant about my ability to access joy.&quot;\n\nOn Being A 'Bubblegum Pop Star':\n\n&quot;I was, for a very brief period of time, a bubblegum pop star. I was in the No. 1 pop band of the late-\u201990s, early 2000s. I\u2019ve had 18 or 19 top 10 songs. I have the right to call myself a singer, I guess. But I would never want to turn my back on being a guitar player and a songwriter. And the reason why is because I have a fan base that loves me for being a guitar player and a songwriter, and it\u2019s very separate from the people who like my pop songs.&quot;\n\nHis Concert Review Regrets:\n\n&quot;I was like, \u2018Look, we\u2019re going to be under these bright lights, and it\u2019s going to be this one moment in time where we\u2019re going to play it for you. And we want to give you the most you can get for your dollar.\u2019 And then I think I had a bad couple of years. And I lost myself and tried to get myself back, and now I\u2019m playing shows where, like, I\u2019m not having a good time.&quot;\n\nOn Live Music Criticism:\n\n&quot;I think we\u2019re overintellectualizing music. The discussion about music is so myopic and minuscule. \u2018He did this, and it\u2019s a disservice to this particular song.\u2019 And I\u2019m like, well, maybe you don\u2019t like this song. If you don\u2019t like it, just don\u2019t buy it. But that\u2019s not the way it works anymore. So I think we\u2019ve been eating the meat and spitting out the bones. You just want to get the marrow.&quot;\n\nThe New York Times concludes by highlighting the vocal coaching he undertook in 2015 in order to overcome his vocal issues:\n\n&quot;I went and saw a vocal coach who I had seen when I was on my 'John Mayer's Journey Continues' world tour. She was a big hero of mine. She\u2019s the one who said to me, 'John, you just don\u2019t like how you sound. Why don\u2019t you do something about it?' I said, 'I can\u2019t afford to do anything about it.\u2019 She said, 'That\u2019s not true.\u2019 &quot;\n\nWatch our video of John's recent live cover of 'A Hard Day's Night' here:]" time="0.066"><properties><property name="score" value="0.0837766351630589" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08377664&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08377664
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[From the outside, the Taurus is a boring, bland looking sedan that's unapologetically designed to appeal to people who are price- and value-conscious, but still want to be able to put some stuff in the trunk. But it has a lot going on under the hood, and it's a great way]" time="0.013"><properties><property name="score" value="0.022065149456340638" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02206515&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02206515
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[COLD SPRING, Ky. (WKYT) - The Kentucky Department of Fish and Wildlife Resources says an investigation into the shooting of a horse at the Country Pleasures horse farm has been completed.\n\nAccording to KDFWR, a citation has been issued to a person involved in the incident. The citation will be mailed to the person's last known address. The citation is for the willful waste of wildlife.\n\nThe investigation by KDFWR has revealed that on June 11th, 2013, three horses were shot on a farm off of Logan Road in Bath County. One horse, a white and gray Paint Mare was found dead. The mare, named &quot;Sissy,&quot; had a history of being aggressive towards other horses. A chest wound indicated the horse had been shot at close range.\n\nThe second horse, a dark brown and white mare named &quot;Bonnie,&quot; had been shot in the rump. Bonnie was treated by a veterinarian for the wound, and released. The third horse, a dark brown and white gelding named &quot;Ghost,&quot; had a wound on his shoulder. He was treated by a veterinarian and released.\n\nKDFWR says the investigation concluded that the shooting of the horses was accidental.\n\nKDFWR says no additional charges will be filed in this case.]" time="0.013"><properties><property name="score" value="0.023903030364530453" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02390303&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02390303
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Please enable Javascript to watch this video\n\nST. CHARLES, MO (KPLR) - The now former police chief of St. Charles has been arrested for drunk driving. A spokesperson for the St. Charles Police Department says Officer Robert Hertzell was on duty and was driving a marked police vehicle.\n\nSt. Charles Police Chief David Kinkead has put Hertzell on unpaid administrative leave and has initiated an internal investigation.\n\nSt. Charles Mayor Sally Faith says Hertzell will remain on unpaid leave until the investigation is complete.\n\nChief Kinkead released a statement saying, \u201cI was very disappointed to learn about this incident. As a police chief, I hold myself and my command staff to the highest standard. I expect my officers to be role models and to demonstrate professionalism and respect. We have a lot of hard working men and women in the St. Charles Police Department and they are to be commended for their work on a daily basis. The actions of one officer, however, is not reflective of the entire command staff or the great work that our officers perform each day. We must continue to move forward with the work that we have been doing to create a safe and vibrant community for all of our residents and visitors.\u201d\n\nHertzell has been with the St. Charles Police Department since 1995.]" time="0.013"><properties><property name="score" value="0.02310414695346208" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02310415&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02310415
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[What happened\n\nShares of Applied Optoelectronics Inc (NASDAQ:AAOI) were getting hammered on Thursday after the networking specialist reported a quarterly loss and revenue that missed estimates. The news sent the stock down 15% at the time of this writing, putting it on pace to be the stock's worst day since late 2016.\n\nSo what\n\nInvestors were expecting Applied Optoelectronics to lose money in the fourth quarter, as it has lost money in each of the past two quarters. But the loss this time around was bigger than expected, with Applied Optoelectronics reporting a non-GAAP net loss of $0.40 per share on a 14% decline in revenue to $93.9 million. Analysts, on average, were only looking for a loss of $0.33 per share on $92.1 million in revenue.\n\nNow what\n\nCEO Paul E. Jacobs called the results a &quot;solid performance&quot; and said the company's overall sales were &quot;driven by market growth&quot; and its plan to &quot;increase long-term shareholder value.&quot; Applied Optoelectronics also noted that its total customer backlog rose 20% year over year to $347 million, which is more than analysts were looking for.\n\nWith the company's stock now trading down 45% over the past year, however, it will need to show much more than &quot;solid performance&quot; to prove that its turnaround is working. To start with, it needs to stop the losses and start producing profits again. If it can do that, then the rise in the backlog will be a positive sign, and investors may start to take a fresh look at the company again.]" time="0.014"><properties><property name="score" value="0.06188511363389715" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06188511&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06188511
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[These two teams have met eight times before with the Houston Dynamo winning five, the most recent meeting coming in last season\u2019s knockout round where Houston won 1-0 in Dallas.\n\nPROBABLE LINEUPS\n\nFC Dallas (4-4-2): Hartman; Loyd, John, Benitez, Morrow; Castillo, Jackson, Michel, Shea; Perez, Cooper.\n\n\n\nHouston Dynamo (4-3-3): Hall; Ashe, Taylor, Horst, Chabala; Moffat, Boswell, Creavalle; Bruin, Weaver, Davis.\n\n\n\nDID YOU KNOW?\n\nThis is the fourth time Dallas has made it to the semifinals. They've reached the final twice, losing to the New England Revolution in 2005 and D.C. United in 2007.\n\nDallas has won the U.S. Open Cup five times, the last coming in 1997. They are the only team in MLS with five Open Cup championships.\n\nDallas has reached the quarterfinals in three of the last four seasons.\n\nOnly one other team, the New England Revolution, has more postseason appearances (8) in the last eight seasons than the Dynamo.\n\nThe Dynamo\u2019s last win in the quarterfinals was over FC Dallas in 2008 when the Dynamo won 2-0 at Robertson Stadium.\n\nThe Dynamo\u2019s last two wins in the quarterfinals came on the road. They defeated Seattle Sounders FC 2-0 at Starfire Sports Complex in 2012, and the Portland Timbers 1-0 at PGE Park in 2011.\n\nForward Will Bruin has scored 12 goals in his last 18 matches in all competitions, including six in his last eight matches in MLS play.\n\nSuspended : None\n\n: None Suspended after next caution : None\n\n: None International duty : None\n\n: None Injury Report: None\n\nProjected Starting XI: Kevin Hartman (GK), Zach Loyd, Matt Hedges, George John, Michel, Jackson, Andrew Jacobson, Fabian Castillo, David Ferreira, Brek Shea, Kenny Cooper\n\nNotes: FC Dallas are undefeated in eight games at home this season. The last time they were shutout at home was against Real Salt Lake in May.]" time="0.024"><properties><property name="score" value="0.1305138903022188" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.13051389&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.13051389
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Sorcerous Origins\n\nUnbound Dancer\n\nUnbound Dancer\n\nSometimes called veils, a particular style of dancer that have sprung up in different cultures across the planes.\n\nCeaseless Dance\n\nAt 1st level you learn a ritual that allows you to spend 8 hours dancing and stamping out a particular pattern. This can be done during a short rest and while you do this you can spend up to half your sorcerer level (rounded up) in sorcery points.\n\nYou must finish a long rest before you can perform this dance again.\n\nWhen you complete the dance you gain the following benefits:\n\nYou have a base movement speed of 15 feet and your maximum movement speed increases by 5 feet.\n\nYou can add half your proficiency bonus (rounded up) to all dexterity ability checks.\n\nYou gain resistance to poison damage.\n\nYour skin becomes extremely smooth and soft and you have advantage on all charisma ability checks.\n\nYou have a +1 bonus to AC.\n\nThe following additional effects apply based on the number of sorcery points you spent on this ability:\n\n4-7 sorcery points: At the end of each of your turns you gain 5 feet of movement.\n\n8-11 sorcery points: You gain resistance to fire and cold damage.\n\n12-15 sorcery points: You gain a flying speed of 30 feet.\n\n16-19 sorcery points: You gain a second resistance to fire and cold damage.\n\n20 or more sorcery points: At the end of each of your turns you gain 10 feet of movement.\n\nPerform the Dance\n\nYou can only perform this dance once per long rest.\n\nStarting at 6th level you gain the ability to copy other creatures movements. You can use your action to study a creature within 5 feet of you for up to one minute. You can use your bonus action to then attempt to imitate that creature for one minute, using one of your known dance forms.\n\nYou must complete a long rest before you can use this ability again.\n\nExpanded Dance Forms\n\nAt 14th level you gain two additional dance forms. These can be from the expanded list on the next page. You can only perform each of these dances once per long rest.\n\nDance Forms\n\nDancing Viper: While dancing you can spend up to 4 sorcery points on the following effects:\n\n2 sorcery points: Gain a burrow speed of 20 feet.\n\n4 sorcery points: Gain a climbing speed of 30 feet.\n\n6 sorcery points: You gain resistance to piercing and poison damage.\n\n8 sorcery points: Gain a fly speed of 20 feet.\n\nDancing Tiger: While dancing you can spend up to 6 sorcery points on the following effects:\n\n4 sorcery points: You gain a climb speed of 30 feet.\n\n6 sorcery points: Your unarmed strikes are considered magical and gain the light property.\n\n8 sorcery points: You gain advantage on all dexterity ability checks.\n\n10 sorcery points: You gain resistance to slashing damage.\n\n12 sorcery points: Your movement speed increases by 10 feet.\n\nDancing Monkey: While dancing you can spend up to 6 sorcery points on the following effects:\n\n4 sorcery points: You gain a climb]" time="0.027"><properties><property name="score" value="0.06775742023460982" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[It's hard to imagine a wittier book than one subtitled A Primer on the Tactics of Scientific Research. But, then, David M. Kaplan and Robert M. May are not your ordinary scientists. At one time or another, they've all but single-handedly invented chaos theory, championed the concept of selfish genes, coined the phrase &quot;keystone species,&quot; and discovered the homeotic genes responsible for the formation of the vertebrate body plan. For sheer force of intellect, they're among the world's most eminent biologists.\n\nYou might, then, think that there would be little for these geniuses to learn from a primer in scientific methodology. But, as it turns out, that's just not so. In the same style of mordant humor that makes &quot;The Ant and the Peacock&quot; one of the best-selling books on evolutionary biology, Kaplan and May delightfully explain why it's difficult for a scientist to break away from deeply ingrained habits of thought.\n\nThey also introduce the problem of citation errors--a problem that has become all too apparent in the aftermath of The Bell Curve, whose authors committed a blizzard of them, citing everything from studies that were never performed to nonexistent journals. Kaplan and May suggest ways to prevent citation errors and to identify them when they do occur. They reveal that it's possible to fabricate a scientific journal, and to fool people into believing it's real.\n\nBut they don't confine their attention to egregious examples of scientific malfeasance. On the contrary, &quot;The Art of Scientific Research&quot; is written as a series of serious and not-so-serious instructions on how to conduct research, how to get it published, and how to cite it accurately. There's even a section on the art of collaboration, complete with sage advice on how to make it work. &quot;The Art of Scientific Research&quot; is a book for anyone who does science, or wants to.]" time="0.027"><properties><property name="score" value="0.07665006831655632" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07665007&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07665007
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[What I love about travelling is you get a chance to taste local food, pick up a language and meet new people.\n\nTravel is a great way to expand your cultural awareness and try new things.\n\nBe open-minded to the experiences.\n\nYou\u2019ll see a lot more on a tour, than just your average solo tourist.\n\nI love travelling as much as I love food, so naturally I\u2019m going to blog about them both.\n\nI\u2019ve created this blog to share some of my experiences, places, and reviews.\n\nHere you can find my best recommendations for travel destinations and hotels.\n\nI want you to enjoy your holiday as much as I did.\n\nSo, keep an eye out for my next travel blog.]" time="0.012"><properties><property name="score" value="0.031182610938879085" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03118261&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03118261
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Today, in the form of yet another unfortunate casualty of homelessness, a man\u2019s dead body was found on the street just outside of a McDonalds. Just a few hundred meters from the student dorms, where a guy was recently stabbed to death.\n\nI\u2019m really getting sick and tired of living in this city, and I\u2019m beginning to seriously think about moving somewhere else. A friend of mine has been living in Boston for the past 2 years, and he\u2019s constantly telling me how amazing Boston is, and how beautiful it is there.\n\nI\u2019m going to head to Boston for the summer, just so I can get a better idea of what it\u2019s like living there. I\u2019m really starting to lose interest in living here, and I want to try out living somewhere else before I settle on doing so. If Boston is not the place for me, I want to at least have an idea of where I should move.\n\nI do think, however, that Boston might be the place for me, and I\u2019m going to be there for the summer, and maybe even beyond. That is what I\u2019m hoping for at least, but we\u2019ll see how it goes. I think I\u2019m going to take a long trip back to India during this time, as well.\n\nI\u2019m going to try to take a lot of pictures, and post them up here on the blog. I will be a bit busy with work and school and my internship, so I don\u2019t know how much time I will have to do that, but I will do my best.]" time="0.013"><properties><property name="score" value="0.042702658817223016" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04270266&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04270266
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Our friends over at Road and Track have the scoop on what will be the new Mustang. You can read the entire article here, but it goes something like this.\n\nFord Mustang will have a larger, more powerful engine with a V8. A V6 will be offered, but as a two-seater.\n\nThe V8 will be a 5.0 liter, not a 3.7, and the V6 will have the same displacement as the current 3.7, but with the variable valve timing engine from the Ecoboost engine.\n\nIt will be called the Mustang GT, and is the car most of us are interested in.\n\nThe rear-drive GT model will be gone, replaced by an all-wheel drive GT.\n\nFord\u2019s CVT will be the only transmission.\n\nTo be honest, I\u2019m disappointed. The Mustang has always been about high performance.\n\nWhat is the reason for the V8 to lose power? Will we even feel the change? It is possible the V6 will make the most horsepower of all Mustang models, while still being smaller than the 5.0.\n\nDoes anyone want an all-wheel drive Mustang? Sure, you could argue that AWD will help with acceleration in the snow, but it\u2019s pretty rare for the Northeast to get that much snow, and I don\u2019t need all-wheel drive for an off-road trip to the Catskills.\n\nThe only reason I would consider buying an all-wheel drive Mustang is for the style. I think it will look very good.\n\nSo what do you guys think? Are you upset like I am, or are you okay with the changes?]" time="0.013"><properties><property name="score" value="0.06851422008007924" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06851422&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06851422
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[La Asamblea, organo pol\xedtico m\xe1ximo de la Administraci\xf3n Municipal, ejerce su funcionamiento en dos periodos:\n\na) Asamblea Ordinaria.\n\nb) Asamblea Extraordinaria.\n\nSe encarga de:\n\na) Intervenir en la conformaci\xf3n de las Autoridades de los Organismos Locales, ejerciendo la tutela de la Ley Org\xe1nica del R\xe9gimen Electoral y del Sufragio, Ley Org\xe1nica de los Partidos Pol\xedticos y dem\xe1s leyes y disposiciones aplicables.\n\nb) Establecer los niveles de las remuneraciones mensuales y de prestaciones de los servidores p\xfablicos municipales, por cualquier concepto.\n\nc) Aprobar el Plan Anual de Gobierno Municipal, el Presupuesto Municipal y su Reglamento.\n\nd) Resolver las quejas, reclamos y peticiones de los habitantes de la Municipalidad de Hualqui.\n\ne) Dictar las normas a que debe sujetarse la actividad administrativa del Municipio, en su ordenamiento jur\xeddico interno y en sus relaciones con otras Administraciones P\xfablicas.\n\nf) Elegir a sus integrantes y reglamentar su funcionamiento interno.\n\ng) Autorizar la emisi\xf3n de los t\xedtulos representativos de la deuda p\xfablica municipal.\n\nh) Dictar resoluciones administrativas que no impliquen la creaci\xf3n, modificaci\xf3n o supresi\xf3n de tributos municipales.\n\ni) Establecer el r\xe9gimen general para la designaci\xf3n de personas naturales o jur\xeddicas, como encargadas de la prestaci\xf3n de servicios p\xfablicos municipales, as\xed como el r\xe9gimen de contrataci\xf3n y supervisi\xf3n de dichas personas.\n\nj) Dictar resoluci\xf3n de acuerdo con el contenido de los proyectos de acuerdo presentados por la Contralor\xeda General del Municipio, as\xed como en el caso de la resoluci\xf3n de la auditor\xeda de las cuentas municipales.\n\nk) Dictar resoluciones que contengan las decisiones y pol\xedticas municipales, las cuales sean de trascendencia y sean definitivas, y que deban ser obedecidas por todos los habitantes del Municipio.\n\nl) Autorizar el otorgamiento de exenciones tributarias, por resoluci\xf3n motivada de la misma Asamblea.\n\nm) Dictar las normas sobre las sanciones a los concejales o autoridades de los organismos y empresas municipales que cometan irregularidades en la gesti\xf3n de la Municipalidad.\n\nn) Dictar resoluci\xf3n sobre las sanciones a los servidores p\xfablicos de la Municipalidad.\n\no) Elaborar el proyecto de Ordenanzas Municipales y otros reglamentos administrativos]" time="0.026"><properties><property name="score" value="0.12225131376778035" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.12225131&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.12225131
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Australian Energy Regulator highlights South Australian electricity network overspend\n\nThe Australian Energy Regulator has released a draft report highlighting the network cost of electricity distribution in South Australia. The report contains the results of an investigation into South Australian electricity distribution costs undertaken by the Australian Energy Regulator (AER).\n\nThe AER\u2019s report found that the electricity distribution networks in South Australia, namely ElectraNet and SA Power Networks, are materially overspending compared to the amount of revenue they receive from consumers.\n\nThe report states that network costs in South Australia have been growing rapidly, due to a range of factors, including but not limited to:\n\nLow and decreasing real term retail revenue\n\nPoor use of time based network costs\n\nTime based revenue regulation\n\nChanging end-user demand for electricity\n\nInvestments in the network that are not supported by regulation.\n\nThe report further states that distribution charges represent around a quarter of the total cost of electricity, and are ultimately passed on to consumers through higher electricity bills. The report notes that the three largest state owned electricity distribution companies (including ElectraNet and SA Power Networks) have all seen rapid increases in their distribution costs since 2013.\n\nThe report recommends that the South Australian Government take action to:\n\nUndertake an urgent review of electricity retail regulation\n\nUndertake an urgent review of the existing regulation of distribution charges\n\nExamine mechanisms to enable local distribution networks to invest in assets that improve reliability for customers, and facilitate the integration of renewable energy into the network\n\nThe report also recommends that the Australian Energy Regulator should seek to work with industry to seek additional reforms to be considered by the Australian Energy Market Commission (AEMC).\n\nThe report is open for submissions and can be found here.\n\nMore information on how energy networks operate and cost can be found here.\n\nSee our website for further information on energy and power regulation.]" time="0.029"><properties><property name="score" value="0.07464541040757054" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[A special thanks to all the people who have read this story and have offered me constructive criticism.\n\nA Big Thanks to The Prankster and Sage, also to Herr and themuseforlove for providing such helpful and encouraging critiques.\n\nEnjoy.\n\nHis life had been pretty good lately. The old man had finally started letting him do more things on his own, such as make deals with the city of Chicago.\n\nAt the moment he was in his office getting a briefing from Mr. Fredrickson, his assistant.\n\n&quot;The latest bounty is on something that is causing a lot of damage to the people in the city. It's probably best to explain the damage to you in person,&quot; Fredrickson suggested.\n\nThey both headed to the police station. Fredrickson had a theory as to what was causing the damage and had contacted the police.\n\nFredrickson's theory turned out to be correct. It was a monster, and it had been causing a lot of damage and injuries, but they were on the right track to capturing the monster.\n\n&quot;You mean that thing is going to be mine,&quot; the old man said when he got word that the monster had been captured.\n\n&quot;Yes sir, we are on our way to get it now,&quot; the officer informed him.\n\n&quot;I'll meet you at the stadium,&quot; the old man said.\n\n&quot;You want to come to the stadium,&quot; Fredrickson asked.\n\n&quot;I do, I want to see this thing,&quot; the old man said with a grin.\n\n&quot;Why,&quot; Fredrickson asked.\n\n&quot;I just want to see it,&quot; the old man answered.\n\nAs the two walked back to the old man's car Fredrickson suggested he do something a little different with the bounty this time.\n\n&quot;This thing is more dangerous than the last two that were captured, why do you want to go see it,&quot; Fredrickson inquired.\n\n&quot;Let's just say that I want to make sure that the right person is getting the bounty,&quot; the old man answered.\n\n&quot;Fine, I'll call them and tell them to make sure to bring the creature in the arena,&quot; Fredrickson said.\n\n&quot;Good, call them now,&quot; the old man said.\n\nFredrickson called and told the team that they were bringing in the monster and it was going to be brought into the arena and they had to get the area ready.\n\nAfter the call was made, the old man and Fredrickson drove to the arena.\n\nWhen they got there the old man went to the back area and took off his suit and put on a set of referee's gear.\n\n&quot;You sure you want to do this,&quot; Fredrickson asked.\n\n&quot;I'm sure, now go call them and tell them we're ready,&quot; the old man ordered.\n\nFredrickson called and then the two waited for the team to show up.\n\nAfter a few minutes the team showed up with the monster in tow.\n\n&quot;It looks mean,&quot; Fredrickson said as the team pulled the creature into the arena.\n\n&quot;It does look mean,&quot; the old man replied.\n\nAs the two watched the creature in the arena the old man asked a few questions of the men.\n\n&quot;When did it show up here,&quot; he asked.\n\n&quot;It came in about six months ago,&quot; the leader of the team answered.\n\n&quot;How did you capture it,&quot; the old man asked.\n\n&quot;We don't know, we just heard a scream and when we went to check on it, it was already bound up,&quot; the leader answered.\n\n&quot;Do you have a picture of what the thing looks like,&quot; the old man inquired.\n\n&quot;Yes, we have a picture of it,&quot; the leader answered.\n\n&quot;May I see it,&quot; the old man asked.\n\nThe leader handed the old man a picture of the creature.\n\n&quot;This doesn't look like the same one,&quot; the old man said as he looked at the picture.\n\n&quot;We don't know, maybe this is a different one,&quot; the leader suggested.\n\n&quot;This looks more like the picture of the monster I fought a long time ago,&quot; the old man said as he handed the picture back.\n\n&quot;It's not the same thing, the one you fought was green and this one is yellow,&quot; the leader said.\n\n&quot;Yes, I know what I'm talking about,&quot; the old man said.\n\n&quot;If it's not the same thing then why are you concerned,&quot; the leader asked.\n\n&quot;This monster could be very dangerous, why haven't you done more research,&quot; the old man inquired.\n\n&quot;I don't know,&quot; the leader admitted.\n\n&quot;You need to study this thing, do some research and if you can't find anything, then maybe you should move it to another city,&quot; the old man suggested.\n\n&quot;I'll talk to my superiors,&quot; the leader said.\n\n&quot;Good,&quot; the old man replied.\n\n&quot;I'm sure that I can find more information if I can get this one back to my lab,&quot; Fredrickson said.\n\n&quot;I'm sure you can,&quot; the old man replied.\n\nThe team was about to move the creature into the arena to meet with their local hero when a man walked up and started talking to the leader.\n\nThe old man looked at the man and recognized him as a former wrestler, and a very dangerous one at that.\n\n&quot;Hey, that guy is the one who can beat this monster,&quot; the former wrestler said.\n\n&quot;How can you be sure,&quot; the leader asked.\n\n&quot;I can beat this monster because I once fought a monster like this one and he defeated it,&quot; the former wrestler explained.\n\n&quot;Why don't you do it,&quot; the leader suggested.\n\n&quot;Fine, but I want the same money that you were going to give the other guy,&quot; the former wrestler requested.\n\n&quot;Okay, let's get this over with,&quot; the leader said.\n\n&quot;Yes sir,&quot; the former wrestler said.\n\nThe former wrestler walked over to the monster and started talking to it. The monster got angry and threw a punch at him.\n\n&quot;It's angry,&quot; Fredrickson said.\n\n&quot;I can see that,&quot; the old man replied.\n\n&quot;If he can't beat it, we can always use it,&quot; Fredrickson said.\n\n&quot;Let's see what happens,&quot; the old man said.\n\nThe former wrestler dodged the punch and was able to get his arms around the creature's neck and was able to pin it to the ground.\n\n&quot;This could be bad,&quot; Fredrickson said.\n\n&quot;It's not going to be bad,&quot; the old man replied.\n\nThe former wrestler held the creature down and then twisted its head around until it snapped off.\n\n&quot;That's one dangerous man,&quot; Fredrickson said.\n\n&quot;He is, let's go talk to him,&quot; the old man said.\n\n&quot;Okay,&quot; Fredrickson said.\n\nThe two walked over to the former wrestler and talked to him.\n\n&quot;You're good, I want to talk to you,&quot; the old man said.\n\n&quot;I'm listening,&quot; the former wrestler said.\n\n&quot;What are you doing now,&quot; the old man asked.\n\n&quot;I'm working on the docks,&quot; the former wrestler answered.\n\n&quot;I have a friend that owns a gym, and I'm sure that I could get him to help you out,&quot; the old man suggested.\n\n&quot;If I go to this gym, who will I be fighting,&quot; the former wrestler asked.\n\n&quot;That depends on who you think is the right person,&quot; the old man answered.\n\n&quot;I know, but how do]" time="0.074"><properties><property name="score" value="0.015103343709797393" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[\u201cJurassic World\u201d is going bigger, with bigger stars and a bigger budget.\n\nJ.A. Bayona (\u201cA Monster Calls\u201d) is in final negotiations to direct the sequel to \u201cJurassic World,\u201d the hit dino tentpole that earned $1.67 billion worldwide last summer. Colin Trevorrow and Derek Connolly, who penned the script for \u201cJurassic World,\u201d penned the script for the sequel.\n\nChris Pratt and Bryce Dallas Howard are both expected to return, but not necessarily as the leads. Frank Marshall and Pat Crowley are producing the pic.\n\nWhile Bayona and Trevorrow are considered to be a package, the studio is trying to decide whether to retain Trevorrow as a producer or have Bayona work with an experienced producer. That will likely be decided in the next week, with a new writer possibly coming aboard, but for now Bayona is poised to make his English-language feature directing debut.\n\nPlot details are vague, but sources say the film will follow in the footsteps of 2015\u2019s \u201cJurassic World\u201d in focusing on the next generation of the dino franchise. The follow-up will be produced by Frank Marshall and Patrick Crowley. The movie is set for release on June 22, 2018.\n\nBayona broke out in Hollywood with his 2007 thriller \u201cThe Orphanage,\u201d and directed the ghost story \u201cThe Impossible,\u201d which earned star Naomi Watts an Oscar nomination.\n\nBayona is repped by WME, Anonymous Content and attorney Marielle Tenghe.\n\nDeadline Hollywood first reported the news.]" time="0.018"><properties><property name="score" value="0.12447215703103587" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.12447216&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.12447216
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Malcolm Butler played a game with a fractured left fibula in Super Bowl LII, according to a league source.\n\nAfter being carted off the field in the fourth quarter with the injury, the cornerback stayed in the game to finish the Patriots\u2019 win over the Eagles. Butler recorded a crucial interception of Eagles quarterback Nick Foles late in the fourth quarter, which helped the Patriots run out the final 3:21 to seal the victory.\n\nButler had been in noticeable pain after being injured while defending a throw to Eagles receiver Alshon Jeffery.\n\nHe missed practice Wednesday and Thursday last week because of an illness. Butler was not listed on the injury report this week. He practiced Friday, but sat out Sunday\u2019s walkthrough.\n\nThe cornerback was not made available for comment after the game.]" time="0.013"><properties><property name="score" value="0.056816112226532045" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05681611&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05681611
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[An 18-year-old man was killed in a crash at Interstate 4 and Kirkman Road in Orlando on Monday morning, officials said.\n\nNo other injuries were reported in the crash, which was reported at 9:24 a.m., said Lt. Robert Terrell of the Florida Highway Patrol.\n\nHe said Juan Bautista-Martinez of Orlando was driving a 1997 Nissan Maxima west on Kirkman and ran off the road into the median.\n\nThe car, which was heavily damaged, then went over the barrier and landed on its roof in the eastbound lanes of I-4, Terrell said.\n\nAll eastbound lanes were closed for a short time as the vehicle was removed.\n\nThis is a breaking news story. Check back for more information.\n\nglotan@orlandosentinel.com or 407-420-5774]" time="0.017"><properties><property name="score" value="0.0922553475562582" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.09225535&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.09225535
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Jules &quot;GhostOf&quot; Wang is the jungler for Team Dragon Knights.\n\nBiography\n\nJules \u201cGhostOf\u201d Wang hails from Nanjing, China. He was introduced to League of Legends by a friend in his last year of high school. From the moment he started playing he was hooked on the game and quickly moved into playing professionally.\n\n2015 Preseason\n\nOn November 24, 2014 it was announced that along with Cloud9 Tempest's LCS roster and Nientonsoh, they would be the main team at the new Cloud9 Challenger roster.[1] After that he played in some amateur tournaments along with the rest of the Cloud9 Challenger roster.\n\n2015 Season\n\nAfter playing for Cloud9 Challenger in the NACS Spring Qualifier, GhostOf joined Team Dragon Knights in February, when the team was invited to the Spring Season of NACS. They finished in third place in the regular season, with a 6-4 record. In the playoffs, they lost in the semifinals to eventual tournament winner Renegades, but won the third place match against Cloud9 Tempest to secure a spot in the summer promotion tournament. However, they were unable to advance to the summer split, as they lost to Team 8 in the first round of their promotion match-up.\n\nTeam Dragon Knights competed in the Summer Promotion Tournament, where they lost 3-0 to Team Fusion in the semifinals. The team also lost their placement matches against Team Coast, meaning that they would be playing in the 2016 NACS Spring Season.\n\n2016 Season\n\nAfter winning the NACS Summer Qualifier, Team Dragon Knights qualified for the 2016 NA LCS Spring Season. In the spring season, they posted a 6-12 record in the round robin and did not qualify for playoffs. However, with the disqualification of Renegades (due to visa issues), TDK received the top seed into the Summer Promotion Tournament, where they were eliminated in the first round by Team Dignitas.\n\nTrivia\n\nLoves to watch anime.\n\nTournament Results\n\nInterviews\n\nRedirects\n\nThe following pages redirect here. There are three types of pages that may appear on this list:\n\nNames formerly used by the player in competition\n\nNicknames or alternative spellings or capitalizations of the player's name\n\nCommon typos that are frequently searched for\n\nThe list is generated automatically. To request an addition to the list, you may use this form.\n\n\n\nNo results found]" time="0.025"><properties><property name="score" value="0.03248376542798437" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03248377&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03248377
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[To compare the electrochemistry of a one-step and a two-step biosynthetic pathway, the co-culture of methanogenic archaea and the acetogenic bacterium Butyrivibrio fibrisolvens was examined. Acetate production from glucose in batch culture was investigated, comparing cells from co-culture with pure B. fibrisolvens and pure Methanosarcina barkeri cells. After 22 h, glucose was depleted in pure M. barkeri cultures but acetate production was not complete. After 22 h, glucose was depleted and glucose uptake and acetate production were completely ceased in pure B. fibrisolvens cultures. When cells from the two pure cultures were co-cultured, the cells from M. barkeri appeared dead in the co-culture after approximately 10 h. Cell growth and acetate production occurred in the co-culture, but was lower than the growth rate and the acetate yield in pure B. fibrisolvens cultures. Pure M. barkeri cells grown with pure B. fibrisolvens cells exhibited no growth after 8 h, but some cells were still viable. However, after 18 h all M. barkeri cells were dead and the co-culture was dominated by pure B. fibrisolvens cells. The results indicated that co-culturing of M. barkeri cells with B. fibrisolvens was not an efficient way of acetate production. Acetate production was possible when pure B. fibrisolvens cells were first grown to a certain population and then M. barkeri cells were added. However, complete utilization of glucose was not possible in such co-culture. In this study, the electrode setup was not optimized to detect early anaerobic cells.]" time="0.023"><properties><property name="score" value="0.07990742939976044" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[If your personal philosophy is a little bit Taoist and a little bit Marxist, there's a new place in LA to scratch that particular philosophical itch. Tao Mao is the only East-meets-West neo-communist tiki bar in LA, a three-story Chinatown bordello of poi dogs and palm trees. To get there, you enter from a rear alley and walk past the washboard abs of dancers in the Bumpin' Uglies Go-Go Bar to the karaoke bar called Red 7, and from there climb a red-lit staircase to the restaurant.\n\nTao Mao, &quot;the great origin,&quot; is the Chinese name for the Milky Way galaxy, as seen from the Earth. But it's also the name of a star in that galaxy. And in Mandarin, Tao Mao is pronounced &quot;Dow Mao,&quot; which sounds like &quot;D.M.A.,&quot; or &quot;doctor of philosophy.&quot; And that is exactly what Tao Mao is\u2014a tiki bar where all the drinks have the names of philosophers.\n\nYou may be wondering: Is Tao Mao a tiki bar for philosophers? Or a bar for philosophers that serves tiki drinks? Or is it both? Does it matter?\n\nActually, no, it doesn't. Tao Mao is a philosophical tiki bar, that's all you need to know. The more you think about it, the less sense it makes.\n\nBefore I try to explain Tao Mao, I should mention that the bartender told me the bar's name is pronounced &quot;Dow Mao.&quot; And he is a white dude. And yet... all the cocktail names are in Mandarin. And he only speaks English. So what the fuck do I know? Maybe I'm the crazy one.\n\nI can tell you that the people who seem most at home at Tao Mao are little clusters of friends\u2014mainly white dudes who look like they'd be hanging out at some over-designed tiki bar that was actually inspired by Tao Mao. For example, I met two TV writers who were there for the first time. The owner was there with his mom. The first time I visited, the woman in the next booth to me was there with her sister, and]" time="0.024"><properties><property name="score" value="0.08164560352083665" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0816456&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0816456
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Considered a New York classic, the Fourth Avenue Deli has been a staple of the city\u2019s Lower East Side for decades. Deli aficionados and tourists alike have been lining up outside the restaurant\u2019s doors for more than 50 years, eagerly awaiting their first taste of Katz\u2019s famed pastrami sandwich. Katz\u2019s special blend of spices, smoked over pinecones in traditional ovens, and the deli\u2019s commitment to quality meat make the perfect recipe for mouthwatering pastrami. Every bite of the succulent, slow-cooked meat is an explosion of flavors.\n\nKatz\u2019s pastrami, made with fatty beef from the belly, is a little softer than that of the more common brisket, but has a much richer, beefier flavor. The deli also offers hot pastrami, which is much more spicy than its cold counterpart. But it\u2019s Katz\u2019s hearty slices of hand-carved corned beef, a delicacy with a history dating back to the 1700s, that has helped make the deli a household name.]" time="0.013"><properties><property name="score" value="0.0739566429963201" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07395664&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07395664
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[IceBerg Originally Posted by No she was not a &quot;sexy voice&quot; in any way, shape or form. I've heard many, many callers on various airwaves that were/are far more seductive and far more sexy sounding than this woman ever will be.\n\n\n\nWhat did you ever think was going to happen when you allowed &quot;a show host&quot; to be able to call into the show and be able to communicate with the host and be able to flirt with the host? I mean, that's pretty much all that happened. It didn't take a rocket scientist to see what was going to happen.\n\n\n\nAnyhow, whatever the circumstances were, I think it's pretty cool that she came back to apologize. It's pretty easy to judge when you're not in that situation.\n\n\n\nAnd to the person that said &quot;You have to look for and find bad things that can be found in anything to be unhappy&quot;, you're absolutely right, but it doesn't take a rocket scientist to look for the bad and find it.]" time="0.016"><properties><property name="score" value="0.044497790659209224" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04449779&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04449779
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Boeing to Cease Orders for Canadian 737 Max 8\n\nThe sudden grounding of Boeing\u2019s 737 Max aircraft is causing a major disruption to the global airline industry.\n\nThis morning, Canadian plane manufacturer Bombardier announced that it was ending its contract to sell 75 of its CSeries aircraft to the lessor, and parent company of Norwegian Air Shuttle.\n\nAnd now, it has been revealed that U.S. aerospace giant Boeing has cancelled its orders for the Canadian plane, putting a hold on any future deliveries.\n\nThe grounding of the Max model of the aircraft has thrown the airline industry into turmoil. While Boeing does not have a huge presence in Canada, it is estimated that the company has invested $6.4 billion into the country in the last two years alone.\n\nIn a statement, Boeing said it \u201ccancelled a previously announced purchase order for 16 Next-Generation 737-800 passenger aircraft from a U.S. customer and added a new order for 16 Next-Generation 737-900ER passenger aircraft from the same customer.\u201d\n\nIt added, \u201cthis change resulted in a reduction of eight 737-8 MAX deliveries from the previously announced 64 deliveries through 2023.\u201d\n\nWith the order having been cancelled, it is unclear if it will be restored in the future. Boeing said the cancellation of the order was part of its routine actions \u201cin response to market opportunities, customer preferences, and our results in a competitive environment.\u201d\n\nBoeing says it remains committed to its relationship with Canada, but the news comes amid concern over the safety of the aircraft. The model was involved in two deadly crashes in Indonesia and Ethiopia, where 346 people died. The cause of the crashes remains unclear, but both involved issues with the new anti-stall system.\n\nAfter the Ethiopian crash, U.S. President Donald Trump said the plane was \u201ccertainly\u201d unsafe to fly.\n\nFollowing the grounding of the aircraft, Boeing has grounded its 737 Max planes across the globe. In addition, countries and airlines have pulled their planes from the skies.\n\nCanadian airline WestJet has also removed its Max models from its fleet, as did Air Canada and Lufthansa.\n\n\u201cThe decision to temporarily remove our Max 8 fleet from service was made as a precaution and is consistent with actions by regulators around the world, including the FAA,\u201d said WestJet in a statement.\n\n\u201cWhile we have faith in the safety and airworthiness of our Max 8 aircraft, this temporary action is as a precaution while we await the findings from both the FAA and Boeing\u2019s ongoing investigations into the cause of the Ethiopian Airlines Flight 302 crash.\u201d\n\nFeatured Image: Pixabay]" time="0.024"><properties><property name="score" value="0.03007231320020296" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03007231&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03007231
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Image Credit: Angie B., Albuquerque, NM\n\nPeople feel alone when they don't have a purpose. They feel like they are nothing when they don't have something to do. They feel insecure because they feel like their life doesn't have a meaning. Without these feelings people will feel like they are a nobody. They will feel like their life is worthless. Without a purpose to live for it will be hard for people to be happy and be themselves. When people have no purpose to life they can be so unhappy. These feelings come up in people when they are very upset. People are less likely to be happy when they are alone. With being alone it makes it harder to have fun. Without fun in your life it makes it harder to have a positive attitude. With a positive attitude people will be able to live without fear. They will be able to live like they are in a good mood. With having a good mood people will not be sad and they will feel good about themselves. People will also feel happy when they have a good life. When people feel good about their life it will help people have a positive attitude. They will also be able to have a better life. It will be easier to have a better life without bad feelings. When people have a bad life it will make it harder for them to be happy. This is because their life is not the way they want it. It is very hard for people to be happy when they don't have the life they want. Without having the life they want they will not be happy. People who have a positive attitude will not feel so sad when their life isn't the way they want it. With having a positive attitude people will be able to feel happy. When they are happy it will help them to live a better life. With a better life they will be happier. It will be hard for people to be happy when their life isn't the way they want it. This is because people feel like their life is going wrong. When people feel like their life is going wrong they are less likely to be happy. People who are happy are more likely to be happy. When people are happy they will not be sad. People will not be sad when they are happy. With having a good attitude it will help people be happy. It will also help people to have a good life. People who have a good attitude will feel good about themselves. They will feel good about their life and will have a good life. With a good life they will feel good about themselves. People who feel good about themselves will not feel so sad. This is because they have a good life. They also have good friends. Having good friends will help people be happy. With having good friends people will be happier. People will be happier with good friends. They will also be happier with a good life. With a good life and good friends people will feel happy. It will be hard for people to be happy without a good life. With a good life people will be happier. It will also be hard for people to be happy without good friends. Without good friends people will feel less happy. This is because they have no one to talk to. People with a good life and good friends will feel more happy. It will be easy for people to be happy with a good life and good friends.]" time="0.064"><properties><property name="score" value="0.19040078357306164" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Court rules out taking blood samples for cell phone surveillance\n\nIn the United States, the Supreme Court ruled out taking blood samples for cell phone surveillance, the second verdict in just a few days.\n\nThe Supreme Court in Washington (USA) ruled that the police is not allowed to search a person's body for a cell phone's location and information.\n\nIt was decided in a 4-to-3 vote. This is the second verdict in a week in which the Supreme Court ruled that blood samples must not be taken from a person arrested for driving under the influence of alcohol.\n\nAccording to the relevant law, the police can take a blood sample for alcohol from an arrested person if there are grounds for suspicion. It also specifies that if a person refuses, they can be forcibly taken.\n\nA man was arrested for driving under the influence of alcohol in 2011. He was ordered to give blood samples. He refused, but the police took the samples forcibly. The court found out that the defendant was under the influence of alcohol. He appealed, but his appeal was rejected in 2012. He then appealed to the Supreme Court.\n\nIn a 4-to-3 vote, the Supreme Court overturned the verdict and ordered the verdict to be re-tried.\n\nIn the meantime, the verdicts are difficult for police. They can't take blood samples from suspects in alcohol-related accidents, nor take blood samples of suspected drug addicts, nor use DNA to find out the suspects in rape cases. They have to be very careful about searching people on the street.\n\nOther websites also commented on the news.]" time="0.023"><properties><property name="score" value="0.08129635216825519" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Define ridged. ridged synonyms, ridged pronunciation, ridged translation]" time="0.014"><properties><property name="score" value="0.006865202530004222" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0068652&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0068652
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Another Apple Exec Talks About Bringing Down The Cost Of Macs\n\nOn the heels of some recent remarks by Apple\u2019s Phil Schiller about the company\u2019s view on entry-level pricing, the current vice president of worldwide marketing, Phil Schiller, was quoted today in a Wall Street Journal article about PC prices.\n\nSchiller says that for the sake of keeping PC prices down, the platform needs to avoid fragmentation by maintaining one version of Windows that\u2019s supported by all vendors. \u201cThat helps the consumer know that if I buy a PC in the next six months, it\u2019s going to run that software. That\u2019s good for consumers, and it helps keep prices down,\u201d he told the newspaper.\n\nThose remarks are similar to what Schiller had to say about Apple\u2019s Mac sales earlier this month, when he told analysts during a conference call that the company\u2019s philosophy is to not \u201cgive customers a lot of choices.\u201d\n\nHe also said that Apple has taken a different approach to pricing compared to its competitors. \u201cIt\u2019s like when you go into a restaurant, you know what the food is going to cost. You don\u2019t have to go through five steps of saying I want the steak for this much money, the chicken for this much money, the pasta for this much money.\u201d\n\n\u201cWe\u2019re not trying to segment or isolate ourselves from anybody,\u201d Schiller said. \u201cIt\u2019s more like, if you want a Mac, this is the price. If you want a PC, go over here.\u201d\n\n[Source]]" time="0.014"><properties><property name="score" value="0.07906424859869254" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07906425&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07906425
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[OK, this is all you need to know about this whole &quot;Ryan Fitzpatrick's against the Browns&quot; narrative that is now in full force, after his dismal performance against the Browns last year, and has been growing in fervor ever since news of his re-signing broke last week:\n\nIt's not going to happen again.\n\nYou may not believe it now, and it may not be true on Sunday when the Browns take on the Bucs in their Week 1 matchup. But just like the Browns have a way of winning the games they're supposed to, Fitzpatrick has a way of having performances like he did against the Browns last year.\n\nIt just happens in reverse.\n\nThe defense was ranked 30th in the league in pass defense and would give up more yards in the first quarter than in the entire game the year before, yet it's Fitzpatrick who throws for 5,367 yards and 40 touchdowns that year. That's not a typo. And it's not because he was the quarterback of the future and Josh Gordon was the No. 1 receiver of the future, either. (We'll ignore the fact that the Browns traded for Jason Campbell at the end of the season to play quarterback and traded away Trent Richardson for basically nothing.)\n\nFitzpatrick and the Browns were matched up against each other in the third week of the 2012 season, and by that point in the season Fitzpatrick had thrown for 4,008 yards and 27 touchdowns in just 12 games. He finished the season with an astronomical 8.18 yards per attempt, which was the highest in NFL history. (Since 2001, at least.) He averaged two touchdowns per game for the first three weeks of that season before throwing just five more in the final nine games.\n\nAnd that was in his first season with the Bucs. They acquired him on the last day of August, after he'd spent the last few months being cut and re-signed by the Bills, which was the latest example of the Bills treating him like a starting quarterback without actually starting him. When the Bucs traded for him, it was widely considered a sign that they'd be moving on from Josh Freeman, but Fitzpatrick's job was only to hold the fort. He didn't even start his first game until the Bucs' eighth game of the season.\n\nThis time around, Fitzpatrick is more than holding the fort. The Bucs have put him on the short list of best quarterbacks in the NFL, with his name appearing on every single MVP ballot last year, and after a few years of analysts treating him like a highly replaceable backup, he has emerged as a bonafide starter. And like most No. 1 quarterbacks, he also seems to thrive against certain teams.\n\nIt seems like a good rule of thumb to get out of the way when he's playing his best.\n\nThe Bucs will almost definitely start out slow against the Browns on Sunday, but they're not going to go through another slow start and blow a lead against the Browns for a second time. It just won't happen.\n\nThis is a different year, a different Browns team, and a different quarterback. It's also a different Ryan Fitzpatrick.]" time="0.025"><properties><property name="score" value="0.06573769647708322" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0657377&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0657377
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Get daily updates directly to your inbox Subscribe Thank you for subscribing We have more newsletters Show me See our privacy notice Could not subscribe, try again later Invalid Email\n\nSunderland AFC striker Steven Fletcher has had his \xa320,000 Bentley vandalised and stolen, sparking a massive police hunt.\n\nThe \xa345,000 luxury motor, which was parked up outside his home, was spotted being driven away by an offender, believed to be male, wearing a black crash helmet and driving a silver Ford Mondeo.\n\nPolice say it was not a random attack and is being treated as a targeted theft.\n\nThe car has the registration number CN55 ANT and Sunderland AFC can confirm the vehicle belongs to Fletcher.\n\nThe vehicle is a black Bentley Continental GT Speed with blacked out windows.\n\nA spokeswoman for Northumbria Police said: \u201cWe received a report of a car theft in Sunderland on Monday night.\n\n\u201cThe vehicle was parked outside a house and was seen being driven away by a man, who was wearing a black crash helmet.\n\n\u201cThe vehicle was a Bentley Continental GT Speed with blacked out windows.\n\n\u201cThe car had the registration number CN55 ANT.\n\n\u201cThis was not a random incident. It was a targeted theft.\n\n\u201cWe are keen to hear from anyone who has any information about this incident, or who has been offered a similar vehicle for sale.\n\n\u201cWe are particularly keen to hear from anyone who saw the vehicle between 9.30pm and 10pm on Monday night.\u201d\n\nThe theft of Fletcher\u2019s Bentley follows the theft of Sunderland\u2019s team bus in 2008.\n\nThe club\u2019s coach was stolen from outside the Stadium of Light on the evening of the final game of the season against Fulham.\n\nA club spokesman said at the time: \u201cWe\u2019re concerned and disappointed that someone could stoop so low as to steal the team bus and all of the club\u2019s transport plans.\n\n\u201cThis is an act that should shame football fans.\n\n\u201cWe would appeal to whoever stole the bus to think again and drop it back to us.\u201d\n\nAnyone with information is asked to contact Northumbria Police on 101 ext 69191.]" time="0.024"><properties><property name="score" value="0.01674977333243506" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01674977&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01674977
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Another episode of Tetsuo: The Iron Man! It\u2019s getting there. I will be enjoying this episode a bit less than the previous few because the creators seem to be in a big rush to make the series come to an end. Well, at least they will finish the story. At any rate, enjoy the recap!\n\nRead the rest of this entry \u2192]" time="0.012"><properties><property name="score" value="0.025017727536741363" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02501773&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02501773
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Cross-posted at The Huffington Post.\n\nThe recent announcement that, for the first time in three decades, a new strain of avian influenza is spreading from one species to another is concerning but not, as some news outlets have claimed, evidence of a \u201cfatal pandemic in the making.\u201d A strain of H5N1 avian influenza (called H5N1-V) was detected in two apparently healthy wild ducks found in Germany. These ducks, according to the World Health Organization, most likely caught the virus from chickens, because H5N1-V has not previously been detected in wild birds, and H5N1 has not previously been found circulating among wild birds anywhere in the world. The two ducks apparently had not recently come into contact with any chickens. The H5N1-V strain found in the ducks is one of at least 14 genetically distinct H5N1 subtypes. While the H5N1-V strain can infect chickens, it cannot efficiently spread among them or to other species.\n\nH5N1 is an avian influenza virus, meaning that it naturally infects birds. Over the past 14 years, H5N1 has caused at least 555 infections and 282 deaths in humans. While those numbers are higher than any other human infection since influenza was first identified in humans in 1918, they are still a small fraction of the estimated 5 billion to 8 billion birds that have been infected with H5N1 over the past decade. The World Health Organization says that \u201cit is difficult to estimate accurately the number of cases that have gone undetected.\u201d Since the virus has spread to so many different countries, it is likely that many cases have gone undetected. Nevertheless, the case-fatality rate (defined as the percentage of people who die from a given disease) from H5N1 in humans is estimated at 60 percent. It is difficult to compare H5N1\u2019s case-fatality rate to those of other viruses. The case-fatality rate from the H1N1 influenza virus that caused the 1918 influenza pandemic, which killed an estimated 50 million to 100 million people worldwide, was only 2.5 percent, yet that virus killed about 2 percent of the world\u2019s population.\n\nGiven the high death toll, it is important to find the best ways to limit the spread of H5N]" time="0.023"><properties><property name="score" value="0.011124269187662813" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01112427&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01112427
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[As word spread of the new church\u2019s intentions, many Muslims came to have a look. \u201cThey were surprised to see Christians.\u201d\n\nKoberg said he was surprised to see the modest construction was built as quickly as it was.\n\n\u201cWe have enough of everything we need. We have enough food, we have shelter, and there is security. We\u2019ve built what we could,\u201d said Mohammed. \u201cThey are very good, very nice. We\u2019re working together.\u201d\n\nBut Muslim leaders in the area are concerned that the church is being built on a spot where a mosque stood for about 80 years.\n\nEarlier this year, tensions between Christians and Muslims flared up when a Catholic hospital in the nearby city of Bauchi shut down its maternity ward, arguing that the government did not give it enough money to operate. As Muslims felt that the ward\u2019s closing was unfair, they attacked the hospital, which is run by the Italian-based congregation of Daughters of Mary Immaculate.\n\nKoberg said he was surprised that no one has reacted to the construction of the church so far.\n\n\u201cNobody has said anything.\u201d\n\nThe Protestants in Maiduguri had been looking for a new place to worship for months. They are not the only Christians in town, but they are the only ones without a church. Now they can freely worship their God.]" time="0.012"><properties><property name="score" value="0.061954447591378026" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06195445&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06195445
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Bungie has confirmed a new competitive mode for Destiny 2, with a new ranking system and playlists.\n\nFollowing the release of the Curse of Osiris expansion, Destiny 2 will receive some new updates to the Crucible, as well as the addition of private matches for the first time. In a recent blog post on Bungie's website, it was confirmed that Crucible rank is returning to Destiny 2, and that it'll be more straightforward than it was before.\n\n&quot;We believe that to earn your highest rewards, you should be playing the most challenging activities possible for your skill level,&quot; the post reads. &quot;With that in mind, we\u2019re making the following changes to the player investment required to reach your highest rank:\n\n- Strike-specific rewards (Medallions, Boots, Class Items, etc.) will no longer be tied to your highest achieved rank\n\n- You will be required to reach a specific rank (HM5) to earn the Nightfall-specific rewards\n\n- You will be required to reach a specific rank (HM5) to earn the Trials of the Nine reward weapons\n\n&quot;In addition to the above changes, we\u2019re also increasing the overall player investment in the Crucible by requiring a minimum Light of 251 to participate in the most challenging endgame activities.&quot;\n\nIn the same post, Bungie has confirmed a new ranking system and matchmaking for the Crucible, which will include new playlists.\n\n&quot;To celebrate our launch of private matches and the new matchmaking system, we\u2019re implementing a new system for ranking up in the Crucible. The highest rank you can achieve in Competitive play is now \u201cLegend,\u201d and your Glory will be available to view on our new Competitive playlist HUD. Glory earned in the Competitive playlist will be used to rank up within this new system. We\u2019ll continue to expand on these improvements in the coming months.&quot;\n\nIt's interesting that the new rank system is being implemented for just the Competitive playlist, as well as the fact that your Glory is limited to a specific playlist. Hopefully, this doesn't mean there will be no more way to rank up in the Crucible outside of Competitive.\n\nAs part of the changes to the Crucible, Bungie is adding Private Matches to Destiny 2, which will be coming with the Curse of Osiris expansion in December.\n\n&quot;If you\u2019ve played our April Update, you\u2019ve had a taste of what a competitive, matchmade experience can be like in Destiny 2. Now, we\u2019re ready to give you the power to host your own private Crucible matches. This new feature will become available in the next update.&quot;\n\nPrivate matches can be played with two to twelve players, and will have matchmaking, as well as private lobbies. There will be a wide range of options available to players, with &quot;full playlist and customization support&quot;. Private matches will also come with their own rewards.\n\nPrivate matches will be available to all players who own Destiny 2 and have access to Curse of Osiris. You'll be able to make them as accessible as you like, whether it's making it a full-scale party affair, or something more intimate. Whatever your needs are, we\u2019re looking forward to you showing us what you\u2019re made of in your own custom arenas.&quot;\n\nDestiny 2 is available now on Xbox One and PlayStation 4.\n\nSource: Bungie]" time="0.062"><properties><property name="score" value="0.022333777191883647" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Approximate size 4.75\u201d x 3.5\u201d\n\nUse both sides\n\nHeavy-duty material for superior durability\n\nProtect your firearms and valuables\n\nMade in USA\n\nColor \u2013 Black\n\nUse both sides to take notes or hold your gun with a large holster. Black finish. Approximate size 4.75\u201d x 3.5\u201d. Heavy-duty material for superior durability. Made in USA.]" time="0.013"><properties><property name="score" value="0.0054343050854333" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00543431&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00543431
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[My wife is a big fan of using oils in her hair. So, I bought a couple different ones to try and compare. This is one of them. It smells very good and it makes her hair nice and soft. If you want a lot of shine, this one is not going to do it. But if you want a lot of softness, this is your oil.]" time="0.012"><properties><property name="score" value="0.014174394547172842" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01417439&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01417439
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Dear Reader, As you can imagine, more people are reading The Jerusalem Post than ever before. Nevertheless, traditional business models are no longer sustainable and high-quality publications, like ours, are being forced to look for new ways to keep going. Unlike many other news organizations, we have not put up a paywall. We want to keep our journalism open and accessible and be able to keep providing you with news and analyses from the frontlines of Israel, the Middle East and the Jewish World.\n\nThe proposal for a binational state on Israel\u2019s future borders with the Palestinians, put forth by outgoing Prime Minister Ehud Olmert, is no longer an acceptable solution, former Knesset speaker Avraham Burg said Wednesday.\n\n\n\nSpeaking in an interview with Channel 10, Burg said that the proposal, which the Olmert presented last week to Palestinian Authority President Mahmoud Abbas, was \u201cna\xefve and counterproductive.\u201d\n\n\n\n\n\nJPOST VIDEOS THAT MIGHT INTEREST YOU:\n\n\u201cIsraelis don\u2019t want to live in a binational state. One can\u2019t create a state of Jews and Arabs, and not even a state that is a nation of all its citizens,\u201d Burg said. \u201cI don\u2019t see any logic in it, even less so now than five years ago.\u201dThe former Kadima party leader argued that a similar situation exists in Ireland, with the Irish living alongside the English, but not in a single state.Similarly, he said, one could not build a future Palestinian state within the borders of Israel, but one can build a state of all its citizens in the West Bank, Gaza Strip and eastern Jerusalem.If the Israeli government doesn\u2019t agree to this, he warned, the two-state solution is no longer attainable.\u201cWhat we need is an understanding that [an agreement with the Palestinians] will be built on borders and not on a state of all its citizens. We\u2019ll build two states on the 1967 borders, one Jewish and one Palestinian,\u201d Burg said.\u201cIsraelis need to know that [in the future] there will be two states,\u201d he continued. \u201cThe Palestinian people needs to know that [in the future] there will be a Palestinian state.\u201dThe two-state solution is the only way to prevent the State of Israel from becoming the State of the Jewish people, the Knesset speaker warned. \u201cThe future of the State of Israel depends on the existence of two states,\u201d he said.\u201cI\u2019m for a binational state only as a stage, not as a solution,\u201d he added. \u201cIn a binational state, we will never live together, but there will be an agreement. If there\u2019s no agreement, there will be a civil war. We have to prepare for a binational state as a transition stage to the two-state solution.\u201dWhen asked about the present leadership\u2019s decision to allow MK Otniel Schneller (Kadima) to bring a vote of no-confidence in Defense Minister Ehud Barak, Burg accused Barak of \u201centhusiastic nationalism\u201d and of being \u201cna\xefve.\u201d\n\nJoin Jerusalem Post Premium Plus now for just $5 and upgrade your experience with an ads-free website and exclusive content. Click here&gt;&gt;\n\n\n\n]" time="0.027"><properties><property name="score" value="0.0634853516752982" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06348535&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06348535
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[JOYNER LUCAS \u2013 JOYNER\u2019S WORLD \u2013 PREMIERE\n\n#JOYNERSWORLD PREMIERE:\n\n\n\nI\u2019m telling you right now, JOYNER LUCAS has been fucking with his album so hard. That\u2019s right, his album. What I mean is that we know about his movement, but his album really stands for something. His \u201cJoyner\u2019s World\u201d is about the transition to adulthood. It\u2019s about being a young artist making waves on the charts and the fan support he\u2019s getting. The song \u201cI\u2019m Not Racist\u201d really gives off the vibe of Joyner\u2019s]" time="0.013"><properties><property name="score" value="0.03266581231470833" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03266581&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03266581
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Yet another reason why we love Mary Tyler Moore: The Golden Girl will donate the proceeds of her upcoming appearance on Dancing With the Stars to St. Jude Children's Research Hospital.\n\nMoore's reps confirmed that the donation will come from the profits of Moore's Dancing With the Stars' appearance, in which she and professional dance partner Tristan MacManus perform a cha-cha to Donna Summer's &quot;Hot Stuff.&quot;\n\nWatch Mary Tyler Moore's 'Hot Stuff' Dancing With the Stars Performance\n\n&quot;Mary Tyler Moore and her team have been supportive of St. Jude Children's Research Hospital for many years and she will be donating a portion of her proceeds from Dancing with the Stars to support St. Jude,&quot; the charity said in a statement to E! News. &quot;Her support is critical to the hospital and we are grateful for her help.&quot;\n\nMoore is competing on the show's 12th season with former General Hospital star Ingo Rademacher. Last night, Moore and Rademacher scored the first 10 of the season from judge Len Goodman.\n\nWhile Moore will donate her Dancing With the Stars appearance, some celebrities don't -- even though that's what's requested. They still get a share of the show's profits from voting. Last year, Katie Holmes said she wouldn't do Dancing With the Stars but she still did. She did donate the appearance fee she would have received from the show to the Children's Hospital of Los Angeles.\n\nIn 2010, Jane Seymour donated the proceeds of her Dancing With the Stars appearance to the Cedars-Sinai Medical Center. Also in 2010, season 11 winner Donny Osmond donated his Dancing With the Stars appearance fee to Children's Miracle Network Hospitals.\n\nDancing With the Stars airs Mondays at 8/7c on ABC.]" time="0.025"><properties><property name="score" value="0.01529318899560786" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01529319&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01529319
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[PHILADELPHIA (CBS) \u2013 The 76ers fell to 5-12 overall and 0-5 at home with a 99-94 loss to the Milwaukee Bucks.\n\nThe Bucks outscored the Sixers 22-16 in the fourth quarter to overcome an 18 point deficit.\n\nHollis Thompson led the Sixers with a season high 21 points off the bench.\n\n\u201cI\u2019m happy for him. I think he\u2019s earned the opportunity to get more minutes. I thought he was a pretty complete player tonight, he guarded, he rebounded. I think he\u2019s taking the first step,\u201d Sixers coach Brett Brown said.\n\nThe Sixers held a seven point lead after one quarter but couldn\u2019t hold the momentum.\n\n\u201cWe had such an amazing first quarter. We really put a dent in their defensive scheme,\u201d Brown said.\n\nBucks coach Jason Kidd, who played with the Sixers for a short time, on a night where his team set a franchise record with 14 three pointers made.\n\n\u201cWe were missing a lot early and then the last quarter we really played well and got it going. We were able to put some big points up and get some stops when we needed it,\u201d Kidd said.\n\nMichael Carter-Williams led the Sixers with 19 points and six assists.]" time="0.014"><properties><property name="score" value="0.11166642629755752" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.11166643&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.11166643
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[...Rather than simple self-help, this book represents what it should be - a collaborative work between the writer and his/her therapist to ensure maximum understanding and success. - Robert M. Corry, M.D., in his review in Psychotherapy\n\n\u2026Clear, practical, and insightful. - Frederick Chapman, M.D., Harvard Medical School\n\nDavid Cohen, M.D. has written a manual on depression that will help many who suffer with this debilitating condition to recognize that there is hope for recovery. More importantly, his description of effective treatments available today is evidence of the tremendous advances in this field. He has done a wonderful job in providing the reader with easy to understand information about depression and how to overcome it. - Daniel J. Vitiello, M.D.\n\nEditorial Board Member, Journal of Affective Disorders\n\n...Presents a wealth of ideas that are written in an easy to read and accessible style. - Library Journal\n\n...From description to treatment, there is something in this book for all readers. - John Guilford, M.D., Clinical Professor of Psychiatry and Behavioral Sciences, Duke University Medical Center, Durham, NC\n\n...Dr. Cohen's book is the first on the market to deal in detail with the nonmedical therapies available for the treatment of depression and to be accessible to both patients and their families. It is a valuable contribution to the literature and deserves to be widely read. - E.E. Riley, Ph.D., in his Foreword\n\n...Cohen discusses in simple terms the pathophysiology of depression, the cognitive/behavioral theories of depression, the treatment of depression, and provides information about organizations that may be able to help. The author's descriptions of medications, both antidepressants and anti-anxiety medications, are accurate and detailed. The reader can feel confident that the medication section is accurate, without the endless descriptions found in other books. The key point to remember is that the medications are not the &quot;cure&quot; for depression. Instead, they are used to provide a way of breaking through a &quot;state of learned helplessness&quot; and providing a method to help the patient engage in a treatment program. This is the key point in the book. In order to help those suffering with depression, Dr. Cohen believes that the medical profession has a responsibility to refer patients to a &quot;team&quot; which will help them get over their depression and learn to cope with the problem. This book will give the layperson a much-needed introduction to the field of depression, the various treatments, and how they work. It is written in an easy-to-understand language and provides references for the reader to pursue if the reader wishes to know more about the topic. - Richard H. Price, Ph.D., Past President, Academy of Cognitive Therapy\n\n...All those who are dealing with depression will appreciate this practical guide. It is written in a clear, direct manner. Those who are serious about becoming well will find in Dr. Cohen a knowledgeable and empathic guide. - Rev. Marybeth Twichell, Ph.D.\n\nThe excerpt from the Foreword, &quot;The Practical Guide for Persons With Depression,&quot; by Dr. E.E. Riley is reprinted with permission of the author. The excerpt from the Foreword, &quot;A Friend to Help Along the Way,&quot; by Dr. Daniel J. Vitiello is reprinted with permission of the author. The excerpt from the Foreword, &quot;The Working Alliance,&quot; by Dr. Robert M. Corry is reprinted with permission of the author. The excerpt from the Foreword, &quot;A Different Perspective on Depression,&quot; by Dr. John Guilford is reprinted with permission of the author. The excerpt from the Foreword, &quot;Keeping the Ball Rolling,&quot; by Dr. Frederick Chapman is reprinted with permission of the author. The excerpt from the Foreword, &quot;A Review of Common Medications,&quot; by Dr. Richard H. Price is reprinted with permission of the author. The excerpt from the Foreword, &quot;Excerpts from 'Depression - What Every Clinician Should Know',&quot; by Rev. Marybeth Twichell is reprinted with permission of the author. The excerpt from the Foreword, &quot;Fantastic Resource for Persons Who are Coping with Depression,&quot; by Dr. Daniel J. Vitiello is reprinted with permission of the author. The excerpt from the Foreword, &quot;The Treatment of Depression,&quot; by Rev. Marybeth Twichell is reprinted with permission of the author. The excerpt from the Foreword, &quot;Dr. Cohen's Story of Success,&quot; by Dr. E.E. Riley is reprinted with permission of the author. The excerpt from the Foreword, &quot;Respecting Depression and its Sufferers,&quot; by Dr. Robert M. Corry is reprinted with permission of the author. The excerpt from the Foreword, &quot;A New Paradigm of Depression,&quot; by Dr. John Guilford is reprinted with permission of the author. The excerpt from the Foreword, &quot;Facing Depression Together,&quot; by Dr. Frederick Chapman is reprinted with permission of the author. The excerpt from the Foreword, &quot;Depression: The Last Great Challenge,&quot; by Dr. Richard H. Price is reprinted with permission of the author.]" time="0.065"><properties><property name="score" value="0.06827792412225234" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[City workers fill up sandbags for residents after torrential rains on January 21, 2019 in Biloxi, Mississippi. Getty Images\n\nThe storms killed at least six people across the U.S. Gulf Coast.\n\nTropical Storm Barry formed on Thursday evening, and at 11 p.m. ET it was a tropical depression, according to the National Hurricane Center.\n\nBut it's expected to regain strength as it continues to move across the warm waters of the Gulf of Mexico.\n\nThe storm, which has maximum sustained winds of 45 mph, is expected to dump anywhere between 4 to 8 inches of rain across Louisiana and neighboring states.\n\nThe National Hurricane Center issued a hurricane warning for southeastern Louisiana and coastal Mississippi as the storm nears land.\n\nLouisiana Governor John Bel Edwards declared a state of emergency on Thursday.\n\n&quot;We will do all that we can to help people prepare for this storm,&quot; Edwards said at a news conference.\n\nHe added: &quot;I urge all Louisianians to remain vigilant and to stay informed on weather conditions and possible storm impacts through the entire weekend.&quot;\n\nEdwards warned residents to be prepared for potential road flooding and power outages.\n\nThe Weather Channel reported that flooding from the storm is a concern for a number of towns along the Louisiana coastline, including Lafayette and Morgan City.\n\nThe NHC says that &quot;life-threatening storm surge is likely along portions of the Mississippi and southeastern Louisiana coastlines&quot; as the storm nears landfall on Saturday night.\n\nWeather models have]" time="0.013"><properties><property name="score" value="0.05447978122273267" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05447978&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05447978
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[#1 Maintaining momentum. While this might be my blog, I know you guys have a lot going on. We all do. That\u2019s why it\u2019s important to make sure you\u2019re building on what you\u2019ve started and staying on track. It\u2019s so easy to get derailed. The most important thing is to be consistent and you\u2019ll find that the results you want will come naturally.\n\nIf you enjoyed the article please share it on social media. I am the author of 4 books, all available on Amazon.com or Barnesandnoble.com. For more information, please visit www.wright-well.com.]" time="0.013"><properties><property name="score" value="0.03769523832705124" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03769524&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03769524
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Surging.\n\nI was a fan of Steppenwolf's Magic Carpet Ride when I was younger, and my tastes have not changed.\n\nI have been listening to Steppenwolf non-stop.\n\nBut this is not a tribute band, they are completely on their own turf.\n\nThere is a burning wildness in the vocals, an urgency in the songs, that you don't see in tribute bands.\n\nAnd the lead guitar is stellar. I don't even have to describe it because the sound is unmistakable, like the perfect mix of Carlos Santana and David Gilmour.\n\nThe group members hail from all over the planet: Stacey Blades is from New York City, Marco Panascia from Brazil, Marco Alfonso, of Caracas, Venezuela, and Juan Pablo Torres from Guayaquil, Ecuador.\n\nTheir bandcamp page gives you a good overview of their sound:\n\nhttps://steppenwolfband.bandcamp.com/]" time="0.012"><properties><property name="score" value="0.1015753441851952" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.10157534&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.10157534
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[As global oil prices continue to rise, some observers are warning that rising gas prices may cause a major economic downturn in 2008. What's more, the recent natural gas pipeline explosion in California, which prompted a moratorium on new development, will exacerbate the problem.\n\nOver the next five years, the United States will have to import nearly half of the energy it uses, according to the most recent forecast by the U.S. Energy Information Administration. As of now, the United States is the largest energy importer in the world, importing more energy than any other country. We import 3.5 million barrels of oil per day, and more than 80 percent of our oil comes from foreign countries. This puts us at the mercy of OPEC nations, and increasingly Russia, who we are now trying to court to invest in our own energy infrastructure.\n\nThis dependence is a matter of great concern for many energy analysts. Some say that it's time to cut our losses and invest in the most energy-efficient technology possible while we still can.\n\nOur continued dependence on foreign oil could have a long-term impact on the economy, they say, because the price of oil will continue to rise with the demand, and the price of energy could skyrocket in the next decade.\n\n'We Need to Start Thinking Now'\n\n&quot;Energy is not just a topic that's come up in the last few months,&quot; says Matthew Ranson, an analyst with the American Council for an Energy-Efficient Economy. &quot;It's been discussed in a serious way for decades now.&quot;\n\nAccording to Ranson, for a long time, people weren't really paying attention to energy issues in this country. But then, he says, &quot;the economy collapsed in the last year or so and the [energy] debate started up again in full force.&quot;\n\nBut even now, he says, &quot;we need to start thinking about it now. We can't just wait.&quot;\n\nThe economy will be hurt most by high energy prices, says Ranson. &quot;Energy, unlike other goods and services, is necessary for people to have jobs and make a living and grow the economy,&quot; he says.\n\nThe crisis is already in the works. &quot;Gas prices are expected to increase at the rate of $3 a year,&quot; Ranson says. &quot;At some point, it's going to make people reconsider car ownership, which we are seeing already.&quot;\n\nOther analysts have noted that the worst economic effects will come in the next few years as oil prices begin to hit consumers' wallets.\n\n&quot;It's really important for people to understand that we don't have to sit around and wait for it to hit us,&quot; Ranson says. &quot;We can do things now to change the way we make energy choices.&quot;\n\nThe United States needs to cut back on our reliance on oil, he says, and start developing clean, domestic energy resources like solar and wind power.\n\n&quot;We need to think differently about the choices that we're making,&quot; he says. &quot;If we have a choice to import oil from places that don't like us, or if we have a choice to invest in clean domestic energy that's good for us, we need to make the choice.&quot;\n\nThe Golden Years\n\nRanson admits that it's not going to be easy to move toward a clean energy economy. But there are things that can be done, such as improving efficiency and investing in new energy technologies. &quot;We can't do everything at once,&quot; he says.\n\nTo help reduce energy demand, Congress has passed a law to require the federal government to improve energy efficiency in its buildings.\n\nThere are also efforts under way to address the energy crisis in cities, where people can be more dependent on energy, according to Ranson.\n\n&quot;A lot of cities are thinking about this in different ways,&quot; he says. &quot;In New York, they're talking about ways to reduce car ownership, like getting people out of cars and into public transportation, making public transportation more convenient, and adding bike paths. You can start with that and see what happens.&quot;\n\nThe Natural Gas Supply\n\nAnother important energy issue to keep an eye on is the country's natural gas supply. Natural gas accounts for nearly 20 percent of the country's energy, and nearly 60 percent of the nation's electricity generation.\n\n&quot;We don't really have a clear understanding of how the supply of natural gas will be affected by what's going on in California,&quot; says Fred Millar, senior fellow at the Washington-based think tank, Energy Security Initiative.\n\nMillar believes that the California situation may signal &quot;troubles ahead&quot; for the country's natural gas supply. &quot;There are real questions about how much natural gas is going to be produced by the industry in the coming years,&quot; he says.\n\nThere's been a dramatic drop in natural gas prices in the last few years, and that has encouraged a lot of companies to start new drilling operations, says Millar. But natural gas prices were going up last week, Millar says, and &quot;the question now is, has the bubble burst?&quot;\n\nA recent explosion at a Southern California natural gas pipeline, which took the lives of two workers, has put a halt to natural gas production in the state. The pipeline was owned by the Texas-based company Sempra Energy, and the blast occurred near Bakersfield. It is still unclear what caused the explosion.\n\nThe blast caused a temporary halt to the development of the Puente Valley Pipeline, a major new pipeline project.\n\nMillar says that the halt on the pipeline will have a major effect on the state's energy supplies. &quot;It could cause a major problem for California's electricity supplies,&quot; he says.\n\nMillar adds that the price of natural gas is likely to rise in the coming weeks and months because of the halt.\n\nOil and the Economy\n\nHigh energy prices are also a major concern for economists. &quot;The most immediate effect of a large price increase is inflation,&quot; says David Hughes, a geologist who has written extensively on energy issues. &quot;We can expect energy prices to increase for gasoline, electricity, natural gas, and other fossil fuels.&quot;\n\nBut economists are still debating how the economy will be affected. Some economists say that higher energy prices could have a positive impact on the economy, while others say the impact could be negative.\n\n&quot;It's too early to tell,&quot; says Millar. &quot;We're still in the stage where the price increases are in the process of playing out.&quot;\n\n&quot;The economy is more fragile than it was in the 1970s, and energy is more integrated into the economy, and prices are higher,&quot; says Ranson.\n\nBut he believes that the economy is better positioned to deal with energy prices now than it was in the 1970s. &quot;I don't think the impact of high energy prices will be nearly as bad this time,&quot; he says.\n\nA slowdown in the economy and a possible recession, Ranson says, could cause energy prices to spike even more.\n\nBut even if the economy takes a downturn, Ranson says that energy prices will keep going up. &quot;People may stop spending on some things,&quot; he says, &quot;but energy prices are going to keep going up.&quot;\n\nRanson says that people will have to make choices about how they live, and the government will have to decide how much of a subsidy it's willing to provide for the alternative energy industry.\n\nHughes believes that, ultimately, high energy prices will bring about some kind of economic crisis. &quot;In my view, this is a symptom of a much bigger problem, which is our oil-based economy,&quot; he says. &quot;We can no longer afford the fossil fuel based economy we have had for the past century.&quot;\n\nCopyright 2007 E&amp;E Publishing. All Rights Reserved.]" time="0.071"><properties><property name="score" value="0.01624908292154878" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01624908&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01624908
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[On the first play from scrimmage, Washington quarterback Jake Browning went deep for Dante Pettis, who picked up 50 yards on the play, but then injured his knee. Browning threw a touchdown pass to Chico McClatcher on the next play, but Browning later came out of the game and it was announced that he had a right leg injury. No further details were given on his condition, but his return]" time="0.013"><properties><property name="score" value="0.05671331337818164" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05671331&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05671331
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[I\u2019m back in Shoreditch and work starts again tomorrow so I\u2019m trying to get things sorted before that. However, this was my second big trip of the year and I\u2019m feeling a bit worn out. It feels like it was longer than a week though.\n\nI went to Seoul with Jess for two weeks. We flew to Seoul Incheon on Cathay Pacific and stayed at the Hyatt in Seoul Station, just outside the hotel station (it was actually our last day before we flew out). We visited the Han River and the Hongdae area (including bars, restaurants and music venues) and went to the Seoul Film Festival. We spent a lot of time walking around too as we don\u2019t really know the city that well and were pretty tired by the end of each day.\n\nWe visited a temple near Cheongdam called Angukdaegyo which was beautiful. We also visited the grounds of the 63 Building which is quite close to the hotel. The following day, we went to Jogyesa Temple which is where the head of the Jogye Order of Korean Buddhism has his residence. We also went to Cheongwadae which is the residence of the President of Korea and the Blue House. I really wanted to go inside the Blue House but it was closed when we were there. The area around the Blue House was really beautiful.\n\nWe had a really nice lunch at the Daehangno area. We had a really nice lunch at the Dongmyo Station near Gyeongbokgung Palace. The area is quite nice. We then went to Insadong where there were lots of small shops. We also visited Namsangol Hanok Village and had a nice lunch at the Hanok Village. It was also near Gyeonghuigung. We walked up the hills near Gyeonghuigung, which were quite steep and also quite a nice walk. We also visited Seoul Tower which was lovely and there was a great view of Seoul.\n\nWe then travelled to Busan for a day to see the Ulsan Bijudong Cable Car and to take the ferry to Geoje Island. Busan was nice and I can definitely see why people live there. There was a great view of the coast from the cable car. There were quite a lot of islands in Geoje Island. We stayed in the Geoje Island Ferry Cruise which is a small hotel. It was very pleasant there and the food was good. We walked around the island and saw the Minho lighthouse. It was quite cloudy so we couldn\u2019t see the islands around. We stayed for dinner and saw the sunset before we returned to Busan.\n\nOur next destination was Jeju Island. We flew from Busan to Jeju Island on a flight that took about 1 hour. There was also a harbour just outside the hotel. We stayed at the Full Moon Hotel. It was quite beautiful. We visited Hallasan Mountain and the Jeju Haenyeo Museum of Women\u2019s Contributions to Life in the Jeju Island in which there was a lovely view of Hallasan Mountain. We also visited Samcheok Beach. We had a great dinner at the Full Moon Hotel in the Samcheok Bay. There was a beautiful view of the moon rising over the beach from our room.\n\nThe next day, we went to Seongsan Ilchulbong Peak, which was very beautiful. It is a volcanic crater on a small island that was completely formed from lava. It was cloudy and rainy while we were there. We went to Seogwipo and had a lovely lunch at Jin Jin Gop. We had a really nice seafood lunch and it was just a couple of minutes away from Seogwipo station. We went to the Seogwipo Lotte Shopping Center and bought some gifts.\n\nWe then went to Hallim Park. It was a very hot day but it was very pleasant there. We went on a cable car to Mount Baekhwa, which was really nice. It is the highest mountain on Jeju Island and the views were nice. We also went to Mokcheongyeon cave and we visited the memorial museum and the huge cave. The cave was really nice and it was a good experience. We then went to Jeju Olle Trail which is a series of pathways that cross Jeju Island. They have been designed to showcase the most beautiful spots on Jeju. We went to the Jeju Olle trail which was near the entrance of Hallim Park. The weather was really nice so it was pleasant to be outside.\n\nWe went to Seongsan Sunrise Peak which was a nice place to see the sunrise. There was a really nice sunset too. The weather was very nice and it was a pleasant place to be. We also went to Mt Sine, which was very pretty. It was a bit foggy and we couldn\u2019t see Mt Halla but it was a nice walk. We also went to the 7 treasures of Jeju which was nice. We then went to Biyang-ri, Seogwipo, which was a bit more rural and very nice.\n\nOur last destination was the Jeju World Cup Stadium. It was a really nice stadium with lots of seating. It had lots of places to eat and drink too.\n\nWe then flew back to Seoul to fly home. It was a lovely place and I can definitely see why people would want to live there.]" time="0.068"><properties><property name="score" value="0.03315561641190452" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03315562&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03315562
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[In her interview with Popnography, Kate Micucci discussed her two new projects, BoJack Horseman and another animated series on Yahoo! Screen called Other Space. Other Space will be an anthology comedy following the crew of the space shuttle Venturi 2, created by Paul Feig.\n\nOther Space will be the first show I do where the cast is mostly male. I\u2019m a big feminist and I\u2019m not sure how I feel about that, but my character, Natalie, is the only girl on the ship and the youngest, which makes her feel isolated from the rest of the crew. She\u2019s shy and reserved. She\u2019s a huge sci-fi nerd and very book smart but not street smart. She\u2019s obsessed with books and fantasy but there are too many people on the ship and she has a hard time focusing.\n\nKate plays one of the main characters, Natalie, a woman with aspirations of space travel who works on the shuttle, Venturi 2. The series has been compared to an early 90s show called Space Cases, which also takes place on a space shuttle, however, it seems that Other Space will be a little bit more adult and will likely also delve into the science fiction genre as well as the comedy. Micucci will be joined by a great ensemble of actors including The Hunger Games actor, Sam Richardson, as Commander Glen. Richardson and Micucci previously worked together on Saturday Night Live, and Kate shared how much she loves working with him.\n\nI love working with Sam. He\u2019s so wonderful. He\u2019s so cute. I think my character Natalie really brings him out of his shell. She\u2019s the only one who really has the courage to push him. I think she also brings out a really kind, sensitive side in him that he probably doesn\u2019t have a lot of. It\u2019s been a really lovely experience to get to work with him. He\u2019s just a very warm, kind person.\n\nAlso joining the cast are Rebecca Romijn as the ambitious first officer, Kelly, and Enrico Colantoni as Captain Zalian. Other Space will be directed by Joe Russo and will be a half hour comedy show with ten episodes. The show is set to premiere on Yahoo! Screen on April 14, 2015.]" time="0.025"><properties><property name="score" value="0.07840539239826111" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07840539&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07840539
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The 24th annual World Press Photo Contest has announced its winners. One of the most prestigious and respected contests for photography, winners were selected from 82,951 images submitted by 5,775 photographers from 124 countries. Below are some of the winners in the photojournalism category. All captions were written by the individual photographers and are published as they were received.]" time="0.012"><properties><property name="score" value="0.029467131570370952" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02946713&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02946713
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Scoop First, an economic-development nonprofit organization based in Mobile, will present \u201cShoals Business Snapshots\u201d at 11 a.m. March 28 at the Crowne Plaza Hotel, 4880 Montlimar Drive.\n\nThe program features short presentations about businesses throughout Mobile and Baldwin counties, including businesses, schools and nonprofit organizations.\n\nThe event is free and open to the public.\n\nA light lunch will be provided.]" time="0.013"><properties><property name="score" value="0.029331011849426103" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02933101&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02933101
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Home &gt; Table Lamp Base &gt; Pearl / Crystal &gt; Round Crystal Table Lamp Base\n\nRound Crystal Table Lamp Base Code: 316 Price: US$ 20.00 New from: US$ 20.00 30 day money back guarantee\n\nMade to order, usually ships within 10 to 14 days Quantity in Basket: None Round Crystal Table Lamp Base\n\n\n\nA round base with polished silver or gold finish and a bright clear crystal knob that adds a sparkle to any table lamp.\n\n\n\nTo order:\n\nEnter the number of bases you wish to order into the box above the quantity selection and press the 'Add to Cart' button.\n\n\n\nFor quantities of less than 12 we will invoice you on receipt of your order and ask for payment to be made before your order is sent.\n\nThere are no products to list in this category.]" time="0.012"><properties><property name="score" value="0.07092321544214797" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07092322&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07092322
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Post by waltzmania \xbb Sun Nov 24, 2017 7:25 pm\n\nJudging by what you're hearing, I'm not sure the recording was ever mixed properly. My best guess is that it was mixed for the wrong kind of playback device. The first movement of the symphony is long and in four sections. If there are abrupt changes in sound quality, it is almost certain that]" time="0.013"><properties><property name="score" value="0.0452457390333838" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04524574&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04524574
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Der Pr\xe4sident des Verfassungsschutzes, Hans-Georg Maa\xdfen, ist von Bundeskanzlerin Angela Merkel (CDU) entlassen worden. Wie das Bundeskabinett am Mittwoch nach Teilnehmerangaben beschloss, wird Maa\xdfen zum Innen-Staatssekret\xe4r im Bundesinnenministerium bef\xf6rdert. Seine bisherige Stellvertreterin, Pr\xe4sidentin des Bundesamtes f\xfcr Verfassungsschutz (BfV), Susann Wimmer, wird dann seinen Posten als Beh\xf6rdenleiter \xfcbernehmen. Wie die &quot;S\xfcddeutsche Zeitung&quot; berichtet, haben die drei Parteichefs der Gro\xdfen Koalition, Angela Merkel (CDU), Horst Seehofer (CSU) und Andrea Nahles (SPD), die Personalie Maa\xdfen am Dienstagabend einstimmig beschlossen. Merkel habe demnach in einem &quot;Auftragskommando&quot; an die drei Parteichefs auf die schnelle L\xf6sung gedrungen. Im Laufe des Tages soll der Bundestag Maa\xdfen noch abschlie\xdfend \xfcber die Bef\xf6rderung abstimmen.\n\nMaa\xdfen soll dann als Staatssekret\xe4r f\xfcr europ\xe4ische und internationale Aussagen im Bundesinnenministerium eingesetzt werden, wie es am Mittwoch aus Regierungskreisen hie\xdf. Demnach wird er Staatssekret\xe4r f\xfcr die Bereiche Bauen, Wohnen und Energie.\n\nMerkel hatte am Dienstag erkl\xe4rt, die Koalition wolle den Fall Maa\xdfen noch heute abschlie\xdfend kl\xe4ren. Nachdem der Verfassungsschutzpr\xe4sident angeblich Zweifel an den Aussagen des j\xfcdischen Mannes ge\xe4u\xdfert hatte, der in Chemnitz gefilmt worden war, hatte der &quot;Spiegel&quot; berichtet, Maa\xdfen habe die AfD-Spitze in Sachsen gewarnt, eine Einbindung des Verfassungsschutzes in die Aufarbeitung des Chemnitzer T\xf6tungsdeliktes anzustreben. Zudem soll Maa\xdfen im Zusammenhang mit den Vorf\xe4llen in Chemnitz gegen\xfcber der &quot;Bild&quot;-Zeitung gesagt haben, dass ein Video von dem \xdcbergriff auf einen Mann &quot;gef\xe4lscht&quot; sein k\xf6nnte. Die Generalstaatsanwaltschaft Dresden bezeichnete diese Aussagen zu einem angeblichen Fake-Video als &quot;irref\xfchrend&quot;.\n\nAm Dienstag hatte der Fall Maa\xdfen bereits f\xfcr Unruhe in der Bundesregierung gesorgt, nachdem das Kabinett offenbar \xfcberraschend \xfcber die Zukunft des Verfassungsschutzpr\xe4sidenten entschieden hatte. Nach Informationen von NDR, WDR und &quot;S\xfcddeutscher Zeitung&quot; sollen SPD-Chefin Nahles und CSU-Chef Seehofer die Entscheidung der Kanzlerin zum sofortigen Rauswurf Maa\xdfens umgehend blockiert haben. Laut Informationen der &quot;S\xfcddeutschen Zeitung&quot; hat Merkel den Koalitionspartnern gegen\xfcber zudem ihren Willen bekr\xe4ftigt, Maa\xdfen als Staatssekret\xe4r in ihrem Kabinett zu halten. Merkel sei ver\xe4rgert gewesen \xfcber die Weigerung der SPD-Parteispitze, den konservativen Kritiker der Migrationspolitik als Staatssekret\xe4r im Innenministerium zu akzeptieren, schreibt die Zeitung unter Berufung auf SPD-Kreise.\n\nMaas: &quot;Irrelevante Debatte&quot;\n\nVor dem Hintergrund der Debatte um Maa\xdfen hat Au\xdfenminister Heiko Maas (SPD) die Bundesregierung aufgefordert, eine &quot;tiefgreifende und un\xfcbersehbare Antwort auf die Herausforderungen der Migration und die Probleme mit der Integration&quot; zu geben. &quot;Es geht nicht um eine Personaldiskussion, sondern um das Schicksal der Demokratie&quot;, sagte Maas am Mittwoch in Berlin. Die Debatte \xfcber Maa\xdfen sei nicht nur eine rein parteipolitische, sondern vor allem eine &quot;irrelevante&quot;.\n\nDas Verh\xe4ltnis der gro\xdfen Koalition zueinander sei &quot;durch die Vorg\xe4nge in der letzten Woche in keiner Weise verbessert&quot;, sagte Maas. &quot;Was wir in den letzten Tagen erlebt haben, ist ein erschreckender Mangel an Selbstverst\xe4ndnis, der manchmal nur noch unterstreicht, wie wenig sich die gro\xdfe Koalition selbst wertsch\xe4tzt.&quot; Dass es &quot;in den vergangenen Wochen&quot; dar\xfcber diskutiert worden sei, &quot;ob sich Maa\xdfen noch halten l\xe4sst&quot;, k\xf6nne niemanden erfreuen. Die Entscheidung, Maa\xdfen entlassen zu wollen, sei dar\xfcber hinaus von Personaldebatten \xfcberlagert worden.\n\nAufgeregte Stimmen\n\nBei der Debatte zum Haushalt des Bundesinnenministeriums im Bundestag wurde im Anschluss an die Haushaltsrede von Minister Horst Seehofer (CSU) die Stimmung in den Reihen der Regierungsparteien am Mittwoch sehr aufgeregt. Die Opposition im Bundestag sieht in den Spannungen der Koalition \xfcber die Zukunft des Verfassungsschutzpr\xe4sidenten Hans-Georg Maa\xdfen auch ein eigenes Versagen beim Streit.\n\n&quot;F\xfcr das, was hier gerade passiert, tragen Sie alle, die Sie hier sitzen, die politische Verantwortung&quot;, sagte die SPD-Fraktionsvorsitzende Andrea Nahles. &quot;Sie tragen daf\xfcr die Verantwortung, dass es in dieser Woche \xfcberhaupt zu einem solchen Haushaltsgesetz kommen kann, wie Sie es hier machen.&quot; Nahles forderte Seehofer auf, die Vorg\xe4nge \xfcber Maa\xdfen zu erkl\xe4ren, weil sie sich selbst von der Regierung verraten f\xfchle. &quot;Die Regierung und die politische F\xfchrung dieses Landes muss jetzt den Weg der Einheit beschreiten&quot;, forderte sie.\n\nKritik \xfcbte Nahles daran, dass das Bundesinnenministerium bei der Einrichtung von Erstaufnahmeeinrichtungen in Anforderungen an die jeweiligen Bundesl\xe4nder nun &quot;weich&quot; geworden sei. Zudem gebe es zu wenig Pl\xe4tze f\xfcr die Abschiebung abgelehnter Asylbewerber und zu wenig R\xfcckf\xfchrungen nach Afghanistan. Es sei ein &quot;Skandal&quot;, dass so viele Menschen in Deutschland festsitzen.\n\nAfD: Maa\xdfen soll Verfassungsschutz-Chef bleiben\n\nDie AfD h\xe4lt dagegen. Fraktionschef Alexander Gauland begr\xfc\xdfte Maa\xdfen als neuen Staatssekret\xe4r im Innenministerium. Er sei \xfcberzeugt, dass Maa\xdfen in der neuen Funktion seine Erfahrungen als Pr\xe4sident des Verfassungsschutzes in die Politik einbringen k\xf6nne, sagte]" time="0.071"><properties><property name="score" value="0.24675457806605347" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.24675458&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.24675458
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Establishing a GitHub Account\n\nLog into your UCSC Gmail account. If you need to create a new UCSC Gmail account, visit this webpage. Once you have logged into your UCSC Gmail account, you can begin the process of creating a GitHub account. Please use your UCSC username as your GitHub username.\n\nTo create a GitHub account:\n\nIn the upper-right corner of the Gmail interface, click the menu icon, then click Settings. On the left side of the settings window, click Accounts and Import. In the lower-left corner of the Accounts and Import window, click Create an account. In the first field, enter your GitHub username. This is your UCSC username. In the second field, enter a GitHub password. This will be your username@github.com email address. In the third field, enter a password for your GitHub account. Please remember to remember this password. This password is only to access your GitHub account from the web. To access your account on your local machine, you will still need to use SSH. In the last field, enter your email address. This will be the email address your collaborators will use to send you commits and other notifications. Click the Sign Up button.\n\nYour UCSC GitHub account will be created. To access your new GitHub account, simply visit GitHub.com and log in with your GitHub username and password.\n\nSet up your GitHub Account\n\nIf you want to allow others to view or contribute to your project, you will need to set up your account to allow this.\n\nTo set up your GitHub account, go to the project you want to contribute to, click on the &quot;Settings&quot; button in the upper right corner of the GitHub page, and under &quot;GitHub profile&quot; click &quot;Edit&quot;. In the profile settings, you will see the &quot;Contributions&quot; tab. You can change the settings in this tab to either allow others to contribute to your project or to view your contributions to others' projects. Click on the &quot;Next&quot; button to save your changes.\n\nGitHub provides a detailed guide for the contribution settings here.]" time="0.023"><properties><property name="score" value="0.06411508213206071" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Myanmar currency and coins\n\nMyanmar Banknotes\n\nMyanmar has a local currency, the Kyat, or K, named after King Thibaw, a former king of Burma. The kyat was introduced in 1923, replacing the Indian rupee which had been in use in Burma for about two centuries. One kyat is subdivided into 100 pyas, and one kyat is subdivided into 100 cents. Coins have been minted in denominations of 5, 10, 25, 50, 100, 500 and 1000 kyats, and 1, 5, 10 and 25 pyas, though the highest four denominations are rarely used.\n\nThe coins and notes currently in circulation are:\n\nBanknotes Coins\n\nKyat\n\n50, 100, 500, 1000 Kyat\n\n5, 10, 25, 50, 100, 500 Pyas\n\n1, 5, 10, 25, 50, 100 Cents\n\n50, 100, 500, 1000 Cents\n\n5, 10, 25, 50, 100\n\nCheck out the latest exchange rates!\n\nMyanmar Currency History\n\nDuring the last half of the 19th century, large quantities of Indian coins circulated in Burma. In 1885, Burma began to issue its own coinage, in a new decimal system. The rupee and the double rupee were equivalent to one kyat. These coins were not for circulation outside of Burma and were not legal tender in India. A similar coinage was issued in the last year of the 19th century, using slightly different weights and diameters for the three silver coins and for the copper coins.\n\nDuring World War I, the Burmese Government began issuing a series of paper notes for 1, 5, 10, 20, 100 and 1000 rupees. They were overprinted on banknotes of the Imperial Bank of India, the currency used for exchange in Burma until the mid-1930s.\n\nDuring the 1920s, the paper money of several countries was in circulation in Burma. Since the demand for small denomination notes was greatest, the British Government began to issue India notes in 1925. They were of low denomination, being in the denominations of 5, 10 and 20 rupees. The existing paper money was then demonetised.\n\nIn 1940, the Japanese occupied Burma and issued their own notes in place of the British issues. These notes continued to be in use until 1945, when the British re-established control of the country. In 1942, emergency notes were introduced, in the denominations of 1, 5, 10, 20 and 100 rupees. These notes were overprinted on paper notes issued by the Reserve Bank of India.\n\nIn 1946, a new series of notes was issued in Burma. This consisted of Indian notes issued by the Reserve Bank of India, overprinted B.E.P. and the State Bank of Burma. These were in denominations of 1, 5, 10, 20, 100 and 1000 rupees.\n\nThe present series of notes began in 1951. These notes are in denominations of 1, 5, 10, 20, 50, 100, 500, 1000 and 5000 kyats.\n\nThe &quot;Burmese&quot; rupee, issued by the British between 1885 and 1896, was a fixed exchange rate at 1 shilling 4 pence (two and one-third rupees).\n\nThe U.S. dollar was established as the national currency of Burma in 1885 and is still the predominant currency in use in Myanmar.\n\nBanking in Myanmar\n\nMyanmar has an international banking system, with at least 7 major banks operating there. These include the state-owned Myanmar Foreign Trade Bank (MFTB) and the private Myanmar Economic Bank (MEB), National Bank of Commerce (NBC), Myanma Foreign Trade Bank (MFTB), Myanma Investment and Commercial Bank (MICB), Myanmar Mayflower Bank (MMB), Myanma Ahlia Bank (MAB), and Myanma Security Bank (MSB).\n\nMost commercial banks use the SWIFT system to transfer funds, and many banks in Myanmar accept electronic banking services through mobile phone banking, Internet banking, and ATMs. Some smaller banks still deal only in cash, but since cash is used mainly for larger transactions and shopping, and with so many ATMs in service now, it is no longer a problem.\n\n\n\nMyanmar currency info\n\nMyanmar info\n\nGuide to Myanmar\n\nCurrency information on country homepages]" time="0.073"><properties><property name="score" value="0.023535615010369937" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02353562&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02353562
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[This week at Marvel there\u2019s an \u201cArmor Wars\u201d as the New Warriors reunite! Will the Illuminati save our universe from the all-powerful \u201cManhattan Projects\u201d? And that\u2019s just the beginning! The Big 2 also has its second full week of August books, including DC\u2019s \u201cInjustice: Gods Among Us\u201d #11, \u201cTrinity of Sin: Pandora\u201d #3, and \u201cBatman Eternal\u201d #19.\n\nDC has the following this week: \u201cGreen Lantern\u201d #24, \u201cInjustice: Gods Among Us\u201d #11, \u201cThe Flash\u201d #24, \u201cJustice League of America\u201d #9, \u201cBatman and Robin\u201d #24, \u201cDetective Comics\u201d #24, \u201cRed Hood and the Outlaws\u201d #24, \u201cCatwoman\u201d #24, \u201cTeen Titans\u201d #24, \u201cThe Green Team: Teen Trillionaires\u201d #3, \u201cNew 52: Future\u201d #3, \u201cThe Movement\u201d #6, \u201cStormwatch\u201d #24, \u201cHarley Quinn\u201d #7, \u201cHe-Man: The Eternity War\u201d #2, \u201cInjustice: Gods Among Us\u201d #12, \u201cTrinity of Sin: The Phantom Stranger\u201d #19, \u201cSuperman/Wonder Woman\u201d #8, \u201cDeathstroke\u201d #24, \u201cSinestro\u201d #3, \u201cAll-Star Western\u201d #24, \u201cWorlds\u2019 Finest\u201d #19, \u201cThe Multiversity\u201d #1, \u201cBatgirl\u201d #24, \u201cGreen Lantern Corps\u201d #24, \u201cNew Suicide Squad\u201d #9, \u201cEarth 2\u201d #22, \u201cCatwoman\u201d #25, \u201cThe Green Team: Teen Trillionaires\u201d #4, \u201cConstantine\u201d #4, \u201cThe Movement\u201d #7, \u201cJustice League Dark\u201d #24, \u201cGreen Lantern New Guardians\u201d #24, \u201cRed Lanterns\u201d #24, \u201cAquaman\u201d #24, \u201cWonder Woman\u201d #24, \u201cHe-Man: The Eternity War\u201d #3, \u201cBatman \u201966\u201d #12, \u201cAll-Star Western\u201d #25, \u201cAction Comics\u201d #24, \u201cTeen Titans\u201d #25, \u201cCatwoman\u201d #25, \u201cStormwatch\u201d #25, \u201cBatgirl\u201d #25, \u201cAll-Star Western\u201d #26, \u201cHarley Quinn\u201d #8, \u201cDeathstroke\u201d #25, \u201cAquaman\u201d #25, \u201cJustice League 3000\u201d #8, \u201cSupergirl\u201d #24, \u201cSmallville Season 11\u201d #10, \u201cBatman \u201966\u201d #13, \u201cSuperman/Wonder Woman\u201d #9, \u201cSwamp Thing\u201d #24]" time="0.014"><properties><property name="score" value="0.04382902211408357" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04382902&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04382902
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[FIG. 1 is a perspective view of a virtual reality system in accordance with a preferred embodiment of the present invention;\n\nFIG. 2 is a top plan view of the virtual reality system of FIG. 1;\n\nFIG. 3 is a front plan view of the virtual reality system of FIG. 1;\n\nFIG. 4 is a side elevational view of the virtual reality system of FIG. 1;\n\nFIG. 5 is a perspective view of a two-handed tracking device in accordance with a preferred embodiment of the present invention;\n\nFIG. 6 is a perspective view of the two-handed tracking device of FIG. 5 with an array of LEDs mounted thereto;\n\nFIG. 7 is a perspective view of a three-dimensional position sensing device in accordance with a preferred embodiment of the present invention;\n\nFIG. 8 is a perspective view of a two-handed tracking device in accordance with another preferred embodiment of the present invention;\n\nFIG. 9 is a perspective view of a three-dimensional position sensing device in accordance with another preferred embodiment of the present invention;\n\nFIG. 10 is a perspective view of a three-dimensional position sensing device in accordance with another preferred embodiment of the present invention;\n\nFIG. 11 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 12 is a top plan view of the virtual reality system of FIG. 11;\n\nFIG. 13 is a front plan view of the virtual reality system of FIG. 11;\n\nFIG. 14 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 15 is a top plan view of the virtual reality system of FIG. 14;\n\nFIG. 16 is a front plan view of the virtual reality system of FIG. 14;\n\nFIG. 17 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 18 is a top plan view of the virtual reality system of FIG. 17;\n\nFIG. 19 is a front plan view of the virtual reality system of FIG. 17;\n\nFIG. 20 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 21 is a top plan view of the virtual reality system of FIG. 20;\n\nFIG. 22 is a front plan view of the virtual reality system of FIG. 20;\n\nFIG. 23 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 24 is a top plan view of the virtual reality system of FIG. 23;\n\nFIG. 25 is a front plan view of the virtual reality system of FIG. 23;\n\nFIG. 26 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 27 is a top plan view of the virtual reality system of FIG. 26;\n\nFIG. 28 is a front plan view of the virtual reality system of FIG. 26;\n\nFIG. 29 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 30 is a top plan view of the virtual reality system of FIG. 29;\n\nFIG. 31 is a front plan view of the virtual reality system of FIG. 29;\n\nFIG. 32 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 33 is a top plan view of the virtual reality system of FIG. 32;\n\nFIG. 34 is a front plan view of the virtual reality system of FIG. 32;\n\nFIG. 35 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 36 is a top plan view of the virtual reality system of FIG. 35;\n\nFIG. 37 is a front plan view of the virtual reality system of FIG. 35;\n\nFIG. 38 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 39 is a top plan view of the virtual reality system of FIG. 38;\n\nFIG. 40 is a front plan view of the virtual reality system of FIG. 38;\n\nFIG. 41 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 42 is a top plan view of the virtual reality system of FIG. 41;\n\nFIG. 43 is a front plan view of the virtual reality system of FIG. 41;\n\nFIG. 44 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 45 is a top plan view of the virtual reality system of FIG. 44;\n\nFIG. 46 is a front plan view of the virtual reality system of FIG. 44;\n\nFIG. 47 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 48 is a top plan view of the virtual reality system of FIG. 47;\n\nFIG. 49 is a front plan view of the virtual reality system of FIG. 47;\n\nFIG. 50 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 51 is a top plan view of the virtual reality system of FIG. 50;\n\nFIG. 52 is a front plan view of the virtual reality system of FIG. 50;\n\nFIG. 53 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 54 is a top plan view of the virtual reality system of FIG. 53;\n\nFIG. 55 is a front plan view of the virtual reality system of FIG. 53;\n\nFIG. 56 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 57 is a top plan view of the virtual reality system of FIG. 56;\n\nFIG. 58 is a front plan view of the virtual reality system of FIG. 56;\n\nFIG. 59 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 60 is a top plan view of the virtual reality system of FIG. 59;\n\nFIG. 61 is a front plan view of the virtual reality system of FIG. 59;\n\nFIG. 62 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 63 is a top plan view of the virtual reality system of FIG. 62;\n\nFIG. 64 is a front plan view of the virtual reality system of FIG. 62;\n\nFIG. 65 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 66 is a top plan view of the virtual reality system of FIG. 65;\n\nFIG. 67 is a front plan view of the virtual reality system of FIG. 65;\n\nFIG. 68 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 69 is a top plan view of the virtual reality system of FIG. 68;\n\nFIG. 70 is a front plan view of the virtual reality system of FIG. 68;\n\nFIG. 71 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 72 is a top plan view of the virtual reality system of FIG. 71;\n\nFIG. 73 is a front plan view of the virtual reality system of FIG. 71;\n\nFIG. 74 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 75 is a top plan view of the virtual reality system of FIG. 74;\n\nFIG. 76 is a front plan view of the virtual reality system of FIG. 74;\n\nFIG. 77 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 78 is a top plan view of the virtual reality system of FIG. 77;\n\nFIG. 79 is a front plan view of the virtual reality system of FIG. 77;\n\nFIG. 80 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 81 is a top plan view of the virtual reality system of FIG. 80;\n\nFIG. 82 is a front plan view of the virtual reality system of FIG. 80;\n\nFIG. 83 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 84 is a top plan view of the virtual reality system of FIG. 83;\n\nFIG. 85 is a front plan view of the virtual reality system of FIG. 83;\n\nFIG. 86 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 87 is a top plan view of the virtual reality system of FIG. 86;\n\nFIG. 88 is a front plan view of the virtual reality system of FIG. 86;\n\nFIG. 89 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 90 is a top plan view of the virtual reality system of FIG. 89;\n\nFIG. 91 is a front plan view of the virtual reality system of FIG. 89;\n\nFIG. 92 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 93 is a top plan view of the virtual reality system of FIG.]" time="0.078"><properties><property name="score" value="0.4680979850306704" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[To show their support for Venezuela, five prominent Tibetan Buddhist teachers from the United States will be joining a gathering in Caracas, Venezuela on September 24, 2015. The purpose of this gathering is to support the effort of the Venezuelan people to bring peace and prosperity to the country.\n\nWe will be holding teachings and prayers in three different locations. The first event will be held at 1:30 p.m. at the Cine Teatro Teresa Carreno in Caracas. This venue can seat 900 people. The second and third events will be held at the Jos\xe9 Antonio Paez Theatre, which can seat around 400 people.\n\nIf you are in Venezuela, please attend one of these events. And if you cannot come in person, please join us in spirit by thinking prayers for Venezuela, especially during the times of the teaching and prayers.\n\nThis event is being organized by Geshe Tsultrim Gyeltsen, president of the Foundation for the Preservation of the Mahayana Tradition (FPMT). He will be joined by other FPMT teachers, including Ven. Robina Courtin, Ven. Tenzin Chonyi, Ven. Tsondue Tsomo, and Ven. Marta Meana.\n\nThe event is being supported by the Embassy of the Bolivarian Republic of Venezuela to the United States, and the Embassy of the Bolivarian Republic of Venezuela to the United Nations.]" time="0.015"><properties><property name="score" value="0.015573770649240926" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01557377&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01557377
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[more-in\n\nJust ahead of the International Yoga Day on June 21, the Gujarat High Court on Monday observed that the Prime Minister Narendra Modi-led government's decision to declare June 21 as International Yoga Day was against the spirit of the Constitution.\n\nHearing a PIL against the Centre's decision, the High Court said that though there was nothing wrong in promoting yoga, but that should not be at the cost of other forms of physical exercise. The Centre's decision to declare International Yoga Day was against the spirit of Article 51A (h) of the Constitution, which emphasises on the promotion of scientific temper, humanism and the spirit of inquiry and reform.\n\n&quot;We will have to consider various aspects like whether the decision to declare June 21 as International Yoga Day is against the Constitution of India. A decision has to be taken as to whether scientific temper, humanism and spirit of inquiry is being followed by the Centre or not,&quot; observed Justice A.J. Desai.\n\nS.P. Tamang, the petitioner, had sought a direction to the Centre to hold scientific discussions before declaring a holiday. The petitioner also questioned the need for declaring June 21 as International Yoga Day and claimed that various festivals and events were already being celebrated by India on the same day.]" time="0.014"><properties><property name="score" value="0.0625912788774996" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Open file to use in Google Earth (KML)\n\nCreate Waypoint or Locator from the KML loaded.\n\n\n\nFile &gt; &quot;Open file to use in Google Earth&quot;\n\n\n\nExample\n\nC:\\Program Files (x86)\\Google\\Google Earth\\GoogleEarth.exe\n\nC:\\Program Files (x86)\\Google\\Google Earth\\kmz\\Dhahran.kml\n\nC:\\Program Files (x86)\\Google\\Google Earth\\kmz\\Dhahran.kml\n\nC:\\Program Files (x86)\\Google\\Google Earth\\kmz\\Dhahran.kml\n\nC:\\Program Files (x86)\\Google\\Google Earth\\kmz\\Dhahran.kml\n\n...\n\nOptions - Jumps to KML file in file chooser]" time="0.013"><properties><property name="score" value="0.04520045098162506" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Schools\n\n\nThis report focuses on the experiences of LGBT youth in the nation's public and private schools. This report shows that anti-LGBT victimization in school is a serious problem for a large number of students. It also provides data on other forms of bullying and discriminatory treatment that LGBT students experience. Some of the findings of this report are particularly disturbing. For example, over half (53.1 percent) of LGBT students experienced feeling unsafe in school because of their sexual orientation, over one-quarter (26.1 percent) because of their gender expression, and over one-third (35.2 percent) because of both their sexual orientation and gender expression. This harassment and discrimination takes a toll on these students' well-being. The impact of the harassment and discrimination experienced by LGBT students is serious. LGBT students were more likely than non-LGBT students to report several adverse outcomes. For example, they were more likely than non-LGBT students to report high levels of depression (33.1 percent of LGBT students compared to 19.8 percent of non-LGBT students), to have seriously considered attempting suicide (32.2 percent of LGBT students compared to 13.2 percent of non-LGBT students), and to have made a suicide attempt (21.8 percent of LGBT students compared to 4.3 percent of non-LGBT students). Students who were more highly victimized were also more likely to experience these negative outcomes. For example, compared to students who experienced little or no anti-LGBT victimization, students who experienced a great deal of anti-LGBT victimization were more than three times as likely to have been threatened or injured with a weapon on school property (10.0 percent compared to 3.6 percent), more than three times as likely to have been threatened or injured with a weapon off school property (14.7 percent compared to 4.3 percent), more than three times as likely to have experienced school violence (35.2 percent compared to 10.6 percent), more than four times as likely to have attempted suicide (25.0 percent compared to 5.7 percent), and more than six times as likely to have made a suicide attempt that required treatment from a doctor or nurse (10.4 percent compared to 1.8 percent).\n\nIntroduction\n\n\nThis report presents data from the National School Climate Survey (NCS), which collected data from over 7,000 lesbian, gay, bisexual, and transgender (LGBT) students ages 13\u201321 in public and private schools in every state and the District of Columbia during the spring of 2015. The report focuses on the experiences of LGBT students in public schools, but private school students' experiences are briefly discussed to provide context and additional information about the experiences of LGBT students in schools across the country.1 This report also includes a limited analysis of the data on the experiences of bisexual youth. This report provides the first national data on LGBT students' experiences in our nation's schools.\n\n\nLGBT students face a number of serious challenges at school, and this report provides a comprehensive look at the extent of the problem and its impact on LGBT students. Many LGBT students have experienced verbal harassment, physical assault, and sexual violence at school. For example, over half (53.1 percent) of LGBT students have felt unsafe at school because of their sexual orientation, over one-quarter (26.1 percent) because of their gender expression, and over one-third (35.2 percent) because of both their sexual orientation and gender expression. In addition, over one-third (35.2 percent) of LGBT students have been verbally harassed and 19.6 percent have been physically attacked at school in the past year because of their sexual orientation or gender expression. These figures are even higher for transgender students: over half (53.8 percent) have felt unsafe at school because of their gender expression and over one-third (36.2 percent) because of their gender identity. Furthermore, over one-quarter (27.2 percent) of LGBT students have missed at least one entire day of school in the past month because they felt unsafe or uncomfortable, and the prevalence of truancy increases further for transgender students: 40.1 percent of transgender students have missed at least one day of school in the past month because they felt unsafe or uncomfortable.\n\n\nThis report also shows that the harassment and discrimination that LGBT students experience can lead to serious and negative outcomes for these students, including lower school engagement, lower levels of happiness and life satisfaction, and higher levels of depression and even higher levels of suicidal behavior. Over one-third (34.7 percent) of LGBT students reported missing school for at least one day during the month before the survey because they felt unsafe or uncomfortable. Students who experienced more victimization because of their LGBT status were also more likely to report this type of adverse outcome. For example, compared to students who experienced little or no victimization, students who experienced a great deal of victimization were almost four times as likely to have missed at least one day of school in the month before the survey because they felt unsafe or uncomfortable (45.6 percent compared to 12.3 percent).\n\n\nThe negative impact of this victimization on LGBT students can be seen in their reports of lower levels of engagement in school activities, lower levels of happiness and life satisfaction, and higher levels of depression. LGBT students who experienced high levels of victimization because of their sexual orientation, gender expression, or both were more likely than students who experienced low levels of victimization to feel that they could not be themselves at school (68.6 percent compared to 42.2 percent) and to feel unhappy or depressed at school (40.7 percent compared to 19.5 percent). Students who experienced high levels of victimization because of their sexual orientation, gender expression, or both were also more likely to report symptoms of depression than students who experienced low levels of victimization. For example, compared to students who experienced little or no victimization, students who experienced high levels of victimization were almost twice as likely to report feeling unhappy or depressed (38.7 percent compared to 20.2 percent) and almost three times as likely to report feeling less happy than other students (14.3 percent compared to 5.4 percent).\n\n\nData from the 2015 NCS show that many students still face serious and pervasive mistreatment and discrimination at school. These experiences can have a serious negative impact on students' well-being. This is especially true for transgender and gender non-conforming students who are most likely to]" time="0.076"><properties><property name="score" value="0.13252375673529498" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The band's creative soul and ringleader, Robert Pollard, reflects on the wild success of Guided By Voices as well as his own triumphs and tragedies.\n\nRobert Pollard \u2014 the creative soul and ringleader of Guided By Voices \u2014 is a living embodiment of the enduring power of rock \u2018n\u2019 roll. Pollard, whose band just released its 15th album, August By Cake, has lived a life that\u2019s been as varied as it has been eventful. From crafting classic lo-fi pop records and guiding his beloved group to victory after victory for over two decades, to overcoming multiple tragedies that threatened to undo him, Pollard has endured and endured well.\n\nFrom rock \u2018n\u2019 roll heartbreak to musical triumphs and to the stories of his childhood friends who took their own lives, the tales behind Pollard\u2019s life are as varied as the songs he\u2019s written over the years. They\u2019re full of sweetness, heartache and inspiration \u2014 the kind that only music can provide.\n\nHere, Pollard reflects on the enduring and wonderful power of rock \u2018n\u2019 roll.\n\nOn being inspired by Bob Dylan\n\n\u201cHe made rock \u2018n\u2019 roll feel alive to me, you know? He just did things no one else was doing. The first time I heard his music, I was just in awe of it. I thought he was one of the only people doing something different in music and it really inspired me to make a rock band.\u201d\n\nOn working at Tower Records\n\n\u201cI worked at Tower Records on Sunset Boulevard in the Eighties. I was in a band that was signed to a major label and we were going to go on tour. We were going to make our first big move out to L.A. and the band broke up in Pittsburgh. I was heartbroken. I went to L.A. anyway and got a job at Tower and the band started to do really well in Pittsburgh. I started to realize that I was being a moron by working at a job that didn\u2019t have anything to do with music. I should have been playing music.\n\n\u201cThen, I met people that worked at the store \u2014 people that worked in the music section \u2014 and they were people that were into the same kind of music that I was. They were all very into the new wave thing that was going on at that time and I was very into the punk scene and I was discovering all this underground music. That was a good time because I was discovering so many great bands and all these great people.\u201d\n\nOn putting together Guided By Voices\n\n\u201cI was living in this little house on Cass Avenue in Detroit and I would just record these songs in the basement. I didn\u2019t think anyone would be interested. I just thought I was doing this for myself. Then, people started calling me and saying they wanted to play on the songs and asking if I could make copies of them. Eventually, I had enough to make a record. I was putting the band together around the same time.\n\n\u201cIn the beginning, it was just one song. The whole point was to just have one song on a compilation of different artists from Michigan. That was the point of the whole thing. It was like, \u2018Oh, that\u2019s fun. That\u2019s really fun. Let\u2019s do another song.\u2019 It just became something else. Then, I started working with the other people. The other people were my friends. My friends were just my buddies that I\u2019d played music with. The first couple of years was just me and my buddies doing these songs. Then, it started to become a bigger thing and we started to tour and it became something else. It just turned into something else. It was an accident that it happened.\u201d\n\nOn his songwriting approach\n\n\u201cMy songs have always come from an unconscious, naive place. I really have to focus on a song and work on a song in order for it to have a definite melody or a definite idea. The words, I kind of write more like a novel and I try to write songs that work like a story. When I\u2019m doing them, I\u2019m trying to do it like a movie. I\u2019m trying to think of it in terms of a movie and I\u2019m trying to create characters and set the scenes and tell the story and make it a little bit deeper than what you would normally expect from a pop song. I\u2019m trying to take it to a different level in some ways. That\u2019s always been my goal and that\u2019s always been my style.\n\n\u201cI don\u2019t think I\u2019m trying to reinvent the wheel, I just think that I\u2019m trying to take things that are already in existence and make them bigger and more beautiful and a little more complete. I don\u2019t know. I don\u2019t really try to analyze it that much. I\u2019ve always written songs in the same way and the only thing that\u2019s changed is the way I record them. I used to record them all on a boombox and then, as we went along, we started to get more and more into studio work.\u201d\n\nOn recording songs at home\n\n\u201cI like to record in a place that\u2019s not a studio. I\u2019ve always liked to record in a place where I could have access to as much stuff as possible. I really enjoy it when I have a lot of cool things to play with. I always like to record at home. I think it\u2019s important to make the place you record at like a laboratory, you know? I like to have it be a lab. If I could, I\u2019d have every piece of equipment ever made and be able to put it all in one room and be able to experiment with every sound I could possibly make. It\u2019s so much fun and it\u2019s so much work at the same time, too. I love it.\u201d\n\nOn his songwriting process\n\n\u201cThe way I work, I just keep track of my thoughts and feelings. I think about it when I\u2019m walking or when I\u2019m eating or I\u2019m just doing stuff. I just keep track of stuff and I write down a lot of notes and I try to pay attention to things. My life is so boring, really, you know? I don\u2019t do anything that exciting. I just pay attention to things and I think about stuff and I write down little ideas. I do that all the time. I write little ideas all the time and I just try to remember them and try to put them in a folder.\n\n\u201cIt\u2019s kind of like a filing system. The filing system is in my head. I try to remember everything I think about and I try to write it down and then, eventually, it becomes a song. Then, I try to look at the songs that are done and I try to pick the ones that seem to be the best and start thinking about them in terms of a story or a character or a scenario or something that would be entertaining. I have to think about all the things that are going on in my life and try to piece things together like a collage.\n\n\u201cThen, when it comes to the recording, I\u2019m thinking about the whole thing as a complete package. I\u2019m thinking about the sound of it and I\u2019m thinking about how I\u2019m going to mix the album and everything. I\u2019m thinking about the lyrics and the voice. Everything is kind of part of a bigger thing and the sound and the performance is really important.\u201d\n\nOn keeping it simple\n\n\u201cIt\u2019s easier to write when you\u2019re just keeping it simple. When you start to complicate things, it becomes too big of a thing to think about. That\u2019s when the song becomes too precious and you can\u2019t finish it or you start getting scared of it or something like that. Then, you get caught up in the emotion of it and you can\u2019t do it. You have to remember that it\u2019s just something you\u2019re doing to pass the time. It\u2019s just something you\u2019re doing for fun. That\u2019s what it is for me. It\u2019s just fun. I\u2019m not trying to be any big shot or anything like that. I\u2019m just trying to do something that\u2019s fun and I\u2019m just trying to make something that\u2019s going to give me some satisfaction. I\u2019m just trying to do something that\u2019s going to be fun to listen to and fun to make.\u201d\n\nOn his songwriting being like therapy\n\n\u201cYeah. I\u2019m not trying to be a serious writer or anything like that. I\u2019m not trying to be pretentious at all. I just want to make a record that I\u2019d like to listen to.\u201d\n\nOn the joy of performing\n\n\u201cIt\u2019s a good feeling, you know? There\u2019s something very satisfying about it. You just can\u2019t stop it, you know? I\u2019ve been doing it for so long and it just doesn\u2019t go away. You can\u2019t stop it. I just keep on doing it and I can\u2019t stop it. I think that\u2019s the reason why I do it, you know?\u201d\n\nOn discovering The Beatles\n\n\u201cThe Beatles made]" time="0.074"><properties><property name="score" value="0.001664983761440044" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00166498&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00166498
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[CHICAGO (AP) \u2014 An American Airlines pilot has been removed from flying duties after he temporarily parked a plane on a taxiway where four other airliners were waiting to take off at Chicago's O'Hare International Airport.\n\nAn investigation into the Saturday incident is ongoing, airline spokesman Ross Feinstein said Sunday.\n\nFeinstein said the pilot has been taken off flying duty &quot;pending the outcome of the investigation.&quot; He didn't specify how long the pilot would be grounded.\n\nNo other crew members were disciplined, and the incident won't affect the scheduling of the roughly 3,000 flights a day the airline operates around the world, Feinstein said.\n\nHe declined to release the name of the pilot or the number of years he had been flying for the airline.\n\nThe Chicago Department of Aviation said in a statement that the pilot of the American Airlines plane, which had arrived from Nashville, Tenn., was supposed to go to a designated taxiway but instead went to a parallel taxiway where four other airliners were waiting for clearance to take off.\n\nIt wasn't immediately clear how close the plane came to taking off.\n\nAccording to radar information cited in the aviation department statement, the aircraft came within 100 feet of an aircraft on the ground and approximately 200 feet from another aircraft.\n\nThe aviation department is investigating the incident along with the Federal Aviation Administration.\n\nAmerican Airlines has endured two crashes this year and five in 2015. Saturday's incident is unlikely to do much to allay the fears of passengers and the traveling public.\n\n&quot;It's obviously one of those things that happen occasionally. It's a pretty serious mistake,&quot; said Henry Harteveldt, founder of the Atmosphere Research Group. &quot;It's a reminder that airline pilots are human beings and that despite their extensive training, sometimes they make mistakes.&quot;\n\n___\n\nAssociated Press writer Don Babwin contributed to this report.]" time="0.028"><properties><property name="score" value="0.001471302126998965" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0014713&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0014713
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[As a manufacturer, it is easy to produce one-of-a-kind pieces, but even easier to make a product for everyone, because every piece is made to the same specifications. What happens when you decide to make a new product and to have a bigger impact? You might think that you have to start with an already existing product and create something new. But this is not always true.\n\nSometimes you need to start from scratch and create something new \u2013 a new form, a new material, new applications or new customer experiences. This is what happened when Ikea created the L\xd6MSK light collection \u2013 a range of products made of plaster and leather, using two materials that we usually associate with other products.\n\nL\xd6MSK: it\xb4s easy being a one-of-a-kind light\n\nTwo elements were the starting point of the L\xd6MSK light collection: an existing material and a very unusual use for it. Plaster and leather are very traditional materials that are linked to very different products. We know plaster for its use in wallpapers, ceilings and sculptures, but it is not very common to use it in products such as lights or furniture. Leather has a much stronger association with leather goods, clothes, bags or shoes. And in fact, leather is one of the most used materials in the furniture industry.\n\nTwo products, plaster and leather, that are very common to different types of industries but never used together.\n\nIkea wanted to combine these two materials in a collection of lights, lamps and other decorative elements, in order to make a complete lighting system that can be easily incorporated into any interior and made more personal.\n\nA creative team started exploring ways of using these two materials in a complete lighting collection. They started experimenting with different elements such as the light bulb and the suspension, to combine them and to make sure they can adapt to the maximum number of situations. The development of the light was not a process of making a more perfect product, but rather a process of exploring new ways of using existing products.\n\nAs a result, they developed three series of products. One made of plaster and leather, and two others made of plaster. The leather light is one of the most surprising elements in the collection, as it is hard to believe that this kind of leather is the same that we see in jackets or shoes.\n\nThe leather used in the light is dyed in different colours in order to add interest to the design and it is used to create the \u201cU\u201d shape that holds the light bulb. This element was designed to be a flexible element in the light, allowing the user to change the direction of the light and adapt it to different uses.\n\nFinally, the leather light can also be suspended with a wire, making it ideal for reading lights.\n\nThe second series, made of plaster, is the plateroom series, and the third series is a collection of lamps made of plaster with different shades. The aim was to use plaster to make different types of lamps that can adapt to the different sizes of space. This is achieved by combining the plaster with different shades and creating elements such as the plateroom bench and the plateroom table, with different dimensions and shades.\n\nThere is an important message that Ikea wants to communicate with this collection: that it is possible to create something new and original.\n\nYou need to start with something simple but at the same time creative and unusual. This is what Ikea wants to achieve with this collection: creating a new product that is easy to understand and appreciate, but at the same time has an innovative design and can be used in different types of interior.\n\nThis can be achieved by starting with a very simple element and having an open mind, as the designers did with this collection. Ikea has created a very unusual product that is adaptable to different situations and can be adapted to any kind of interior or architecture.]" time="0.066"><properties><property name="score" value="0.03220854382897711" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Good for price\n\nBy Anonymous\n\n06 Apr 2012\n\nShimano PD-T400 SPD Pedals\n\n3 / 5\n\nI recommend this product\n\nCons rubber grip strip wears quickly, not made to wear a cleat off road or in the mud Pros inexpensive, comfortable, easy to adjust, replaceable\n\nMy first set lasted me 4 years on the commuter bike and MTB with up to 50 mile/week rides. The rubber grips are made to wear down quickly, and don't offer a lot of friction to hold on to (that's good and bad), but replaceable. The cleat is replaceable too, which is nice if you walk on the pedals a lot (they do have walkable cleats).\n\n\n\nOverall, a great value pedal for commuter bikes. The tension is adjustable, but it is easy to over-tighten and bend the pedal.]" time="0.015"><properties><property name="score" value="0.068245611438289" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06824561&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06824561
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Movie details\n\n7.5 / 10 ( 189 )\n\nAn old friend of Archie's has been released from prison and needs a place to stay. Archie offers his hospitality to &quot;Leroy&quot; but has a hard time adjusting to this out-of-place visitor. The ...]" time="0.014"><properties><property name="score" value="0.02055853456952028" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02055853&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02055853
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The Pennsylvania governor's race is getting a jolt from two longshot third-party candidates \u2014 Libertarian Ken Krawchuk and Green Party candidate Paul Glover.\n\nThe two men were on the ballot with Republican Tom Corbett and Democrat Dan Onorato until the parties removed them. Onorato sued to keep Krawchuk on the ballot, but the state Supreme Court on Monday said he could be dropped.\n\nKrawchuk says he's energized by the attention. He plans a statewide tour and calls Corbett &quot;the worst candidate for governor&quot; in the state's history.\n\nGlover is campaigning in Philadelphia and plans a statewide tour with the Green Party, but he's unlikely to win.\n\nCopyright Associated Press]" time="0.014"><properties><property name="score" value="0.05897918034036831" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05897918&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05897918
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Karen &amp; Ryan, 9-6-08\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n]" time="0.018"><properties><property name="score" value="0.0117425246955849" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01174252&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01174252
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Jubal Early\n\nJubal Anderson Early (July 14, 1816July 2, 1894) was a Confederate general in the American Civil War and a politician after the war. Although he served almost the entire war in the eastern theater, his most significant contributions were in the Valley Campaigns of 1864, which ended in the stalemate at Petersburg, Virginia, and the capture of the Union garrison at Lynchburg, Virginia. Early is considered among the best Confederate generals who never served in the Army of Northern Virginia.\n\nEarly was born in Southampton County, Virginia, the seventh of ten children of Ruth (Anderson) Early and Judge John Early, a former governor of Virginia and member of the United States Congress. In 1825, Early attended Charlotte Hall Military Academy. He graduated second in his class at the United States Military Academy at West Point in 1838, placing him above such future generals as Stonewall Jackson and Ulysses S. Grant.\n\nEarly was appointed a second lieutenant in the 4th U.S. Artillery. He served as an aide to General Winfield Scott in the War with Mexico and fought in the Battle of Molino del Rey and the Battle of Chapultepec. Early served as the acting chief of artillery in the assault on the city of Mexico, and received two brevets for his bravery in these battles. He was one of the first officers to enter Mexico City, and he served as an artillery instructor at West Point from 1847 to 1852. Early served as a major in the United States Army in the Seminole Wars in 1855, and as a captain in the United States Dragoons in 1858.\n\nEarly entered the Virginia Military Institute in 1852, and became a professor of mathematics at the Institute, where he taught until the Civil War. He was known for his severity as a disciplinarian. Early became known as a superb military tactician and administrator, as well as an able administrator and strategist.\n\nIn 1860, Early entered politics, running as a Democrat for one of the two seats in the Virginia House of Delegates from Southampton County, but he was defeated. The state seceded from the Union on April 17, 1861, and Early was commissioned as a colonel of artillery in the Virginia militia on May 14. When Virginia militia units were activated for Confederate service, Early was appointed a brigadier general of Virginia forces on May 27.\n\nEarly initially commanded a brigade in the brigade of Brig. Gen. Richard B. Garnett. When Garnett died at the Battle of Mill Springs, Early was promoted to brigadier general on January 16, 1862. His brigade was assigned to Maj. Gen. Edmund Kirby Smith's division in the Army of Northern Virginia. Early fought in the Peninsula Campaign, including the Siege of Yorktown, during which his brigade was an important factor in repulsing a Union assault on the right flank of the Confederate entrenchments at Savage's Station. Early received praise for his actions at Hanover Court House.\n\nEarly distinguished himself during the Seven Days Battles. He received his commission as a major in the Confederate Regular Army on June 9, 1862, but was later returned to his previous rank of colonel in the reserves. During the Northern Virginia Campaign, he received plaudits for his actions at the Second Battle of Bull Run.\n\nEarly was promoted to brigadier general in the Confederate regular army on October 11, 1862. He was given command of a brigade in Maj. Gen. Robert E. Rodes's division of A.P. Hill's III Corps. During the Maryland Campaign, Early performed well at the Battle of South Mountain.\n\nDuring the Battle of Fredericksburg, Early was severely wounded in the right thigh and temporarily assigned to administrative duties. He returned to the field in time to defeat Union Maj. Gen. Joseph Hooker's offensive at the Battle of Chancellorsville. At Chancellorsville, Early performed ably in combat and earned praise for his conduct as a brigade commander. Early was sent to western Virginia, where he defeated Union Maj. Gen. William S. Rosecrans at the Battle of Mill Springs. After Chancellorsville, Early was promoted to the permanent rank of brigadier general in the Confederate Regular Army on May 21, 1863, and given command of a division.\n\nEarly fought under Maj. Gen. Thomas J. &quot;Stonewall&quot; Jackson during the Valley Campaigns of 1864. Early's old brigade was reassigned to Maj. Gen. Richard H. Anderson. The Second Battle of Kernstown was his most significant contribution during the Valley Campaigns. Following the Valley Campaigns, Early was assigned to Maj. Gen. John C. Breckinridge's command in the Shenandoah Valley.\n\nHe conducted the Valley Campaigns of 1864. Following Early's victory at the Battle of Lynchburg in June 1864, Early's army became separated by Union Maj. Gen. David Hunter's raid. Hunter's force was driven from the Valley in the Battle of Lynchburg by Early's troops. During the Valley Campaigns, Early defeated Union Maj. Gen. George Crook at the Battle of Rutherford's Farm and the Battle of Tom's Brook, drove Maj. Gen. David Hunter from the Valley in the Second Battle of Kernstown, and defeated Union Maj. Gen. Philip Sheridan at the Third Battle of Winchester.\n\nEarly defeated Union Maj. Gen. David Hunter's Valley Campaign in the Second Battle of Kernstown in July 1864. He then crossed into Maryland and threatened Washington, D.C., forcing Grant to send reinforcements to the capital area. During the subsequent Valley Campaigns of 1864, Early defeated Maj. Gen. Philip Sheridan at the Battle of Opequon, the Third Battle of Winchester, and the Battle of Fisher's Hill. Sheridan's victory at the Battle of Cedar Creek forced Early to retreat back to the Valley.\n\nEarly was assigned to the Department of the West and ordered to reinforce the Confederate Army of Tennessee, but Maj. Gen. John C. Breckinridge convinced him to resist Maj. Gen. William T. Sherman's advance. At the beginning of the Atlanta Campaign, Early was given command of the newly created Third Corps of the Army of Tennessee. During the Battle of Resaca, he launched an unsuccessful assault against a Union fortification on the western bank of the Oostenaula River, suffering heavy casualties. During the subsequent battles of Adairsville and Cassville, his corps attempted to flank Union Maj. Gen. William T. Sherman's left flank, but his subordinates were slow in attacking and failed to coordinate their movements with other Confederate forces. After the Confederates suffered defeat at the Battle of Peachtree Creek, Early retreated with his corps to Lovejoy's Station. He ordered Maj. Gen. Benjamin F. Cheatham's division to open the Battle of Atlanta by attacking the Union force under Maj. Gen. James B. McPherson. Early was in the town of Marietta when he learned that Lt. Gen. John B. Hood was wounded during the Battle of Atlanta, and he returned to Atlanta to take command of the army.\n\nDuring the subsequent Siege of Atlanta, the Army of Tennessee lacked provisions and was forced to withdraw from Atlanta. The army was defeated during the subsequent pursuit, and Early's corps was heavily engaged at the Battle of Ezra Church. In November 1864, the Army of Tennessee was reorganized and Early was given command of the new Second Corps. At the Battle of Franklin, he was killed by a shot through the heart, and his corps suffered significant casualties. The Second Corps was commanded by Maj. Gen. William B. Bate after Early's death. Bate was also wounded and captured, and the corps was virtually destroyed. Bate's capture led to another shakeup of the army's command structure, and Lt. Gen. Alexander P. Stewart assumed command of the Second Corps.\n\nEarly was buried at the Old Methodist Cemetery in Alexandria, Virginia, but the cemetery closed and his remains were moved to Richmond National Cemetery.\n\nThe Virginia Historical Society operates the Jubal Early Memorial Cemetery near New Market, Virginia, where he is buried.\n\nEarly was known as &quot;Old Jubal&quot; to his men and &quot;Fighting Joe&quot; to his friends. Confederate General Robert E. Lee was a friend of Early's, and named his youngest son, Jubal Anderson, after him.\n\nIn the film &quot;Gettysburg&quot;, Early was portrayed by actor Martin Sheen. The movie presented Early as a vain, glory-seeking but weak commander who bickered with fellow Confederate generals Joseph E. Johnston and James Longstreet over who was in command of the Army of Tennessee.\n\nHe is also featured in Jeff Shaara's &quot;The Last Full Measure&quot; trilogy, where he is portrayed by Jon Voight.\n\nThe James River Squadron was a Confederate Navy organization during the American Civil War which operated on the James River, Virginia.\n\n\n\n\n]" time="0.103"><properties><property name="score" value="0.04661751300473742" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04661751&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04661751
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Al-Hamzah, Yemen\n\nAl-Hamzah is a village in western central Yemen. It is located in the San\u2018a\u2019 Governorate.\n\n]" time="0.020"><properties><property name="score" value="0.01760621691104247" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01760622&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01760622
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Bangalore United\n\nBangalore United is a men's team of the Bangalore United Football Club.\n\nIn the past Bangalore United Football Club had two men's teams, Bangalore FC and Bangalore United.\n\nBangalore United FC was a founder member of the I-League in 2007 and competed in the first season of the competition, finishing in third place in the Southern Group. The club withdrew from the 2008\u201309 season for financial reasons and since then have played in the Karnataka State Football Association's Leagues and the Bangalore Super Division.\n\n\nThe following players are registered for the 2014\u201315 I-League.\n\n&quot;As of 5 January 2015&quot;\n\n]" time="0.019"><properties><property name="score" value="0.004148158144318126" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00414816&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00414816
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Cirrochroa hieroglyphus\n\nCirrochroa hieroglyphus, the Hieroglyph]" time="0.016"><properties><property name="score" value="0.0009980230130490764" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00099802&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00099802
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Colorimetric Detection of Bovine Serum Albumin by Sodium Dodecyl Sulfate-Boronate Affinity Chromatography with Conventional Staining: Optimization by Factorial Design\n\nD. J. Lindstedt, F. C. Hanft, and J. P. Weese\n\nCereal Chemistry\n\nVolume 67(4), 380-383\n\nJune 1990\n\nThe detection of serum albumin in bovine milk by sodium dodecyl sulfate-borate affinity chromatography was investigated with the use of Con A stain. After optimization, it was determined that a total of 6.6 ml of bovine plasma (0.5 mg protein/ml) and 1.2 ml of milk sample were sufficient for the analysis of 0.2 ml of milk. A total of 35 ml of buffer A (20 mM sodium phosphate pH 7.0) containing sodium dodecyl sulfate, boronate, Con A, and 0.5 M sodium sulfate were added to an 8 \xd7 100 mm borosilicate glass column. Bovine plasma and milk samples were passed through the column at a flow rate of 1.5 ml/min. The flow-through was collected and fractions were obtained by passing 1.5 ml of fresh buffer A through the column. Bovine serum albumin was detected in the fractions after staining with the Con A. The mean recovery was 2.78 \xb1 0.01% for bovine plasma and 1.28 \xb1 0.07% for bovine milk.\n\nGo to the article in Cereal Chemistry\n\nKeywords: Milk and dairy products, Bioactive compounds, Animals, Mammals, Bovines, Colostrum, Proteins, Bovine, Dried-milk products, Dairy products, Whole milk, Milk fat, Milk-fat globule membranes, Caseins, Sodium dodecyl sulfate, Serum albumin, Affinity chromatography, Sodium dodecyl sulfate-borate, Concanavalin A stain, Affinity chromatography, Affinity separations, Immobilized enzymes]" time="0.031"><properties><property name="score" value="0.08187451725697105" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08187452&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08187452
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Hi Guys,We need your help.We are writing an essay about a topic of our choice and we need help with a paragraph.The topic is:&quot;The more mobile and connected we are, the more isolated we feel. Is it because the more we interact the more we expect? Is it because our friends are not always available? We are flooded with so many information but we don\u2019t know how to manage it and, in a way, are lost in it. The real question is, how can technology provide more human connections and make us less lonely? &quot;Please feel free to share any ideas you have.Thanks and have a great day.Thesis = + http://www.plenglish.com/index.php?option=com_content&amp;view=article&amp;id=9288:2014-04-21-22-52-20&amp;catid=40:tecnologia-ciencia&amp;Itemid=51]" time="0.013"><properties><property name="score" value="0.06380637224008279" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06380637&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06380637
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Junior Potchefstroom Boys\n\nBoys' Junior Cup\n\n04 April 2019\n\nGrade: Gauteng North (South)\n\nDivision: Pool A\n\nPool: A\n\nResults\n\nGame 1 (09:30)\n\nPlayStation/Yello Pockets v Athlone (Kingfisher Blue 2)\n\nWin: 3 wickets\n\nAthlone won the toss and elected to field\n\nPlayStation/Yello Pockets - 75/10 in 10 overs\n\nKhawu v Chamuhiya 2 runs\n\nChirag v Alam 8 runs\n\nQalbya v Singh 8 runs\n\nRyan v Daggupan 6 runs\n\nSantlal v Taufik 5 runs\n\nTiaan v Alam 0 runs\n\nScorecard\n\nAthlone - 68/10 in 9.4 overs\n\nAwal v Ryan 16 runs\n\nSeqib v Santlal 4 runs\n\nImran v Qalbya 4 runs\n\nShahir v Chirag 2 runs\n\nMubarak v Kole 2 runs\n\nAhmed v Chamuhiya 4 runs\n\nFarook v Kabbya 5 runs\n\nNadeem v Ahli 6 runs\n\nScorecard\n\nPoints: Athlone 2, PlayStation/Yello Pockets 2\n\nGame 2 (10:15)\n\nClub Greenheights v Kelvin Stars (Kingfisher Blue 1)\n\nWin: 0 wickets\n\nKelvin Stars won the toss and elected to bat\n\nKelvin Stars - 131/6 in 20 overs\n\nAsela v Johnson 7 runs\n\nSmith v Shakoor 5 runs\n\nPaliwal v Hallahan 4 runs\n\nDipoo v Dabhi 4 runs\n\nFernando v Dabhi 3 runs\n\nJo v Rajoo 9 runs\n\nStrydom v Kumar 10 runs\n\nGoswami v Basheer 1 run\n\nScorecard\n\nClub Greenheights - 76/6 in 12 overs\n\nVusumuzi v Smith 16 runs\n\nAubrey v Jo 6 runs\n\nAlen v Fernando 3 runs\n\nChase v Dabhi 3 runs\n\nHaque v Johnson 3 runs\n\nNjoo v Fernando 4 runs\n\nAhmed v Dabhi 1 run\n\nSheehy v Gupta 3 runs\n\nScorecard\n\nPoints: Kelvin Stars 2, Club Greenheights 2\n\nGame 3 (11:00)\n\nClub Greenheights v Athlone (Kingfisher Blue 2)\n\nWin: 3 wickets\n\nAthlone won the toss and elected to bat\n\nAthlone - 90/7 in 10 overs\n\nAwal v Chamuhiya 2 runs\n\nSeqib v Chase 5 runs\n\nAhmed v Njoo 1 run\n\nFarook v Sheehy 0 runs\n\nAhmed v Basheer 3 runs\n\nShahir v Singh 1 run\n\nFernando v Dabhi 2 runs\n\nMubarak v Kumar 0 runs\n\nScorecard\n\nClub Greenheights - 92/3 in 9.2 overs\n\nVusumuzi v Dabhi 9 runs\n\nAubrey v Kumar 7 runs\n\nChase v Seqib 1 run\n\nAhmed v Fernando 3 runs\n\nKeshi v Gupta 5 runs\n\nNjoo v Patel 2 runs\n\nSheehy v Rajoo 3 runs\n\nHaque v Ahmed 1 run\n\nScorecard\n\nPoints: Athlone 2, Club Greenheights 2\n\nGame 4 (12:00)\n\nPlayStation/Yello Pockets v Kelvin Stars (Kingfisher Blue 1)\n\nWin: 3 wickets\n\nKelvin Stars won the toss and elected to bat\n\nKelvin Stars - 72/10 in 10 overs\n\nDabhi v Sheehy 0 runs\n\nSmith v Hasan 5 runs\n\nPaliwal v Rahman 1 run\n\nFernando v Hasan 5 runs\n\nJo v Chase 2 runs\n\nDipoo v Njoo 1 run\n\nDabhi v Shaheed 5 runs\n\nSmith v Shaheed 3 runs\n\nScorecard\n\nPlayStation/Yello Pockets - 76/4 in 8.4 overs\n\nKhawu v Dabhi 10 runs\n\nChamuhiya v Smith 7 runs\n\nAhli v Dabhi 2 runs\n\nRyan v Dipoo 4 runs\n\nSantlal v Shaheed 7 runs\n\nQalbya v Fernando 5 runs\n\nTiaan v Jo 1 run\n\nKole v Fernando 2 runs\n\nScorecard\n\nPoints: Kelvin Stars 2, PlayStation/Yello Pockets 2\n\nGame 5 (13:00)\n\nPlayStation/Yello Pockets v Kelvin Stars (Kingfisher Blue 2)\n\nWin: 0 wickets\n\nKelvin Stars won the toss and elected to bat\n\nKelvin Stars - 95/7 in 10 overs\n\nAhmed v Chamuhiya 0 runs\n\nFarook v Aucho 3 runs\n\nAhmed v Aucho 5 runs\n\nShahir v Chamuhiya 5 runs\n\nFernando v Aucho 2 runs\n\nMubarak v Hasan 5 runs\n\nSmith v Hasan 0 runs\n\nDabhi v Njoo 1 run\n\nScorecard\n\nPlayStation/Yello Pockets - 98/5 in 8.4 overs\n\nKhawu v Ahmed 3 runs\n\nChamuhiya v Farook 7 runs\n\nAhli v Hasan 7 runs\n\nRyan v Shaheed 3 runs\n\nSantlal v Fernando 3 runs\n\nQalbya v Shaheed 7 runs\n\nTiaan v Dabhi 3 runs\n\nKole v Dabhi 3 runs\n\nScorecard\n\nPoints: Kelvin Stars 2, PlayStation/Yello Pockets 2\n\nGame 6 (14:00)\n\nClub Greenheights v Athlone (Kingfisher Blue 2)\n\nWin: 3 wickets\n\nAthlone won the toss and elected to bat\n\nAthlone - 89/9 in 10 overs\n\nAwal v Shaheed 3 runs\n\nAhmed v Dabhi 1 run\n\nShahir v Rajoo 3 runs\n\nFarook v Shaheed 2 runs\n\nAhmed v Patel 3 runs]" time="0.068"><properties><property name="score" value="0.024947218532766283" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02494722&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02494722
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Andr\xe9 Le N\xf4tre \u2013 Andr\xe9 Le N\xf4tre, also known as Le N\xf4tre, was a landscape architect, responsible for the design of the gardens of many of the most famous French ch\xe2teaux of the 17th century, including Versailles. He is popularly known as the father of the profession and he was born in Paris, the son of Pierre Le N\xf4tre, a painter of miniatures and Marie Le N\xf4tre. He was trained in his fathers occupation, he did some paintings by the time he was sixteen. He travelled to Italy with Pierre Berchet in 1640, and worked with the architect Jacques Lemercier, from 1646 he was occupied with architectural drawing for the Louvre. He was made a member of the Acad\xe9mie at Mans in 1650, Le N\xf4tres first known landscape project was at the ch\xe2teau of Vaux-le-Vicomte, for his distant relative, Nicolas Fouquet. Here he was more open to the imagination, transforming the plan into a Picturesque and Romantic garden. His work showed a tendency towards the natural and asymmetrical, Le N\xf4tre was unable to visit Italy on account of wars with the Italian states. His influence was limited to parts of the northeast of France, Le N\xf4tres first projects in this style were done for the court of the prince de Cond\xe9. The ch\xe2teau of Chantilly, belonging to the princess of Cond\xe9, Fran\xe7ois, Le N\xf4tre achieved the most at the Ch\xe2teau de Saint-Germain-en-Laye, for Louis XIV and his duc de Bourgogne, later the duc dOrl\xe9ans. Le N\xf4tre created a estate at Marly, and, at the Trianon, for the young Louis XIV. His work transformed the old fort into a palace, more suitable for the residence of the king. He is also responsible for the landscapes of the gardens of the Ch\xe2teau de Maisons and Ch\xe2teau de Saint-Cloud, Le N\xf4tre died in 1681, and was succeeded by his son Jean-Charles Le N\xf4tre, to whom he had taught his encyclopedic style, and his son-in-law Fran\xe7ois Dunan. The design of the gardens of the Ch\xe2teau de Marly was a project of Le N\xf4tres. They were designed to be seen from the top of the terrace, the perspectives and size of the gardens were calculated to surprise the spectators. Le N\xf4tre planned the gardens following the Renaissance principles of regularity, symmetry and he was also responsible for the modifications made to the original plan, done by Charles Le Brun and the gardener Martin Charbonnier, who remade the terrace and the parterres. The final result was the ancestor of modern formal gardens, in 1660 and 1661, Le N\xf4tre built two mansions on the Place Royale in Paris, for the jewellers Louis de Lancy and Balthazar Claquin. In 1664, Le N\xf4tre and Claude Perrault collaborated on a project for the west side of the Louvre. Le N\xf4tres last landscape was the garden around the ch\xe2teau de Sceaux, a project which was incomplete at the time of his death in 1681\n\n9. Gardens of Versailles \u2013 The Gardens of Versailles occupy part of what was once the Domaine royal de Versailles, the royal demesne of the ch\xe2teau of Versailles. Situated to the west of the palace, the gardens cover some 800 hectares of land, in addition to the meticulous manicured lawns, parterres of flowers, and sculptures are the fountains, which are located throughout the garden. Dating from the time of Louis XIV and still using much of the network of hydraulics as was used during the Ancien R\xe9gime. On weekends from spring to early autumn, the administration of the museum sponsors the Grandes Eaux \u2013 spectacles during which all the fountains in the gardens are in full play. In 1979, the gardens along with the ch\xe2teau were inscribed on the UNESCO World Heritage List, one of thirty-one such designations in France. With Louis XIII\u2019s final purchase of lands from Jean-Fran\xe7ois de Gondi in 1632 and his assumption of the role of Versailles in the 1630s. Records indicate that late in the decade Claude Mollet and Hilaire Masson designed the gardens and this early layout, which has survived in the so-called Du Bus plan of c.1662, shows an established topography along which lines of the gardens evolved. This is evidenced in the definition of the main east-west. From this point forward, the expansion of the gardens of Versailles followed the expansions of the ch\xe2teau, accordingly, Louis XIVs building campaigns apply to the gardens as well. At every stage the prescribed tour was managed, under the Sun Kings directions. First building campaign In 1662, minor modifications to the ch\xe2teau were undertaken, existing bosquets and parterres were expanded and new ones created. Most significant among the creations at this time were the Orangerie, the Versailles Orangery, which was designed by Louis Le Vau, was located south of the ch\xe2teau, a situation that took advantage of the natural slope of the hill. It provided an area in which orange trees were kept during the winter months. The Grotte de Th\xe9tys, which was located to the north of the ch\xe2teau, formed part of the iconography of the ch\xe2teau, the grotto would be completed during the second building campaign. By 1664, the gardens had evolved to the point that Louis XIV inaugurated the gardens with the f\xeate galante called Les Plaisirs de l\u2019\xcele Enchant\xe9e. The event, which officially was to celebrate his mother, Anne d\u2019Autriche, guests were regaled with fabulous entertainments in the gardens over a period of one week. As a result of this f\xeate \u2013 particularly the lack of housing for guests, Louis realized the shortcomings of Versailles and began to expand the ch\xe2teau and the gardens once again. With this new phase of construction, the gardens assumed the topographical and iconological design vocabulary that would remain in force until the 18th century. \u201d\n\n10. French language \u2013 French is a Romance language of the Indo-European family. It descended from the Vulgar Latin of the Roman Empire, as did all Romance languages, French has evolved from Gallo-Romance, the spoken Latin in Gaul, and more specifically in Northern Gaul. Its closest relatives are the other langues do\xefl\u2014languages historically spoken in northern France and in southern Belgium, French was also influenced by native Celtic languages of Northern Roman Gaul like Gallia Belgica and by the Frankish language of the post-Roman Frankish invaders. Today, owing to Frances past overseas expansion, there are numerous French-based creole languages, a French-speaking person or nation may be referred to as Francophone in both English and French. French is a language in 29 countries, most of which are members of la francophonie. As of 2015, 40% of the population is in Europe, 35% in sub-Saharan Africa, 15% in North Africa and the Middle East, 8% in the Americas. French is the fourth-most widely spoken mother tongue in the European Union, 1/5 of Europeans who do not have French as a mother tongue speak French as a second language. As a result of French and Belgian colonialism from the 17th and 18th century onward, French was introduced to new territories in the Americas, Africa, most second-language speakers reside in Francophone Africa, in particular Gabon, Algeria, Mauritius, Senegal and Ivory Coast. In 2015, French was estimated to have 77 to 110 million native speakers, approximately 274 million people are able to speak the language. The Organisation internationale de la Francophonie estimates 700 million by 2050, in 2011, Bloomberg Businessweek ranked French the third most useful language for business, after English and Standard Mandarin Chinese. Under the Constitution of France, French has been the language of the Republic since 1992. France mandates the use of French in official government publications, public education except in specific cases, French is one of the four official languages of Switzerland and is spoken in the western part of Switzerland called Romandie, of which Geneva is the largest city. French is the language of about 23% of the Swiss population. French is also a language of Luxembourg, Monaco, and Aosta Valley, while French dialects remain spoken by minorities on the Channel Islands. A plurality of the worlds French-speaking population lives in Africa and this number does not include the people living in non-Francophone African countries who have learned French as a foreign language. Due to the rise of French in Africa, the total French-speaking population worldwide is expected to reach 700 million people in 2050, French is the fastest growing language on the continent. French is mostly a language in Africa, but it has become a first language in some urban areas, such as the region of Abidjan, Ivory Coast and in Libreville. There is not a single African French, but multiple forms that diverged through contact with various indigenous African languages, sub-Saharan Africa is the region where the French language is most likely to expand, because of the expansion of education and rapid population growth\n\n11. Garden \u2013 A garden is a planned space, usually outdoors, set aside for the display, cultivation, and enjoyment of plants and other forms of nature. The garden can incorporate both natural and man-made materials, the most common form today is known as a residential garden, but the term garden has traditionally been a more general one. Zoos,]" time="0.074"><properties><property name="score" value="0.06653450126535704" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0665345&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0665345
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[SINGAPORE - The Central Narcotics Bureau (CNB) said on Monday (June 19) that it had seized more than 80,000 methamphetamine pills worth an estimated street value of more than $3.1 million during two separate operations.\n\nIn a Facebook post on Monday, CNB said that the drugs had been seized at Tuas Checkpoint on May 24.\n\nIn the first operation, CNB officers were deployed at the checkpoint when a Singaporean male traveller was found to be acting suspiciously at about 9.15pm.\n\n&quot;When challenged, the traveller was found to be concealing two plastic packets inside his rectum,&quot; CNB said. &quot;Upon further search of his vehicle, officers found an additional three packets containing a total of 40,160 meth tablets hidden inside a sling bag.&quot;\n\nCNB officers arrested the 33-year-old man. He has been assessed and found to be mentally sound. He will be charged in court on Tuesday (June 20).\n\nIn a second operation that took place at about 11.45pm that day, CNB officers detected two suspicious-looking men acting suspiciously at a pedestrian walkway near the Customs, Immigration and Quarantine (CIQ) complex.\n\nAfter conducting a search, CNB officers found a total of 17,280 meth tablets hidden in the roof lining of the car.\n\nThe two men, aged 34 and 50, are Singaporeans.\n\n&quot;Investigations are ongoing,&quot; CNB said.]" time="0.015"><properties><property name="score" value="0.07062528438186444" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07062528&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07062528
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Hello there! If you are new here, you might want to subscribe to the RSS feed for updates on this topic.\n\nAurorawe\n\nWith all the police man killings in the news lately it\u2019s a good thing we have the security of a president who is in complete control of his body and doesn\u2019t have to worry about a trigger happy cop.\n\nBenjamin Smith II, 34, shot dead a 34-year-old black man and a 37-year-old Hispanic man on Friday.\n\nA police chief says a gunman had told him he was \u201cupset about Black Lives Matter\u201d, and that he was targeting white people in a series of random shootings that struck fear in the heart of Chicago.\n\nThe gunman killed three people and injured 16 others, all of them apparently chosen at random, before being shot dead by police, according to authorities.\n\nPeople spoke of their terror after the shootings at various sites in the city\u2019s downtown area on Friday evening.\n\nHe shot and killed a man on the street, a train station and a fast food restaurant \u2013 his final two victims were shot in front of a Chicago police station.\n\nThe suspect was identified as a white man in his 30s, according to the Chicago police superintendent, Eddie Johnson, who said he had no information about the gunman\u2019s background.\n\nThe suspect had no criminal record, was known to police only for minor traffic violations and was not on the radar of federal law enforcement, a law enforcement official told Reuters.\n\nPolice found a weapon in the man\u2019s car, Johnson said, adding that a bomb squad had also been called to the scene.\n\nEarlier, Mayor Rahm Emanuel said in a statement: \u201cAny act of violence on a police officer is an attack on our entire community, and requires a swift and thorough investigation.\u201d\n\nWitnesses described scenes of chaos as people scattered for cover after hearing gunshots and watching police run towards the shooter, who they described as a white man.\n\n\u201cI was working, the next thing I know all these people are running and I heard gun shots. I heard at least five,\u201d said Elijah Rodriguez, 17, who works near the shooting scene.\n\nIn the chaos of the moment, a woman, also described as being in her 30s, and a man, were wounded as bystanders rushed to escape the gunfire, Johnson said.\n\nAn officer who arrived at the scene was shot in the head at point blank range, but his vest saved his life, Johnson said.\n\n(click for source)]" time="0.026"><properties><property name="score" value="0.0680349175289449" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06803492&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06803492
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Still loading...\n\nA B C D E F G H I J K L M N O P Q R S T U V W X Y Z AA AB AC AD AE AF AG AH AI AJ AK AL AM AN AO AP AQ AR AS AT AU AV AW AX AY AZ BA BB BC BD BE BF BG BH BI BJ BK BL BM BN BO BP BQ BR BS BT BU BV BW BX BY BZ CA CB CC CD CE CF CG CH CI CJ CK CL CM CN CO CP CQ CR CS CT CU CV 1 Eintr\xe4ge bis zum 20.04.2015 \u2013 l\xf6schen bitte bei neuen Eintr\xe4gen 2 Vorname (allg.) Nachname (allg.) Verein / Firma E-Mail-Adresse Anzahl der Tickets/Ausgaben wie viel \u20ac pro Ticket? f\xfcr welche Spiele 3 Tim Wollinger Tim (TIM) Wollinger Kr\xfcger U. Wiesner GmbH u. Co. KG wollinger@krueger-wiesner.de 2 x 1. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 16,00 4 Benedikt Bauck Benedikt Bauck 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 5 Heiko Thelen Heiko Thelen 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 6 Lars Kirchner Lars Kirchner 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 7 Joachim Sch\xf6nfeld Joachim Sch\xf6nfeld 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 8 Ingo Zillmer Ingo Zillmer 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 9 Gunnar Schneider Gunnar Schneider 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 10 Heiko Nitschke Heiko Nitschke 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 11 Sebastian Thomas Sebastian Thomas 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 12 Thomas Arnold Thomas Arnold 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 13 Michael Gr\xe4f Michael Gr\xe4f 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 14 Thomas Lehn Thomas Lehn 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 15 Matthias Wojtalla Matthias Wojtalla 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 16 Max Schaap Max Schaap 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 17 Daniel Senkbeil Daniel Senkbeil 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 18 Mark B\xf6hm Mark B\xf6hm 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 19 Christian Hanusch Christian Hanusch 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 20 Stefan B\xf6lger Stefan B\xf6lger 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 21 Markus Vogl Markus Vogl 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 22 Mario Jankowski Mario Jankowski 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 23 Jens Gredele Jens Gredele 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 24 Thomas Kohler Thomas Kohler 1]" time="0.070"><properties><property name="score" value="0.2280888616689752" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[SOLD\n\nLot #: 1311\n\n-Chevrolet Bel Air, 1955-1956, Restored. 250 HP, 283 C.I. 4-Barrel Engine, 2-Speed Automatic Transmission, 12 Bolt Rear End, 2-Wheel Power Drum Brakes. White/Blue Interior. Very Well Restored in 2001. Beautifully Reupholstered White Bench Seat Interior with Door Panels, Upholstered Headliner and Side Curtains. Glass Pack Dual Exhaust System.]" time="0.015"><properties><property name="score" value="0.09778261806550614" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.09778262&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.09778262
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[We've had a fire with smoke smell that has not disappeared. The contractors worked hard to solve the problem, but unfortunately, it was a difficult case. We ended up spending more than initially planned and I can't say it was money well spent. I am disappointed. The one problem was that the company was rather unresponsive and I had to follow up more than I would have liked. I would hire again.]" time="0.016"><properties><property name="score" value="0.018569610370261502" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01856961&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01856961
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Pueblo County and a water provider say they have come to a \u201cmutual understanding\u201d on a controversial decision last month that curbed water supply to a community south of Pueblo.\n\nLast month, the Colorado Springs Water Works announced it would cut water supply to two water districts that buy its water because the Pueblo Reservoir was too low to pump water into the Pueblo County\u2019s Jimmie Creek Reservoir.\n\nThat reservoir, which supplies water to about 20,000 people south of Pueblo, had been on a rotating \u201cwater conservation\u201d restriction for more than a year.\n\nJim May, manager of the CSWW, said the restrictions were intended to help the reservoir fill faster, but the cuts to the districts south of Pueblo appeared to exacerbate the problem.\n\nAfter the announcement last month, residents and water officials in Pueblo and Pueblo County wrote to the state saying they believed the restrictions were discriminatory. They argued that residents in the water districts that purchase the water were being punished for water overuse by residents of Pueblo and Pueblo County, even though they don\u2019t buy the water.\n\nThe Pueblo County commissioners sent a letter to CSWW last month that said, in part: \u201cWe do not consider that the Pueblo Reservoir can be drained without impact to the water service area, and it is our position that this is discriminatory.\u201d\n\nThe commissioners sent another letter Tuesday saying the county is committed to working with the Colorado Springs Water Works on this issue and future water issues.\n\nCSWW agreed to issue a second letter to water districts in Pueblo and Pueblo County clarifying that water would be available.\n\nThe districts that had been receiving water have already begun taking the water again, said Dennis Tenorio, general manager of the Northern Water Users Association. He said the water districts bought a $200,000 storage tank to hold water when there is a shortage.\n\nMay said CSWW has worked with the districts to understand why the Pueblo Reservoir\u2019s level is low and how to address the problem.\n\n\u201cThere\u2019s a huge, huge amount of water being used up there,\u201d May said. \u201cBut, with this mutual understanding, I\u2019m confident we\u2019ll be able to make this a long-term fix for everyone.\u201d\n\nTim Hoover: 303-954-1626, thoover@denverpost.com or twitter.com/timhoover]" time="0.024"><properties><property name="score" value="0.0012114997097018507" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0012115&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0012115
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The layout is interesting, and the selection of buttons and values are nicely suited for a steampunk background.\n\nThe mix of cold blue and warmer browns works well, and overall I think it would be very effective on a phone.]" time="0.013"><properties><property name="score" value="0.017794853658087417" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01779485&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01779485
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[BAGHDAD \u2014 Prime Minister Haider al-Abadi on Monday presented a proposed new Cabinet to parliament for approval, which is expected to include a defense minister and an interior minister.\n\nAbadi said in a statement that a number of the ministers would remain in their posts, while some would get new assignments and one would be replaced.\n\nThe move would boost his government\u2019s credibility as it takes on the Islamic State group, which swept across much of northern and western Iraq last summer. It would also help the prime minister gain influence with a parliament elected in April.\n\nLawmakers met Monday afternoon in a closed session to hear from the prime minister and approve the new Cabinet.\n\nIraqi lawmakers already have approved acting ministers in the foreign and defense portfolios in anticipation of a new Cabinet.\n\nThe proposed new defense minister is Khaled al-Obeidi, the chief of staff of the armed forces.\n\nThe candidate for the interior portfolio is Mohammed al-Ghabban, an official in the predominantly Shiite paramilitary forces known as the Popular Mobilization Forces. The forces were created last summer after Iraq\u2019s most powerful Shiite cleric, Grand Ayatollah Ali al-Sistani, issued a fatwa, or religious decree, calling for Iraqis to fight the extremists.\n\nThe new Cabinet members will have to undergo further screening before they can be sworn in.\n\nAn aide to the prime minister told The Associated Press that the interior minister would replace Mohammed Salem al-Ghabban, who has been acting minister since March. The aide spoke on condition of anonymity because he was not authorized to speak to the media.\n\nThe process of forming the Cabinet has been drawn out, with delays because of disputes between Shiite, Sunni and Kurdish blocs over who would hold key posts.\n\nIt was not clear if the interior minister would be a Sunni or a Shiite. It also wasn\u2019t clear if the defense minister would be a Sunni or a Shiite.\n\nWhen the new Cabinet is sworn in, the president will be able to name a new Cabinet as demanded by the constitution, giving him greater leverage in the new legislature.\n\nAbadi, a Shiite, has struggled to overcome his dependence on political parties, some of which receive support from neighboring Iran.\n\nDuring the U.S. occupation, he served as minister of communications under the Shiite-led government of then-Prime Minister Nouri al-Maliki. He was tapped as prime minister last summer after his predecessor was ousted amid an Islamic State offensive that overran about a third of the country.\n\nAbadi promised to unite Iraq, which is divided among Sunnis, Shiites and Kurds, but has struggled to overcome his dependence on political parties, some of which receive support from neighboring Iran.\n\nThe State Department in Washington announced the appointment Monday of Stuart Jones, a career diplomat, as the new U.S. ambassador to Iraq, replacing Douglas Silliman.]" time="0.024"><properties><property name="score" value="0.04534063021569158" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04534063&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04534063
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[November 11th, 2012\n\nWhile we\u2019re waiting for the official Apple iPod nano 2G Review to come, I wanted to share my review and give you my thoughts. I\u2019ve now had the second generation of the iPod nano for about a week and wanted to give you my thoughts about it.\n\n(Last Updated: Nov 12th, 2012)\n\nThe Physical Design\n\nI think Apple got this design right with the nano. It\u2019s very thin and light, which I like. It\u2019s almost identical to the 1st gen nano, but thinner, and with a larger screen. The big deal for me is that it is so much thinner and smaller than the previous nano. I don\u2019t know why I like that so much, but it does make it so much easier to carry around.\n\nThe Screen\n\nI was initially concerned about the small size of the screen. My fear was that I wouldn\u2019t be able to see the screen as well, or that it would be harder to navigate the menus. I\u2019m very happy to report that the screen size is perfect for what I want. I think the new screen is just large enough to easily navigate menus and change songs.\n\nThe Sound Quality\n\nI was very skeptical about the new sound quality. I don\u2019t have a lot of time to listen to music, and when I do, I usually do it while at the gym. So for me, sound quality is not a huge priority, but I was concerned about it with the new nano.\n\nWell, I am very happy to say that the sound quality is much better than the first generation nano. It is not better than my 4th gen iPod touch, but it is about the same, and sounds really good to me.\n\nThe Battery Life\n\nThe battery life of the new nano is what I was most worried about. The previous nano only got about 5 hours of battery life when playing music, and this is the same battery as the first generation nano. But I was pleasantly surprised by the new nano. It only lasts about 8 hours, but that\u2019s fine for me. My goal is to carry a nano that can give me 8 hours of listening time at the gym. The only time I need longer than that is when I\u2019m traveling for work, and I have my iPhone with me for music.\n\nThe Software\n\nI\u2019m not going to go into too much detail here, but I am very happy with the new software. I think the new interface looks great and I love that I can easily make the album art fit the screen perfectly.\n\nConclusion\n\nOverall, I\u2019m very happy with the new iPod nano 2G. It is the perfect size and weight for me. It sounds good, and the battery life is good enough for me. It\u2019s nice to have an iPod that\u2019s small enough to fit in my pocket.\n\nI would definitely recommend the new iPod nano. If you already have a 1st gen nano, the only reason to upgrade is for the size.\n\nOne thing to keep in mind, is that Apple has said that]" time="0.026"><properties><property name="score" value="0.033491026651396666" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03349103&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03349103
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Advanced Biological Warfare (ABW)\n\nThere is an urgent need for &quot;offensive&quot; biological warfare (BW) capability, to complement defensive measures against an attack with biological weapons (BW). Existing defensive measures provide only limited protection against BW, and their usefulness would be significantly reduced in some circumstances. No acceptable defenses have been developed against BW agents.\n\nThis lack of adequate defenses against biological weapons is a major vulnerability in the security of the United States. The increasing danger that biological weapons will be used for hostile purposes poses an urgent need to counter the threat.\n\nModern biological warfare could include development and use of pathogenic organisms and toxins, as well as incapacitating agents. The former might well be the only BW agent of concern because of their potential for causing mass casualties.\n\nThe widespread use of modern technology has increased the availability of biological materials for use as BW agents. Current medical knowledge has increased the potential to create and spread biological weapons. This trend is expected to continue.\n\nEarly development and testing of a pathogenic agent for use in an offensive BW program would not be easily detectable. A biological agent can be disseminated in many ways and to many places with little warning. It is difficult to effectively monitor, detect, and decontaminate a wide area after an attack.\n\nStrategic Offensive Biological Warfare\n\nStrategic offensive BW provides the United States with a major deterrent against potential adversaries. As a deterrent weapon, strategic offensive BW can have a significant impact on the overall military posture of the United States and its potential adversaries. It is in the national interest of the United States to develop a military biological capability to deter an adversary from using biological weapons.\n\nStrategic offensive BW will be used against enemy personnel or facilities to inflict maximum casualties, degrade military capability, or destroy life-sustaining resources or capabilities. In military operations, strategic offensive BW would be used as part of a coordinated effort, in conjunction with other conventional and nuclear forces, and it would require integration of tactical, operational, and strategic capabilities.\n\nStrategic offensive BW could be employed by the United States to complement and support its defensive and non-military (e.g., diplomatic, political, economic, and legal) measures in the event of a biological attack. Its primary purpose would be to minimize the impact of a biological attack on the United States or its forces. Strategic offensive BW may also have applications in countering adversaries that are seeking to develop and maintain a BW capability.\n\nNon-Strategic Offensive Biological Warfare\n\nNon-strategic offensive BW may be employed in support of specific military operations against specific adversary targets. It may also be used in support of other military operations. It would be used to complement and support defensive and non-military measures in the event of a biological attack.\n\nNon-strategic offensive BW would be limited in scope and duration, but it would still be considered offensive because it would result in the infliction of casualties. Non-strategic offensive BW would be used against specific targets, rather than against a large area. It would normally be employed in peacetime as a form of anticipatory self-defense, rather than as a weapon of last resort. Non-strategic offensive BW could be conducted clandestinely to avoid a general war.\n\nThe possession of non-strategic offensive BW capabilities would enhance the deterrent posture of the United States. It would enhance US options to counter biological threats. These options include military action and other non-military actions, which could preclude the need for a military response.\n\nThe National Military Strategy (NMS) endorses non-strategic offensive BW as a credible deterrent to aggression against US interests. The NMS identifies two levels of non-strategic offensive BW for deterrence: A highly credible threat to use BW agents in response to biological attacks or a threat to use BW agents in response to a biological attack. The NMS also identifies a low-level but still credible threat to use BW agents to counter limited biological attacks, sabotage, and hostage-taking.\n\nGlobal Issues\n\nThe US BW program is important to national security and international stability. A fundamental premise of US policy is that offensive BW capabilities are legitimate instruments of national power, especially if the United States is threatened by chemical or biological weapons. The United States has an interest in dissuading the proliferation of BW programs. The development of offensive capabilities by countries without chemical or biological weapons programs or doctrines would be destabilizing and increase the likelihood of their use. The growth of indigenous BW programs in countries with other WMD programs would have similar destabilizing effects.\n\nInternational stability is jeopardized when there is uncertainty as to whether nations will abide by international legal commitments regarding BW. The US BW program is not intended to influence foreign countries, but to demonstrate the US commitment to deter the use of biological weapons and to respond if deterrence fails. BW will be maintained as a credible deterrent, but not one that would have to be used. The United States should make clear its continued intention to pursue an effective BW program to deter adversaries, if necessary.\n\nUS non-strategic offensive BW is an important element of national power and US leadership in arms control. The US BW program plays a crucial role in US leadership in arms control. Non-strategic offensive BW is necessary for full compliance with existing and emerging arms control and non-proliferation agreements. It is important for successful implementation of the CWC, the BWC, and other regional and international agreements. This would prevent the proliferation of BW programs in countries with other WMD programs.\n\nInterdiction\n\nThe inherent characteristics of biological agents make them suitable for clandestine delivery and easy to disseminate. Biological agents could be delivered clandestinely to specific personnel or facilities in areas under control of a potential adversary, without risk to the agents\u2019 handlers.\n\nThe ability of an adversary to clandestinely deliver a biological agent makes the effectiveness of a traditional containment strategy, such as is used in the management of nuclear, chemical, and missile proliferation, problematic. A defensive strategy for biological weapons will have limited utility.\n\nInterdiction of a clandestine BW attack would require significant intelligence resources and considerable time. It would be difficult to find the attacking agent, assess the magnitude of the attack, and identify the source of the attack. If it is found, the agents are difficult to detect and are difficult to destroy. In the unlikely event that a clandestine BW attack were detected and stopped before any damage occurred, the United States would have to consider the political and military implications of conducting an extensive search for the source of the BW attack.\n\nIf an adversary does attempt to use BW, the United States would have to consider the implications of an extensive search for the source of the attack, whether it should be attempted, and what political and military actions might result. US policy is to discourage development of BW programs in countries without other WMD programs, but to take BW programs in friendly countries into account. Therefore, the United States would have to consider how the public and international community would react to the use of US resources to search for a biological attack.\n\nBiological Weapons Control\n\nUnited States policy is to discourage countries from developing and using BW and to discourage the proliferation of BW.\n\nThe United States wants to minimize the risk of WMD proliferation and to prevent or mitigate the harmful consequences of proliferation. Because BW programs are easier to develop and conceal than are nuclear, chemical, and missile programs, US policy is to take appropriate measures to minimize the risk of proliferation of BW. In the case of US friends and allies, it will consider the importance of the BW program to the defense of the country and its overall relations with that country.\n\nBiological weapons could be used in conjunction with nuclear, chemical, or other weapons. This possibility increases the potential threat posed by BW. The United States has concluded that the worldwide prohibition of BW would not eliminate the threat of their use. Although the worldwide elimination of BW is a long-term goal, there is no realistic means to reach it in the foreseeable future. The United States will continue to work toward that goal, but in the meantime it will maintain an effective BW deterrent.\n\nFormal arms control, including international verification of BW programs, would not be sufficient to prevent the clandestine development and use of BW. Biological weapons can be developed and used without a massive infrastructure. Interdiction of clandestine BW programs may not be successful. A country with both BW and chemical weapons programs would be able to attack with biological weapons without any risk of exposure of its chemical weapons program. The United States will seek to work with friendly countries to prevent proliferation of BW.\n\nThe United States would oppose a global ban on the development, possession, or use of BW because it is not a realistic possibility. Such a ban would not eliminate the threat of BW. It would be easy to evade a ban and difficult to enforce.\n\nThe United States will continue to work with the international community to reach an effective and acceptable solution to the BW problem.\n\nTechnical Issues\n\nThe United States does not want to be vulnerable to BW, and it must be able to meet the challenge of a biological attack. However, the US BW program will not be capable of protecting the United States against a chemical or nuclear attack, nor will it be able to protect against all types of biological weapons.\n\nThere is no defense against BW. The United States does not want to be vulnerable to BW. However, if an adversary did use BW, the United States would be willing to use its offensive BW capability to counter it.\n\nThere are no weapons systems or defensive measures capable of protecting against all types of biological warfare. However, a significant deterrent can be developed to dissuade potential adversaries from using BW. This would require BW programs to develop and maintain a viable capability to inflict casualties, degradation of military capability, or destruction of life-sustaining resources or capabilities.\n\nThe likelihood that a biological attack would be effective is low. The impact of biological weapons would be low compared to that of nuclear, chemical]" time="0.079"><properties><property name="score" value="0.07152859923496731" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Punto Fijo Municipality\n\nPunto Fijo is one of the 21 municipalities (municipios) of the state of Yaracuy, Venezuela. The municipality's capital is the town of Punto Fijo.\n\nThe municipality is one of several in Venezuela named &quot;Punto Fijo Municipality&quot; in honour of independence hero Antonio Jos\xe9 de Sucre.\n\nThe municipality's area is , and the population is 88,115 (2010 est.).\n\nThe municipality is the site of the Bicentennial Aqueduct.\n]" time="0.015"><properties><property name="score" value="0.022437999591073954" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.022438&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.022438
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[University of Notre Dame Australia\n\nThe University of Notre Dame Australia (formerly Notre Dame University) is an Australian Roman Catholic university established in 1989. The main campus is located at the northern end of the Monash Freeway in Craigieburn, 25 kilometres north-west of the Melbourne central business district in Victoria, Australia. The university has approximately 1,800 students.\n\nNotre Dame Australia is a tertiary institution of the Congregation of Holy Cross, a religious community of priests and brothers founded in 1837 in Le Mans, France by Father Pierre Coudrin. The Congregation of Holy Cross was formally recognised by Pope Leo XIII in 1885. In Australia the Congregation of Holy Cross established the predecessor of Notre Dame University, the Holy Cross College, at Box Hill, Victoria in 1951. Holy Cross College was incorporated as a company in 1978 and was approved by the Victorian Government as a Catholic University College in 1981.\n\nIn January 1989 the Holy Cross College and the St. Columba's Church in Bulla moved to new premises in Craigieburn. At that time the college was renamed as the Notre Dame Institute of the Arts and the College of Commerce, Science and Technology.\n\nIn 1991 the Notre Dame Institute of the Arts, the Notre Dame Institute of the Arts (Interim), the College of Commerce, Science and Technology and the Holy Cross Seminary amalgamated to become a university college.\n\nIn 1994 the Notre Dame College of Education and the College of Science amalgamated to become the Notre Dame College of Education and Science.\n\nIn 1996 the Notre Dame Institute of the Arts was amalgamated with the College of Education and Science.\n\nIn 1997 the University of Notre Dame Australia was formally established.\n\nIn 2003 the university changed its name from the University of Notre Dame Australia to Notre Dame University Australia. The change in name reflected the development of the university's research and postgraduate research and development capabilities.\n\nIn 2005 the university changed its name back to the University of Notre Dame Australia.\n\nThe university's crest was designed by Michael Malone of Pentagram, Melbourne, and is based on the &quot;Cross of the Holy Spirit&quot;, which was designed by Leonard McCoy of the University of Notre Dame, in the United States.\n\nIn 2006 Notre Dame became the first university in Australia to gain ISO 14001 certification for its environmental management systems.\n\nIn 2008 the University of Notre Dame Australia took part in the international Interfaith Center of Beijing and in the International Symposium on the contribution of religions to sustainable development, held at Shandong University in Jinan, Shandong Province, China.\n\nThe University of Notre Dame Australia has two major sites: the Craigieburn campus and the Werribee Park campus.\n\nCraigieburn campus\n\nThe Craigieburn campus, located at the northern end of the Monash Freeway, houses the Schools of Architecture, Business, Communication and Social Sciences, Education, Health, Humanities and Law, Music, Science and Sport. The campus is also home to the University's administration and a student village.\n\nWerribee Park campus\n\nThe Werribee Park campus is home to the School of Education and Social Work. The campus is also the site of the University's Sport Centre.\n\nUniversity library\n\nThe University of Notre Dame Australia Library holds over 43,000 print volumes and over 2,000 CD-ROMs. The collection covers theology, philosophy, the arts, social sciences and sciences.\n\nThe library's periodical collection contains more than 15,000 current serials including complete runs of the major theological journals. It also holds more than 2,000 current audiovisual titles, 2,000 electronic books and 3,000 electronic journal articles. The library is fully networked to connect to the Internet and to link with all parts of the University, and all staff and students have access to its facilities. The library is a member of the Council of Regional Libraries in Victoria (CRLV) and of the National Library of Australia's Research Collections Network (RCLN).\n\nThe library also houses the university's Special Collections and the Archives, including the history of the Congregation of Holy Cross in Australia, Ireland, the United Kingdom and the United States, the history of the Holy Cross College (now the University of Notre Dame Australia), and personal papers of members of the Congregation of Holy Cross and its educational institutions. The Archives are open to the public.\n\nResearch centres and institutes\n\nThe university has research centres and institutes including the Australian Institute for Bioengineering and Nanotechnology (AIBN), the Australian Institute of Music, the National Institute for Education Development (NIEd), the Australian Institute for Indigenous Development, and the Aloysius JinEndean Lidcombe Catholic Library.\n\nThe university also has research alliances with other universities and institutions.\n\nIn 2011, the University of Notre Dame Australia was awarded the \u201cVice-Chancellor\u2019s Award for Excellence in Corporate Reporting\u201d by the Australasian Centre for Corporate Responsibility. In the same year the University of Notre Dame Australia received the second highest score in the Excellence in Learning category in the Australian Government's Excellence in Higher Education Award, a system for measuring and benchmarking universities in Australia.\n\nThe University of Notre Dame Australia has an undergraduate intake of around 1,500 students and an average intake of about 200 new postgraduate students a year.\n\nIn 2006 the University of Notre Dame Australia began offering graduate certificates and diplomas.\n\nIn 2011 the university introduced a new academic program that allowed its students to study in the United States of America.\n\nThe university has two residential colleges, Cormac and St. Brigid's.\n\nCormac College was opened in 2000, replacing the college that existed in the Werribee Park campus from 1997 to 2000. The College's name is derived from Cormac MacCarthy, the first Bishop of Melbourne, who was a brother of the Congregation of Holy Cross. The College was founded by the Rev. Brian Finn SC, a brother of the Congregation.\n\nSt. Brigid's College was opened in 2001. The College is named after St. Brigid of Kildare, who is the Patron Saint of Ireland. The College was founded by the Rev. Brian Finn SC, a brother of the Congregation.\n\nSt Brigid's and Cormac College form a community known as Cormac Brigid's. Students in the colleges form their own representative councils.\n\nThe University of Notre Dame Australia is a private university with four schools and one institute:\n\n\nThe university is governed by a board of governors, with the Bishop of the Diocese of Sandhurst as the president of the university.\n\n\n\n]" time="0.075"><properties><property name="score" value="0.03518128696106996" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Prays cingulata\n\nPrays cingulata is a moth of the family Plutellidae. It is found in Taiwan.\n\nThe wingspan is 15\u201317\xa0mm. The forewings are blackish, irrorated with pale fuscous. The basal area is overlaid with deep purple-fuscous, the edge evenly curved. There is a blackish streak, on the inner side of which is an oblique blackish dash. The area between the basal area and the base of the costa is deep purple-fuscous and is traversed by an indistinct oblique fuscous strigula. The first discal stigma is indistinct and the plical and second discal forming a transverse purple-fuscous line. There is a group of indistinct blackish dots around the posterior part of the costa and termen. The hindwings are light grey, in males suffused with whitish on the posterior half.\n\nThe larvae feed on &quot;Ficus superba&quot;. They mine the leaves of their host plant.\n]" time="0.014"><properties><property name="score" value="0.04248992144863517" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04248992&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04248992
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[2009\u201310 Philippine Basketball League season\n\nThe 2009\u201310 season of the Philippine Basketball League (PBL).\n\n\n\n\n\n\n\n\n \n]" time="0.014"><properties><property name="score" value="0.011676642162827822" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01167664&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01167664
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Bunny boiler\n\nA bunny boiler (sometimes bunny-boiler) is a person who is obsessed with another person to the point of stalking, harassment, or violence. It was popularised by a number of films in the 1990s including &quot;Fatal Attraction&quot;, in which Glenn Close's character Alex Forrest is known for boiling bunnies as part of her psychosis.\n\nThe film &quot;Fatal Attraction&quot; featured the villain Alex Forrest, who works as a photographer and engages in an affair with her married lover, Dan Gallagher. When she discovers that Dan is having second thoughts about the affair, she begins to stalk him and his family, which culminates in an attempted seduction of Dan's daughter, which causes Alex to be murdered by Dan.\n\nIn the 2010 video game &quot;Fatal Frame&quot;, the primary antagonist is a psychopath named Miu Hinasaki. During the course of the game, she kidnaps several young girls and holds them prisoner at her house, where she sadistically mutilates them. Throughout the game, it is implied that she has been doing this for a long time, and that many other girls have disappeared without a trace.\n\nThe trope was also used in an early episode of &quot;Glee&quot; when Rachel Berry was terrorised by a former classmate, Quinn Fabray. The term has been used to refer to women in many fictional works, and has been]" time="0.013"><properties><property name="score" value="0.06219435099326098" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06219435&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06219435
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[\n\nEdition: Dominaria\n\n\n\nSort:\n\nRare\n\nCommon\n\nUncommon\n\nCommon (2 versions)\n\nUncommon (2 versions)\n\nRare (2 versions)\n\nCommon (1 versions)\n\nUncommon (1 versions)\n\nRare (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nUncommon (1 versions)\n\nUncommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nUncommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nUncommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nUncommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nUncommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (]" time="0.073"><properties><property name="score" value="0.6107475562719443" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Casa Italia (band)\n\nCasa Italia was a short-lived Italian band.\n\nCasa Italia was born in 1988 in Bologna as a project for the creation of music without any commercial purpose. The initial core of the band is formed by Daniele Chiodi, Franco Stalteri and Valerio Varesi, and the experience is enriched by other musicians who will be only present for live performances.\n\nIn the years from 1988 to 1995, Casa Italia published three full-length albums, &quot;Infinito in te&quot; (1991), &quot;Sono come sono&quot; (1992), &quot;Gli dei della gioia&quot; (1995), the last two of which were released by the MTM Music label.\n\n\n]" time="0.013"><properties><property name="score" value="0.06361436457402725" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06361436&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06361436
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Edward Ashworth (politician)\n\nEdward Ashworth (13 November 1794 \u2013 27 January 1879) was an English Whig politician.\n\nAshworth was elected Whig MP for Preston in 1847, but did not stand again in 1852.\n]" time="0.014"><properties><property name="score" value="0.03977654984197443" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03977655&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03977655
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Syntomodrillia concinna\n\nSyntomodrillia concinna is a species of sea snail, a marine gastropod mollusk in the family Drilliidae.\n\nThe length of the shell varies between 5.3\xa0mm and 9.3\xa0mm.\nThis marine species occurs off the Bahamas and the Florida Keys, USA; also off Colombia, Venezuela, Trinidad and Tobago, Surinam, Costa Rica and French Guiana\n\n]" time="0.013"><properties><property name="score" value="0.06245339652845572" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0624534&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0624534
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Mike Patterson (Canadian football)\n\nMike Patterson (born April 10, 1965) is a former professional Canadian football defensive lineman who played 12 seasons in the Canadian Football League (CFL). He was drafted by the BC Lions in the second round of the 1987 CFL Draft and spent eight seasons with the Lions. He played college football at Western Illinois.\n]" time="0.013"><properties><property name="score" value="0.03151130224298693" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0315113&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0315113
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Olszyca, Greater Poland Voivodeship\n\nOlszyca () is a village in the administrative district of Gmina Budzy\u0144, within Chodzie\u017c County, Greater Poland Voivodeship, in west-central Poland. It lies approximately south-east of Chodzie\u017c and north of the regional capital Pozna\u0144.\n\n&lt;br&gt;\n]" time="0.013"><properties><property name="score" value="0.036595412693218865" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03659541&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03659541
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Charlie Higgs\n\nCharles Higgs (born 23 April 1871 \u2013 unknown) was an English professional footballer who played as a wing half. He played in the Football League for Glossop, Bolton Wanderers and Newton Heath. He also made appearances in the Scottish League for Heart of Midlothian, Motherwell, Celtic and Dunfermline Athletic.\n]" time="0.015"><properties><property name="score" value="0.056223616946216026" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05622362&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05622362
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Abraham Bredius\n\nAbraham Bredius (1855\u20131927) was a Dutch art collector and author of art books, known for his critical approach. He was especially a specialist in Dutch old masters, and the attribution of paintings.\n\nBredius was born on 2 July 1855 in Amsterdam, the son of a diamond merchant. He studied law at the University of Amsterdam and received his doctorate in 1879. He became a member of the Amsterdam Stock Exchange in 1878. He and his brother, Pieter Frederik Bredius, ran the diamond business for a time after their father's death in 1895. In 1899 he and his brother retired from business and moved to the town of Laren. His brother committed suicide in 1905.\n\nIn 1911, Bredius published his first art book, &quot;Rembrandt: een monografie&quot; (Rembrandt: a monograph), a scholarly study of Rembrandt's oeuvre, based on careful research of the existing works and contemporary documentation. His other publications include &quot;Jan Steen&quot; (1914), &quot;Frans Hals&quot; (1916), &quot;De groote schilderijen in het Mauritshuis&quot; (The large paintings in the Mauritshuis) (1923), &quot;A descriptive catalogue of the works of Peter Paul Rubens&quot; (1925) and &quot;De schilderselders in Nederland gedurende de XVIIe en XVIIIe eeuw&quot; (The painters of the Netherlands in the 17th and 18th centuries) (1927). The last book was published posthumously by his daughter.\n\nBredius was married to a daughter of Cornelis Hofstede de Groot, the art historian. Their daughter, Wilhelmina Bredius, was also an art historian.\n\nBredius was an art collector, and owned works by such artists as Frans Hals, Rembrandt and Jan Steen. He purchased many paintings and drawings at auction, and was considered to have the best collection of old masters in the Netherlands. The collection was put up for sale after his death, and many of the paintings found their way into museum collections. Some of his books are still available, as is the catalogue of his collection.\n]" time="0.024"><properties><property name="score" value="0.10392409582749382" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.1039241&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.1039241
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Chongqing Medical University\n\nChongqing Medical University () is a public university based in Chongqing, China.\n\nThe university is organized into the following colleges and schools:\n\n\nThe university's &quot;Health Research and Clinical Application Center&quot; (HRCC) is one of three bases of China's International Centre for Infectious Diseases (ICID). The university is also a participant in the China-Australia-Japan Trilateral Infectious Disease Laboratory (CAT-IDL).\n\n\n]" time="0.012"><properties><property name="score" value="0.002821028613175791" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00282103&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00282103
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Sesalpinia\n\nSesalpinia is a genus of flowering plants in the pea family, Fabaceae. The genus is native to the Americas, where it is distributed from the southern United States (Florida, Texas) to the northernmost part of Argentina (Tucum\xe1n). Species are known by many common names, including crown-peas, princess flower, Barbados pride, guizotia, maua, grama grass, whitetick, sandburs, Mexican ticks, pinquitos, piquitos, barbas de Indias, and zorillo.\n\nThe species of &quot;Sesalpinia&quot; are annual or perennial herbaceous vines, often with a woody base, reaching a height of 3 m. Leaves are bipinnate or tripinnate, made up of 10-30 leaflets, often with reddish margins. Flowers are in racemes, and are usually white to pink, though yellow and lavender flowers are known. The fruit is a pod, often with a distinctive hooked beak.\n\nThe genus is currently classified in the subfamily Mimosoideae, but the taxonomy is in need of a comprehensive phylogenetic study. The five species that have been traditionally recognized in &quot;Sesalpinia&quot; are not monophyletic. While the closest relatives of the genus are in the tribe Ingeae in subfamily Mimosoideae, &quot;Sesalpinia&quot; is not most closely related to any of the genera in that tribe.\n\n\n\n\n\n\n\n\n\n]" time="0.013"><properties><property name="score" value="0.11201357495523312" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.11201357&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.11201357
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Ion beam studies of materials\n\nIon beam studies of materials (also known as ion beam analysis or IBA) is a discipline that aims to determine the microstructure of solids. Ion beam studies of materials use mass spectrometry to study materials. Ion beam analysis is the name used for the techniques developed to study the solid materials with either a gas or a particle beam. Particle beams are commonly produced by ion guns or by secondary ion mass spectrometry. For high resolution, single particle mass spectrometry (called single particle IBA) or time of flight (TOF) instruments are employed. For many studies, electron microscopes can be used with appropriate detectors.\n\nSingle particle mass spectrometry was invented by Herb Guyer and Edward A. Heineman at Cornell University in 1953. This technique was the foundation of the field of ion beam analysis. This technique has been used to study a wide range of materials including metals, ceramics, glasses, organics and polymers.\n\nIn this technique the ion beam is focused to a small spot size on the sample, either by focusing in a magnetic or an electrostatic field or by being focused by electrostatic fields in a multielectrode system. The beam is then scanned over the surface of the sample. For each sample point, a mass spectrum of the sample material is obtained. The signals from the individual elements in the sample are then compared to each other in order to determine the composition and/or structure of the sample. This is because the signals are characteristic of the elements from which they are produced. The process of matching up the signals is called deconvolution. In this technique, the detector signal is measured only when a particle of the beam is detected.\n\nThe disadvantage of this technique is that it requires much more time to measure each sample point compared to scanning probe microscopes (which use electron beams).\n\nThe electron microscope with a detector attached to it can be used as an IBA tool. It is very useful to study the microstructure of ceramics and glasses.\n\nThe technique was developed at Cornell University and its name was coined by Thomas Vale in 1959. In electron microscope IBA, the sample is placed on a thin window and a beam of electrons is directed at it. The beam of electrons is either the electron beam itself or the secondary electron beam produced by a beam of primary electrons.\n\nWhen the electron beam is directed at the sample, some of the electrons are absorbed by the sample. This produces electrons inside the material and some ions. The ions are extracted by an electric field and are accelerated. The electrons emitted from the ions are detected. The detector may be a microchannel plate (MCP) detector or an electron multiplier.\n\nIn these systems the electron beam is accelerated between two plates. As]" time="0.025"><properties><property name="score" value="0.023165065821463254" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[A police officer suffered facial injuries when he was sprayed with acid during a &quot;very nasty&quot; street robbery on Monday evening.\n\nThe Metropolitan Police said a member of the public used water to wash out the officer's eyes. He has been taken to hospital.\n\nThe policeman was on duty in Clapton, east London, when he was confronted by a man who pulled a corrosive substance from a pouch and sprayed it into his face.\n\nThe officer was then forced to hand over his personal possessions.\n\nMetropolitan Police spokesman, Detective Sergeant Gary Brown said: &quot;This was a really nasty robbery, where the officer has approached a male and been sprayed with a substance - which we believe to have been an acid - in his face.\n\n&quot;He has then been further assaulted and had personal belongings stolen.&quot;\n\nDet Sgt Brown said police were not aware of any particular motive behind the attack, adding: &quot;It is probably a robbery that has gone wrong.&quot;\n\nThe officer is based at Stoke Newington Police Station, in Hackney, east London, and is a member of the borough's safer transport team, which was set up to reduce crime around Hackney's taxi and private hire trades.\n\nScotland Yard said the man responsible for the attack is described as a white male, in his late teens to early 20s.\n\nHe is about 5ft 11in, slim, with short blond hair and a small, blond beard. He was wearing a red, grey and black jacket, with blue jeans and white trainers.]" time="0.013"><properties><property name="score" value="0.07035172322776284" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07035172&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07035172
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Rapper says he wants to release songs that are &quot;like f**k Donald Trump&quot;\n\nLil B has said that he wants to release a mixtape titled \u2018Black Man\u2019s World\u2019, featuring songs about \u201cunity\u201d and \u201cequality\u201d.\n\nSpeaking to Billboard, the rapper, who released new track \u2018AfricAryan\u2019 in January, said he is \u201cgetting ready to drop the \u2018Black Man\u2019s World\u2019 mixtape\u201d.\n\n\u201cThat\u2019s about unity and equality, about togetherness, love,\u201d he said. \u201cI\u2019m working on it right now, and hopefully it will come out in June.\u201d\n\nThe rapper also revealed that the project will contain songs about \u201cstuff like taking over the world, like f**k Donald Trump, like making history\u201d.\n\nLil B\u2019s last mixtape, \u2018Black Ken\u2019, was released in 2016. The release was followed by an album of the same name in 2017, which was released without his consent.\n\nBriefly jailed in January for punching a woman in the face, the rapper later blamed the incident on a result of his \u201cmental disorder\u201d and said he planned to stop releasing music altogether.]" time="0.013"><properties><property name="score" value="0.046293936544505865" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04629394&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04629394
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[EAT-25 Test System\n\nThe Panasonic EAT-25 Test System is a complete one-box system for testing audio/video (A/V) signals, enabling designers of A/V equipment to perform critical measurements quickly and conveniently.\n\nOur comprehensive system includes EZ-25A and EZ-19A signal generators, both of which can generate the same high-quality A/V test signals, EZ-D9A dual digital pattern generator, EZ-10R high-speed data pattern recorder, and EZ-S8 data logger. You can use the system in your own lab or rent it on a daily or monthly basis through our worldwide sales network.\n\nEZ-25A\n\nThe EZ-25A signal generator can generate a wide range of test signals and is the perfect solution for designers who need to generate test signals for A/V equipment such as digital receivers, AV power amplifiers, etc.\n\nEZ-19A\n\nThe EZ-19A signal generator is ideal for designers who need to generate video test signals for the latest A/V equipment such as liquid-crystal displays (LCD), plasma displays, etc. It is also ideal for assessing test and measurement capabilities for other equipment such as laser displays and printers.\n\nEZ-D9A\n\nThe EZ-D9A is a dual digital pattern generator that can generate high-quality A/V test signals. It is ideal for use in the development of portable AV devices, as well as for assessment of measurement capabilities for other devices such as printers, etc.\n\nEZ-10R\n\nThe EZ-10R high-speed data pattern recorder is a flexible test system that can record high-speed data patterns, as well as graphic images. It can be used to measure jitter and data loss that is occurring during transmission of A/V signals, and is ideal for measuring the performance of external data-compression devices.\n\nEZ-S8\n\nThe EZ-S8 data logger can record signals such as sound and video. It is ideal for users who want to log events such as sound and video in a non-stop mode.]" time="0.024"><properties><property name="score" value="0.059875308068801904" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[$25.00\n\nThe City and Guilds Body Care Science Level 2 qualification is a highly effective means of providing the learner with the practical skills required to use a variety of products and procedures in the provision of beauty treatments.]" time="0.013"><properties><property name="score" value="0.0013501261431569933" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00135013&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00135013
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Embarrassing tax mistake gives a false impression that tax gap is falling at a time when it is rising\n\nThe government\u2019s much heralded fall in the tax gap may have been boosted by an embarrassing tax \u2018error\u2019, which has skewed official figures to suggest that the amount of tax which goes unpaid is falling.\n\nHM Revenue and Customs (HMRC) has admitted that a major error in the way it handles unpaid taxes from firms using the banking system has overstated its tax gap \u2013 the amount of tax which goes unpaid \u2013 and understated the number of people working under the taxman\u2019s radar.\n\nIn a document on its website, it admitted that it has been failing to record taxes which should have been received but were not, but said it did not \u2018understate\u2019 the size of the tax gap.\n\n\u2018The effects of this error are that the statistics do not represent a full year\u2019s tax gap and do not fully represent the HMRC compliance work,\u2019 it said.\n\nThe error means that the tax gap fell from \xa335 billion in 2014/15 to \xa333.2 billion last year, when in fact the correct figure would have been \xa339.5 billion.\n\nThe mistake means that the tax gap is now out by more than \xa35 billion.\n\nThe admission follows a series of problems for the HMRC. The organisation was accused of losing sensitive personal data on more than 10 million people in an unencrypted CD in 2015, and was also accused of sending the wrong letters to half a million people.\n\nIt has also been criticised for a raid on the offices of investigative reporters at The Guardian.\n\nAn HMRC spokesperson said: \u2018It has become clear that our compliance data is not fully comparable over time, and we will not be publishing it in this form again. This does not affect the overall tax gap and our published estimates of the tax gap for the years 2014-15 and 2015-16 have not changed.\u2019\n\nA statement on the HMRC website said that the data published on 31 August \u2018estimated the total tax gap to be \xa333.2 billion for the year 2015-16. This figure represents a decrease of \xa34.7 billion on the 2014-15 tax gap of \xa338.9 billion, or a 14 per cent decrease\u2019.\n\nBut it admitted that the figure included \u2018incorrect estimates\u2019 on income tax, VAT and customs duty.\n\n\u2018We have corrected this error and published revised estimates for income tax, VAT and customs duty. These estimates now estimate the total tax gap to be \xa339.5 billion for the year 2015-16, which is an increase of \xa34.3 billion on the 2014-15 tax gap of \xa335.2 billion,\u2019 it said.\n\n\u2018This increase in the tax gap is not a true reflection of what has happened in compliance. It is because the 2015-16 estimates have included the correction of the error which has overstated the 2014-15 tax gap.\u2019]" time="0.023"><properties><property name="score" value="0.006060107382458968" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00606011&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00606011
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[9 ATC ? What is an ATC and how do I use one? An ATC is an automatic transfer switch. It is used to transfer power to the boiler in the event of a power failure. These switches are &quot;automatic&quot; in the sense that they monitor the power flow through the generator and as soon as the power goes off, they engage the transfer switch (in parallel to the house breaker) to power the boiler. When the power is restored, the ATC turns off and disconnects the power to the boiler, restoring the house power. (We are not responsible for damages caused by not following these instructions). Before starting to install the system, you need to find the right place to put your ATC switch. Some considerations include: 1. You should put your ATC as close to the back of the generator as possible. This will make it easier for you to wire it up. 2. It is important to find a place that is dry. You can't use an ATC outside. It will just short out if it gets wet. 3. It needs to be in an accessible place (inside your home). It doesn't need to be near the breaker box (but it should be near a wall outlet). 4. The ATC should be installed at least 3 feet from the nearest source of water (sump pump, water heater, etc.) 5. If your generator is in a cabinet, you should put the ATC switch on the side of the cabinet so it is easy to]" time="0.013"><properties><property name="score" value="0.028813208360239928" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02881321&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02881321
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[One of my best friends, Rachel, is pregnant and is due in April, so of course I had to make her something. She loves to read, so I made her a book gift! It has 2 to choose from; The Mad Tea Party by Lauren Child and Mary Poppins.\n\nI used the Victoria Plum and The Grey from the Victoria Plum fabric line.\n\nI found the bookmarks at a scrapbooking store and thought they were adorable.\n\nI wish I could remember where I found the quote, I think I searched Etsy and bought it there.\n\nI thought it would be cute to make little tags to put the names on so she could just hand them out.\n\nI love the bright colors on the back!\n\nThank you for looking!]" time="0.012"><properties><property name="score" value="0.058736174662271656" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05873617&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05873617
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The original photographic image is composed of five photos stitched together with Photoshop. Each photo is on a different curve in order to show the fullest scope of the incredible reach of the fire.\n\n[See &quot;more photos&quot; below for further info and versions.]]" time="0.014"><properties><property name="score" value="0.023974168352297598" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02397417&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02397417
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The map below shows the approximate location of every reported gorilla location between October 2010 and September 2011.\n\nThe red circles show where the gorillas have been reported.\n\nTo see the previous year's map, click here.\n\nMapping gorillas is difficult because of the rugged and thick vegetation in the forests of Congo, where they are found.\n\nOften the trackers do not actually see the gorillas themselves, only hear their distinctive screams, and have to make their estimates based on signs such as footprints and vegetation trampled by the large apes.\n\nHowever, the gorilla range is thought to have expanded significantly in the last five years.\n\nThe good news is that the population is now thought to be in excess of 4,000 gorillas.\n\nThere are many more threats to gorilla survival than loss of habitat, however, and none are yet under control.\n\nThe gorilla is Africa's most threatened mammal, and faces threats from poaching and disease, in addition to the loss of their habitat, and the practice of hunting them for &quot;bushmeat&quot;.\n\nThe Rwandan genocide of 1994 also left a terrible legacy, with the influx of refugees leaving a legacy of poaching and poverty.\n\nThe official policy of the Republic of Congo is that all gorillas should be left in peace in their natural habitat, which is now also under pressure from logging.\n\nThe WWF is helping to train a team of gorilla trackers to continue to monitor the gorillas' progress.\n\nYou can support their work by donating to the WWF.]" time="0.012"><properties><property name="score" value="0.05287562168095339" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05287562&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05287562
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[A neighborhood staple since 2003, Shawarma Grill offers quality Mediterranean cuisine, including an extensive menu of vegetarian and vegan options. Customers rave about the huge portions, reasonable prices, and friendly staff. If you're looking for healthy options, Shawarma Grill has salads, burgers, and baked potatoes on the menu. The vegetarian gyro is another hit, and the shawarma is sure to please any meat lover.]" time="0.012"><properties><property name="score" value="0.027033054607605855" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02703305&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02703305
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Related articles: 1/3\n\n2/3\n\n3/3\n\nAre you looking for a way to improve your website\u2019s search engine rankings? How about better user experience? There are many ways you can accomplish both.\n\n1. Reduce your page\u2019s load time\n\nGoogle uses the page load time as one of the factors when determining your search engine rankings. You can use PageSpeed Insights to measure the page load time for any URL.\n\nYou should aim for a score of 90/100 and lower. It\u2019s worth pointing out that although this factor is being used as a ranking factor, a slow site will have a negative impact on user experience, which in turn will affect your brand image, search results, and conversions.\n\n2. Enable GZIP compression\n\nGZIP compression compresses the HTML, CSS, and JavaScript files in your web pages, which reduces the size of those files and the total data transferred to the user. Google uses this factor to determine page load time.\n\nPageSpeed Insights has a compression test that allows you to measure the level of compression achieved by your website.\n\nFor more information on how to enable GZIP compression, see this article.\n\n3. Enable HTTPS\n\nHTTP Strict Transport Security (HSTS) allows websites to tell browsers to avoid interacting with any sites except the ones that use HTTPS. This is useful in preventing man-in-the-middle attacks.\n\nAs of this writing, Google considers HTTPS as a ranking factor in its search algorithm.\n\nThe problem is that although enabling HTTPS is pretty easy (for WordPress websites), it\u2019s not so easy when you have to replace all of your URLs with HTTPS URLs, and make sure that your backlinks are using HTTPS.\n\nIf you\u2019re not ready to migrate, you can use the free tool from CloudFlare, which allows you to encrypt your WordPress website and even apply SSL to all of your images, making the switch to HTTPS easy and effective.\n\n4. Check your internal links\n\nInternal links are links that point to another page on your own website. Google uses these links to determine the value of the pages in your website, and how well your site is structured.\n\nLink-building specialists will tell you that there are a lot of ways to build good quality links.\n\nFor example, you can use social bookmarking sites, forums, and blogs. For links pointing to your website, you can ask friends, relatives, clients, and business partners to post links to your website on their websites.\n\nBut, you should do this within reason, and avoid looking like a link-building spammer.\n\n5. Include internal links in your blog posts\n\nYour blog posts should include internal links, to ensure that Google can detect the relationships between your posts. This will also improve your readers\u2019 experience by allowing them to quickly find the posts that are most relevant to them.\n\n6. Create an XML sitemap\n\nAn XML sitemap is a page on your website that contains a list of all the URLs on your website. If you\u2019ve been blogging for some time, you\u2019ve probably already created a sitemap to allow search engines to crawl your site.\n\nThe problem is that search engines do not use the sitemap file on your website to crawl your website. In other words, sitemap file is useless to search engines.\n\nFortunately, there\u2019s a free plugin from Yoast that can automatically generate an XML sitemap and also add the sitemap to your website.\n\n7. Install social sharing plugins\n\nSocial media is a great way to market your website and brand. Not only that, you can also use social media as a way to improve your search engine rankings.\n\nWhen you share your content on social media, you can make the content more accessible to your visitors. Not to mention the fact that social sharing plugins also provide users with a link to your content, which is great for SEO.\n\nSome of the best social sharing plugins for WordPress websites include Facebook, Pinterest, and Twitter.\n\n8. Install and enable W3 Total Cache\n\nW3 Total Cache can greatly improve your website\u2019s load speed by enabling caching for your website\u2019s pages, posts, and images.\n\nYou can read my review of W3 Total Cache to learn more.\n\n9. Install and enable Yoast SEO\n\nYoast SEO allows you to configure your search engine settings for each post and page on your website. It\u2019s the most powerful SEO plugin for WordPress.\n\nFor more information on how to configure Yoast SEO for optimal SEO, see my tutorial on SEO for WordPress.\n\n10. Install and configure Google XML Sitemaps\n\nThis is another free plugin from Google that generates an XML sitemap for your website. This sitemap will allow search engines to crawl your website faster and more efficiently.\n\nIt also helps you to build quality backlinks to your website. The links in the XML sitemap are much stronger than the links in your sitemap file, because they are permanent links.\n\nYou can read more about how to configure Google XML Sitemaps for optimal SEO.\n\n11. Install and enable Google Analytics\n\nGoogle Analytics is a free analytics tool from Google that lets you monitor your website\u2019s performance, including the number of visitors you\u2019re getting, where they\u2019re coming from, and what they\u2019re doing on your website.\n\nI have my own tutorial on how to configure Google Analytics for optimal SEO.\n\n12. Install and enable Google Webmaster Tools\n\nWebmaster Tools is another free tool from Google that lets you verify your website and configure a wide range of settings, including your preferred domain, how your website is indexed, what the canonical URLs are, whether you allow Google to crawl dynamic pages, whether you allow web crawlers to follow your links, and whether you allow Google to scan your pages for malware.\n\nWebmaster Tools is especially useful for websites that have a lot of dynamic pages, and pages that are indexed using non-standard URLs.\n\n13. Install and enable Google Analytics by Yoast\n\nGoogle Analytics by Yoast is another free plugin that allows you to synchronize your Google Analytics account with the most popular SEO plugin for WordPress, Yoast SEO.\n\nWith this plugin, you can easily add your Google Analytics tracking code to all your posts, pages, and sidebars, without having to manually add the tracking code for each post.\n\n14. Submit your website to Google\n\nIf you have a new website, you should submit it to Google Webmaster Tools, which will allow Google to index your website and create a site map.\n\n15. Install a contact form plugin]" time="0.077"><properties><property name="score" value="0.0012613615440022251" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[As the fashion world is taking notice of Athleisure (a new style trend of casual attire that incorporates sports-luxe style elements), a new video called The Slits, from G. Garvin (a C-level marketing executive), perfectly illustrates why we love this trend. The video features an athleisure-clad woman who uses the sporty-swag inspired outfit to complete a variety of day-to-day tasks in style.\n\nThe music to the video is performed by Miley Cyrus and written by Cyrus and Pharrell Williams. The music video is part of a larger promotion for the G. Garvin brand.]" time="0.013"><properties><property name="score" value="0.049281759091746036" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04928176&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04928176
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[.\n\n\n\nAttenzione:\n\nL'ingresso al museo \xe8 gratuito.\n\nLa visita ai locali espositivi e ai musei ad esso collegati \xe8 a pagamento.\n\nIl costo della visita \xe8 di \u20ac 6,00 per adulto e di \u20ac 3,00 per bambini fino ai 12 anni.\n\n\n\nOrario\n\nAperto il marted\xec, mercoled\xec, venerd\xec e sabato dalle 9:00 alle 17:30.\n\nGioved\xec dalle 9:00 alle 22:00.\n\nDomenica dalle 10:00 alle 17:30.\n\n\n\n]" time="0.015"><properties><property name="score" value="0.07251195103877284" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07251195&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07251195
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[In today\u2019s digital age, many people look to the internet to learn about and purchase products. These include retail shoppers who look to Amazon for new toys or appliances, travelers who look to Expedia for airline tickets, and even patients who look to the internet to learn about a disease or find a doctor. This presents unique challenges for brand owners. While brand owners can claim ownership of certain content on their sites and apps, much of the internet is controlled by other entities. This includes social media sites, search engines, content syndicators, content aggregators, and other web sites. A brand owner who does not protect its name and trademarks on these sites risks losing control of its brand online.\n\nIn the past, brand owners who found a third-party using its trademarks or similar trademarks had to prove that the third-party was using the mark in a confusing manner. This required proving that the third-party\u2019s use of the mark was likely to cause confusion in the mind of the average consumer, i.e. that it was likely to be mistaken for the mark of the brand owner. Today, however, that burden of proof is no longer necessary.\n\nA little over one year ago, the Ninth Circuit Court of Appeals decided the landmark case of Global-Tech v. eBay, 562 F.3d 982 (9th Cir. 2009). The Global-Tech decision held that a plaintiff can bring a claim against a third-party for cybersquatting by using the new procedures set forth in the Anti-Cybersquatting Consumer Protection Act of 1999 (\u201cACPA\u201d). Previously, under the ACPA, a brand owner had to prove confusion to make a cybersquatting claim. Under the new regime, the burden has been placed on the third-party to show that its use of a domain name is legitimate.\n\nThe burden shift was the most significant change in the Global-Tech case. Before Global-Tech, a cybersquatting plaintiff had to prove confusion. After Global-Tech, a cybersquatting plaintiff only needs to prove that a defendant has registered a domain name similar to the plaintiff\u2019s mark with bad faith intent to profit.\n\nAccording to the Ninth Circuit, \u201cthe ACPA\u2019s [bad faith] requirement means little more than that a domain name registrant must intend to profit from the goodwill associated with the plaintiff\u2019s mark.\u201d\n\nOnce a cybersquatting plaintiff has proven this intent, a plaintiff need only show that it owns a trademark that is similar to the defendant\u2019s domain name in order to prevail on its claim. This is the essence of the burden shift. A plaintiff who successfully proves that the defendant is cybersquatting on a domain name that is similar to its trademark does not need to prove that consumers are being confused. The cybersquatter must show that its use of the domain name is legitimate.\n\nThe Global-Tech case also made it easier for a plaintiff to win a cybersquatting case because of the additional burdens placed on a defendant.\n\nFirst, the Global-Tech case made it harder for a cybersquatting defendant to defend itself on the merits. Before Global-Tech, a defendant could defend a cybersquatting claim by arguing that its use of the domain name was not confusing to consumers. However, after Global-Tech, a defendant can no longer rely on its good faith use of the domain name to show that it is not cybersquatting. Instead, a defendant must show that its use of the mark is legitimate. This has led to claims of cybersquatting where the mark and the domain name are similar, but where there is no evidence that consumers are being confused.\n\nSecond, the Global-Tech case made it easier for a cybersquatting plaintiff to obtain injunctive relief. Traditionally, a cybersquatting plaintiff had to show that it was likely to prevail on its claim in order to obtain an injunction to prevent a defendant from using its domain name. After Global-Tech, a plaintiff need only show that the defendant is likely to profit from its use of the domain name to obtain an injunction. Since an injunction may be obtained before trial, this allows plaintiffs to eliminate the use of their marks by defendants before they can defend their actions in court.\n\nThe Global-Tech case has led to a shift in the burden of proof in cybersquatting cases. This shift has created an environment where cybersquatting cases are becoming more difficult to defend.]" time="0.068"><properties><property name="score" value="0.04186171848509268" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Rated 5 out of 5 by Bullockie79 from The only thing better than this shirt is the price I\u2019m a bigger guy and I like to buy shirts that fit my body. This shirt is long enough in the body and the sleeves, so the shirt fits properly. I\u2019m very pleased with the weight of the shirt. It\u2019s not too heavy. So I don\u2019t get too hot.\n\nRated 5 out of 5 by kramer from great product This is the second shirt I have purchased in this style. I wear them almost every day and wash them many times. They are very durable.\n\nRated 5 out of 5 by JJG1 from Great shirt Great shirts. Great quality and material. Washes great and hangs out on the line to dry. Great with the Carhartt suspenders.\n\nRated 5 out of 5 by Bihangir from Fantastic! Awesome product, awesome quality. Will buy more for sure. I bought this on sales and even then, it is a good buy.\n\nRated 5 out of 5 by Anthony from A go-to for work, play, anywhere! Great fitting shirt, true to size. A comfortable classic for any occasion. It is heavy weight, but does not feel thick, just a nice fit.\n\nRated 5 out of 5 by Carpenter from Best dress shirts These are great shirts they fit well, they are long enough. I would recommend them to anyone.\n\nRated 5 out of 5 by KLJ93 from Great fit These shirts are great for work or everyday wear. They fit perfectly and are of very high quality.]" time="0.013"><properties><property name="score" value="0.04536172943914174" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04536173&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04536173
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[One of my favourite times in the barbershop. I love conversations with guys about life, business, sex, sports and news.\n\n\n\nA friend of mine said, &quot;The best conversations happen over a good cup of coffee&quot;.\n\n\n\nThat\u2019s what I\u2019ve done in my new book... \u2018Good Coffee, Bad Conversations\u2019.\n\n\n\nIf you\u2019ve got a few minutes, I\u2019d like to share a story with you.\n\n\n\nIn the late 90\u2019s, my business partner, Jeff \u2013 a good friend of mine and also a barber \u2013 was having problems. He was having lots of conflict with his wife and daughter and he was angry all the time.\n\n\n\nI asked him if he wanted to talk about it.\n\n\n\nHe said, \u201cNo. That would be good coffee, bad conversation.\u201d\n\n\n\nIt\u2019s true, Jeff was scared to open up and share what was really going on in his life. He was afraid of what he would hear.\n\n\n\nI remember this time because I wanted to help Jeff, but I didn\u2019t know what to do. So I did the only thing I knew how to do.\n\n\n\nI started writing.\n\n\n\nAs I wrote, I realized that coffee was a lot like conversation.\n\n\n\nA cup of coffee was good.\n\n\n\nA cup of coffee was bad.\n\n\n\nThat\u2019s when it hit me... \u2018Good Coffee, Bad Conversations\u2019.\n\n\n\nIt\u2019s about the conversations we have and how they can make or break our lives.\n\n\n\nIn my book, I show you the seven bad conversations you need to stop having. You\u2019ll also learn the seven good conversations you should start having instead.\n\n\n\nAs a bonus, I also share the six good conversations I\u2019ve had with some amazing people including Grant Cardone, Kody Deis, Andy Frisella, Gary Vaynerchuk and others.\n\n\n\nThere\u2019s more, but you\u2019ll have to read it to find out.\n\n\n\nThe book is short and concise. It\u2019]" time="0.026"><properties><property name="score" value="0.03818275365604124" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03818275&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03818275
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Magic: the Gathering has launched the Dominaria campaign. Dominaria is a plane steeped in rich lore and ancient magics. The world is home to several important factions, including Benalia, home to the Knights of Benalia, and the nation of Corondor. Corondor is ruled by Baron Sengir and his brother, Baron Sardu.\n\nDominaria is also home to some of Magic\u2019s most iconic characters. The Weatherlight is crewed by a legendary host of characters, including Captain Sisay, Karn, and Teferi. But the Weatherlight\u2019s roster is far from limited to famous crew members.\n\nThis is the history of the Weatherlight\u2019s crew members who weren\u2019t famous enough to get their own card.\n\n(Special thanks to Jason Carl\u2019s Comprehensive Rules FAQ and to YSDC\u2019s flavor text wiki for their incredible flavor text sources.)\n\nVenser, Shaper Savant\n\nVenser is an artificer who discovered his latent Planeswalker abilities while recovering from an injury on the plane of Shandalar. After being recruited to join the crew of the Weatherlight, Venser eventually fell into a coma and later revived himself on the plane of Koilos.\n\nKarn, Silver Golem\n\nKarn is a silver golem who used to live on the plane of Argentum. He came aboard the Weatherlight after his homeworld was destroyed by a time riptide.\n\nAven Windreader\n\nAven Windreader is a scout who used to serve in the Feather clan of Aven, a nomadic race. She is joined on the Weatherlight by two siblings: Aven Brigadier and Aven Trooper.\n\nSisay, Captain\n\nSisay is a captain from the nation of Jamuraa, and is the Captain of the Weatherlight.\n\nSquee, Goblin Nabob\n\nSquee is a goblin on the Weatherlight crew. His fellow goblin cohorts on the crew are Rishadan Cutpurse, Rishadan Footpad, and Greel.\n\nIrina, Great Glass-Spinner\n\nIrina is an illusionist who used to serve the Imperial Court of Otaria. She was exiled after the court was overthrown and joins the Weatherlight in search of answers.\n\nSharuum the Hegemon\n\nSharuum is an artificer who, in her past life, became a planeswalker. Sharuum built a stronghold called Koilos on the plane of Koilos, and served as a guardian to the mysterious stronghold.\n\nEdric, Spymaster of Trest\n\nEdric was a spy who was arrested after his cover was blown. Edric\u2019s home nation of Trest was destroyed, and he was freed from his cell aboard the Weatherlight in exchange for his assistance on the ship\u2019s infiltration of Phyrexia.\n\nSquee, Goblin Nabob\n\nSquee is a goblin who served as the goblin court advisor to the Supreme Arbiter of the Imperial Court of Otaria. He joins the Weatherlight crew after being abandoned in Krov on Dominaria, and joins the crew in order to find a way home.\n\nTahngarth, Talruum Hero\n\nTahngarth is a Tahngarth is a minotaur who hails from the nation of Talruum. Tahngarth was captured by Sarpadian pirates and sold into slavery. He was rescued from his servitude by the crew of the Weatherlight.\n\nKarn, Silver Golem\n\nKarn was a silver golem who lived on the plane of Argentum. He boarded the Weatherlight after his homeworld was destroyed by a time riptide.\n\nKarn, Silver Golem\n\nKarn is a silver golem who used to live on the plane of Argentum. He came aboard the Weatherlight after his homeworld was destroyed by a time riptide.\n\nStarke of Rath\n\nStarke of Rath is a master thief who joins the Weatherlight crew after discovering a conspiracy to overthrow the current government of his native country of Rath.\n\nKarn, Silver Golem\n\nKarn was a silver golem who used to live on the plane of Argentum. He came aboard the Weatherlight after his homeworld was destroyed by a time riptide.\n\nGryff, Knights of Benalia\n\nGryff was a member of the Knights of Benalia who died in a battle against an Onakke Ogre. He was later revived by Tiana, Ship\u2019s Caretaker, who repurposed his damaged body as a messenger for the ship.\n\nKarn, Silver Golem\n\nKarn is a silver golem who used to live on the plane of Argentum. He came aboard the Weatherlight after his homeworld was destroyed by a time riptide.\n\nTiana, Ship\u2019s Caretaker\n\nTiana is a ship\u2019s caretaker who was attached to the Weatherlight before it was even built. She was responsible for the upkeep of the ship, including a number of tiny, flying custodians.\n\nTiana, Ship\u2019s Caretaker\n\nTiana is a ship\u2019s caretaker who was attached to the Weatherlight before it was even built. She was responsible for the upkeep of the ship, including a number of tiny, flying custodians.\n\nAven Envoy\n\nAven Envoy was a captain who used to serve in the Knights of Avon. She lost her entire company during a battle against an Onakke Ogre. Aven Envoy was revived by Tiana, Ship\u2019s Caretaker, and took up a new post as envoy for the Weatherlight.\n\nUrza Planeswalker\n\nUrza was a planeswalker who, after his death, was merged with a powerstone and transformed into an artificial planeswalker. He was the first person to travel to the plane of Phyrexia, and later forged a coalition of planeswalkers against the Phyrexians.\n\nSquee, Goblin Nabob\n\nSquee is a goblin who served as the goblin court advisor to the Supreme Arbiter of the Imperial Court of Otaria. He joins the Weatherlight crew after being abandoned in Krov on Dominaria, and joins the crew in order to find a way home.\n\nIcatian Lieutenant\n\nIcatian Lieutenant is a soldier from the nation of Icatia. She joined the crew of the Weatherlight in order to track down her fellow Icatian soldiers who were captured by Rathi forces.\n\nKarn, Silver Golem\n\nKarn is a silver golem who used to live on the plane of Argentum. He came aboard the Weatherlight after his homeworld was destroyed by a time riptide.\n\nBarren Glory\n\nBarren Glory is a captain who came from an apocalyptic future. He used to captain a ship called the Weatherlight, which is (unsurprisingly) a ship not unlike the Weatherlight.\n\nJhoira, Weatherlight Captain\n\nJhoira is a master artificer who was captain of the Weatherlight for a time.\n\nGryff, Knights of Benalia\n\nGryff was a member of the Knights of Benalia who died in a battle against an Onakke Ogre. He was later revived by Tiana, Ship\u2019s Caretaker, who repurposed his damaged body as a messenger for the ship.\n\nCaptain Lannery Storm\n\nCaptain Lannery Storm was a member of the crew of the Weatherlight who was killed during a mission to Rath. She was later revived by Tiana, Ship\u2019s Caretaker, and now serves on the Weatherlight as a crew member.\n\nNimbus Maze\n\nNimbus Maze was an artificer who served the Knights of Benalia. He was sent to accompany Squee on a diplomatic mission to Icatia, but was killed by a phyrexian vat lord during the visit.\n\nSquee, Goblin Nabob\n\nSquee is a goblin who served as the goblin court advisor to the Supreme Arbiter of the Imperial Court of Otaria. He joins the Weatherlight crew after being abandoned in Krov on Dominaria, and joins the crew in order to find a way home.\n\nAjani, Mentor of Heroes\n\nAjani is a leonin planeswalker who was once the leader of the Capashen clan on the plane of Alara. His pride led to a war between the clans of Capashen and Kumena. After the war, Ajani renounced his leadership and traveled to Dominaria.\n\nSquee, Goblin Nabob\n\nSquee is a goblin who served as the goblin court advisor to the Supreme Arbiter of the Imperial Court of Otaria. He joins the Weatherlight crew after being abandoned in Krov on Dominaria, and joins the crew in order to find a way home.\n\nBarren Glory\n\nBarren Glory is a captain who came from an apocalyptic future. He used to captain a ship called the Weatherlight, which is (unsurprisingly) a ship not unlike the Weatherlight.\n\nSquee, Goblin Nabob\n\nSquee is a goblin who served as the goblin court advisor to the Supreme Arbiter of the Imperial Court of Otaria. He joins the Weatherlight crew after being abandoned in Krov on Dominaria, and joins the crew in order to find a way home.\n\nKarn, Silver Golem\n\nKarn is a silver golem who used to live on the plane of Argentum. He came aboard the Weatherlight after his homeworld was destroyed by a time riptide.\n\nCaptain Lannery Storm\n\nCaptain Lan]" time="0.074"><properties><property name="score" value="0.017287799424114302" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[BestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disinc]" time="0.078"><properties><property name="score" value="0.5331452595662242" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[All the new games for the week\n\nOnce upon a time, the first week of the month meant a new Switch eShop release was imminent, but recently, Nintendo has shown an admirable level of restraint. The first week of this month is no different, as there's just a couple of new games on the way to the system's digital store.\n\nComing to North America this week is the sequel to the surprise hit of 2017, Wonder Boy: The Dragon's Trap. The platformer has a rich history on the PC and consoles, but this remake proved to be well worth the time spent by the fans. With a new look and modern gameplay, this is certainly worth your time if you haven't played it yet. The Dragon's Trap releases in North America on February 9.\n\nFor our European readers, this week brings us into the dangerous world of SteamWorld Dig 2. While the first game was critically acclaimed and earned quite a bit of love, this sequel looks to improve on the formula by adding a lot more variety and some online multiplayer options. Dig 2 is a bit more than a week away, and will be available on February 9.\n\nIf that's not enough for you, then you'll be happy to know that the Nintendo Switch's Virtual Console service has something new in store for you. For the week of February 13, the Nintendo 64 classic Donkey Kong 64 will be available on the system's store.\n\nAs always, we'll keep an eye on the eShop and update you on new releases if and when they become available. If you're holding out for some new physical releases, we should have some news later today.\n\nThis article may contain links to online retail stores. If you click on one and buy the product we may receive a small commission. For more information, go here.]" time="0.026"><properties><property name="score" value="0.07204011375176646" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07204011&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07204011
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[mikeshea.net\n\nScribblechat: Thoughts and Notes on Black Mountain\n\nby Mike Shea on 3 October 2003\n\nThis morning I saw the movie &quot;Black Mountain&quot; in a theatre near me. I had been really looking forward to it since I heard about it months ago. I enjoyed it a lot, but the final scene didn't end with the punch I was expecting. The entire movie was a quiet, intense build-up to that one final scene and, as much as I enjoyed the movie, I found myself disappointed at the end.\n\nThe film is good, I highly recommend it, but be ready for an intensely quiet movie. You might even want to bring a set of earplugs to help block out the noise from the rest of the theatre.\n\nSend comments to mike@mikeshea.net or follow @mshea on Twitter. If you enjoyed this article, please use this link to Amazon.com for your next online purchase.]" time="0.013"><properties><property name="score" value="0.0720209023908902" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0720209&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0720209
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[View Full Version : How can I identify potential variables?\n\nMikeJackness Hi all,\n\nI'm trying to learn about and identify some potential (input) variables for a simple device.\n\n\n\nI've looked at the classic approach using the Analyser to get a range of DC voltages (input values) and their corresponding output values but that's not what I'm looking for.\n\n\n\nI want to find out what variables may affect the final values I am looking for and why they have an effect on the outputs.\n\n\n\nAn example would be, lets say I am looking for DC values of between 2V and 4V. The device I am testing has a button which, when pressed, will alter the value of something between 1% and 100%. How can I find out what that value is? I know that it's 1% - 100%, but I want to know the value of the device and how it affects the outcome.\n\n\n\nThanks\n\nMike\n\nIan Smith You need to identify what it is you are actually measuring.\n\n\n\nAre you measuring the current through a certain circuit? If so, you need to measure the voltage drop across it (and also the current through the circuit itself)\n\nAre you measuring the temperature of a thermocouple?\n\nAre you measuring the force produced by a tensile tester?\n\netc etc.\n\n\n\nThe only way to find out is to identify what is actually changing (when the button is pressed) and what is being measured.\n\n\n\nHope that helps.\n\nMikeJackness You need to identify what it is you are actually measuring.\n\n\n\nAre you measuring the current through a certain circuit? If so, you need to measure the voltage drop across it (and also the current through the circuit itself)\n\nAre you measuring the temperature of a thermocouple?\n\nAre you measuring the force produced by a tensile tester?\n\netc etc.\n\n\n\nThe only way to find out is to identify what is actually changing (when the button is pressed) and what is being measured.\n\n\n\nHope that helps.\n\n\n\nIt does help a lot, thank you. I'm looking at the values in the actual device rather than the sensor values.\n\n\n\nI'm trying to test a stepper motor, and I want to see if the resistance of the motor changes when it turns.\n\n\n\nI guess I'll have to figure out what the resistance is when the motor is not turning and what it is when it is, then identify the variables that affect it.\n\n\n\nAny more ideas would be great.\n\nIan Smith I'm trying to test a stepper motor, and I want to see if the resistance of the motor changes when it turns.\n\n\n\nAh! Well, you've gone and identified the system (stepper motor) and the variables (resistance). Now all you have to do is find out the actual resistance and how it changes when the motor is turned.\n\n\n\nAlso, have you determined whether the motor is powered from an AC mains supply or from a DC battery (or similar)? You need to know this because AC mains supplies may vary the current they provide, depending on the load (which may include the motor), and some DC supplies may have a current limit (which you may need to bypass if it's too low).\n\n\n\nHope this helps.\n\nMikeJackness Hi all,\n\nI've been doing some reading and thinking about how to approach this, and it looks like my answer is to use two multimeters.\n\n\n\nOne to measure the voltage of the stepper motor (when turned) and the other to measure the current flowing through it (when turned).\n\n\n\nThe questions are;\n\n\n\nWhat kind of multimeter should I use?\n\nShould I use a panel mount multimeter for ease of use, or a DMM?\n\nShould I try to build a sensor for the current measurement?\n\nIan Smith Hi all,\n\nI've been doing some reading and thinking about how to approach this, and it looks like my answer is to use two multimeters.\n\n\n\nOne to measure the voltage of the stepper motor (when turned) and the other to measure the current flowing through it (when turned).\n\n\n\nThe questions are;\n\n\n\nWhat kind of multimeter should I use?\n\nShould I use a panel mount multimeter for ease of use, or a DMM?\n\nShould I try to build a sensor for the current measurement?\n\n\n\n1. Both, to begin with. You may only need one, but you may need two, and it's better to have two of them.\n\n\n\n2. As above. If you can't build the current measurement, then either use one of the basic models that has a current setting, or use the voltage reading on the meter and then calculate the current from that. (e.g. multiply by 0.6 to get approximate amperes)\n\n\n\n3. Again, as above. If you can't build a current measurement, then you can get a panel meter to do the job for you, or you can try one of the fancier meters that have current measurement as one of their functions.\n\nMikeJackness Thanks for the replies. I am currently using two multimeters, a panel meter and a handheld meter.\n\n\n\nI've built an adjustable resistance using one of the 5W, 12V heater elements from a microwave. When I put it in series with the motor it showed a resistance of around 400 Ohms.\n\n\n\nI was then trying to identify how much current it drew when it was switched on, and I couldn't get a good result because the panel meter wasn't reading it properly.\n\n\n\nI think that's where I'm at with it at the moment. I may try using my handheld multimeter to see if it can read the current draw, but that may prove difficult because I'm not sure where the current is going through the stepper.\n\n\n\nThanks again for all the advice.\n\nIan Smith Thanks for the replies. I am currently using two multimeters, a panel meter and a handheld meter.\n\n\n\nI've built an adjustable resistance using one of the 5W, 12V heater elements from a microwave. When I put it in series with the motor it showed a resistance of around 400 Ohms.\n\n\n\nI was then trying to identify how much current it drew when it was switched on, and I couldn't get a good result because the panel meter wasn't reading it properly.\n\n\n\nI think that's where I'm at with it at the moment. I may try using my handheld multimeter to see if it can read the current draw, but that may prove difficult because I'm not sure where the current is going through the stepper.\n\n\n\nThanks again for all the advice.\n\n\n\nI can't help with the current measurement, I don't know much about motors.\n\n\n\nBut, if you]" time="0.071"><properties><property name="score" value="0.09615104370112332" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Now you can print documents, or send a fax, from virtually anywhere. If you have an Internet connection and a phone, you can create, send, and receive faxes from the convenience of your home or office.]" time="0.015"><properties><property name="score" value="0.007465299995862429" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0074653&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0074653
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Unbelievably cute.\n\n\n\nThis is a review of the artbook for Lucky Star. I'm not going to write a full review, because you can find that on other sites.\n\n\n\nLucky Star is a 4-panel comedy manga that revolves around the lives of four girls. In this artbook, it also includes background sketches, 3D illustrations of some of the characters, and even some concept art of the outfits.\n\n\n\nThe artwork is cute, and will make you smile, and for me, that's what makes this a good book. It has all the artwork for the series, and all the side-materials as well.]" time="0.016"><properties><property name="score" value="0.023603420632278826" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02360342&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02360342
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[LINGUIST List 12.2282\n\nThu Apr 26 2001\n\nReview: Discourse Studies\n\nEditor for this issue: Terence Langendoen &lt;terry linguistlist.org&gt;\n\nDirectory\n\nMessage 1: Genre Analysis and Discourse Analysis\n\nKen Hyland, Introducing Discourse Analysis (Blackwell Introductions to Applied Linguistics) , xii+267pp, paperback ISBN 0-631-22512-8, hardback ISBN 0-631-22511-X, $47.95 / ??35.00 The blurb on the back: This textbook introduces students to the main traditions of discourse analysis, from structuralism to modern approaches, and includes many examples of actual texts for study and analysis. Hyland has developed his own approach, based on a number of leading methods, to show the student how to apply them to language use. Divided into four parts, this book explores the key concepts, applications and findings of these methods, in the first section looking at language and discourse as a unit, in the second examining patterns of text structure, in the third explaining how discourse is used to construct particular kinds of meaning, and in the fourth looking at how we study discourse. The first three sections introduce three different approaches to discourse analysis - the macro-structural approaches of Saussure and Gee, the micro-structural approach of Halliday, and the interpersonal approaches of Brown and Hodge. The last section explores ways in which different kinds of discourse are studied, including conversation, literature, power, work and the media. What I liked: This is an excellent textbook. It provides a good survey of a range of discourses analysis approaches and provides clear descriptions of the methodological principles of these approaches. This is an approachable introduction to a complex and diverse field. The topics covered include the analysis of conversation, text and the various modes of literature. Hyland shows how the same principles can be applied to these different types of text. He draws on a wide range of literature to show how discourses can be studied and includes many examples of actual texts. The book is very clearly written, contains many well-chosen examples, and provides a good introduction to a subject which is likely to be of increasing interest to Linguists. What I didn't like: The chapter on the analysis of conversation and the chapter on the analysis of discourse in fictional texts would benefit from a more rigorous discussion of the key concepts used in this type of analysis. For example, the chapter on conversation provides an excellent introduction to this type of discourse, but it would be better if there was a clearer discussion of the concepts of turn-taking and overlap, and a more careful explanation of the use of 'look-back' studies and the main types of transcription. The chapter on discourse in fictional texts provides an introduction to a type of analysis which has not been covered elsewhere in the book, and this is an important addition to the text. However, this chapter does not include the discussion of the key concepts of discourse analysis which is provided in the other chapters. It would have been better if this chapter could have followed the same format as the other chapters, and discussed how a particular approach was developed, what its key concepts are, and how these concepts can be applied. I think that this is an important addition to the book. Another minor point is that there are several typos in the book, but I don't think that this detracts from the overall quality of the book. What is missing: A discussion of the key concepts of discourse analysis which can be applied to all types of text.]" time="0.126"><properties><property name="score" value="0.0204358066053969" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Home Articles &amp; Publications\n\nHOME ARTICLES &amp; PUBLICATIONS\n\nFind out more about the Ark Community Foundation and our many achievements here.\n\nIN THE NEWS\n\nWHAT'S AHEAD\n\nSee what exciting events and programs the Ark Community Foundation is involved in this year!]" time="0.014"><properties><property name="score" value="0.03396992015777811" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03396992&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03396992
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[More than 600 students in the CSU system are waiting for their transcripts to be cleared.\n\nTranscripts at CSU campuses are frozen for over 600 students who attended the Pueblo campus last semester, and if they don't complete their credits and final exams, they could lose their scholarships and their seats.\n\nAccording to the CSU-Pueblo website, 615 students have not yet completed their spring semester. They will lose their financial aid eligibility if they don't finish.\n\n&quot;It's like they're on a waiting list,&quot; said Dale Lange, who works at the registrar's office. &quot;We're waiting for the teachers to grade their finals and the professor to submit the grades and then they can release the transcript.&quot;\n\nOn March 20, the Pueblo campus was told to cease instruction and close by the chancellor's office.\n\n&quot;It's just horrible,&quot; said CSU-Pueblo student Julea Prentice. &quot;I don't know if I'm graduating or not.&quot;\n\nPrentice said she's upset she's paying for classes that she can't use, but she feels she doesn't have a choice.\n\n&quot;If I want to go to graduate school I have to get a certain GPA, so I can't afford to fail this semester,&quot; said Prentice.\n\nMeanwhile, her son is suffering because he needs a caretaker for his medical condition.\n\n&quot;That's why I'm a mother, that's why I'm a mom, that's why I'm a father,&quot; said Prentice.\n\nPrentice said she feels alone and doesn't know what to do.\n\n&quot;If I don't have an answer, I can't do anything,&quot; said Prentice. &quot;I just have to wait.&quot;\n\nIt's a wait-and-see game for CSU students.\n\n&quot;I just hope we all pass,&quot; said Prentice.\n\nSome students, like Adrienne Longoria, have called out for help.\n\n&quot;Help us,&quot; she said. &quot;Help us students.&quot;\n\n&quot;We would have probably been able to catch up if they had let us finish our classes,&quot; said Longoria. &quot;It's just frustrating that they can't let us finish our last few weeks.&quot;\n\nLongoria and Prentice said they want to go to graduate school, but if they don't pass this semester, they could lose their scholarships.\n\n&quot;I'm trying to be hopeful,&quot; said Prentice.\n\n&quot;It's frustrating, but I don't know what to do,&quot; said Longoria. &quot;I'm hoping we get our money back and I'm hoping we can finish this semester.&quot;\n\nCSU-Pueblo officials said they don't know when transcripts will be released because each student's professor has to grade the final exams and send them to the registrar's office.]" time="0.024"><properties><property name="score" value="0.027123767833278332" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02712377&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02712377
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Download jyrd.ptc.me Nude Selfies free porn video, hd xxx mobile porn.\n\nTorrentz will always love you. Farewell.\n\nCategory: Misc. Tagged With: hayley kiyoko jyrd nude, hayley kiyoko nude, hayley kiyoko porn.\n\ng l a m is a one stop shop for celebrity gossip news &amp; celebrity entertainment including gossip news on music gossip news on. In an interview with paper magazine on november 1, 2016, she.]" time="0.015"><properties><property name="score" value="0.055333015666314544" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05533302&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05533302
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Justin Fox\n\nIf you are the type of person who needs a car that can turn on a dime in the city but also tow your boat to the lake, the 2014 Lexus ES will probably fit the bill. When Lexus first introduced the current ES in 2006, the formula was right. But by 2012, it had become dated. The original ES was slow, thirsty and not very capable. The second generation added power and speed and was the first ES to tow. With the third generation, Lexus has taken that formula one step further and brought in all-wheel drive and a more refined driving experience. But it also kept the old formula.\n\nStyling\n\nThe 2014 Lexus ES shares its looks with the outgoing model and in a way with the new IS sedans as well. So it is a blend of a wedge shape and something more traditional. From some angles, it is vaguely reminiscent of the LFA. Its grille is pronounced, its side character lines are muscular and its curves on the hood and trunk are attractive and not overdone.\n\nIts front end is more Lexus-like than before and its tail is more sedan-like than before. It has a more modern look without being overly aggressive and its design is not overly bland like some of its competitors. It is a good looking sedan but it will be a little too plain for some people.\n\nTruck Trend's 2014 Best Interiors trophy went to the new Lexus ES. It has an attractive dash that has been borrowed from other Lexus models but it is still nice and functional.\n\nTechnology\n\nI was surprised at how good the 2014 ES was. It was comfortable and quick and the transmission was a work of art. It was easy]" time="0.035"><properties><property name="score" value="0.0894358506536356" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08943585&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08943585
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Pastry chef Kayoko Uematsu is a cake virtuoso. She creates works of art from her home-kitchen studio. She shared with us how to make delicious gourmet cakes. Watch and learn!\n\nHow to make gourmet cakes from your home kitchen]" time="0.015"><properties><property name="score" value="0.018255438356062914" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01825544&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01825544
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Did you think there\u2019s no such thing as free lunch? Wrong!\n\nI was an old punk. I ate cheaply. I ate unhealthily. I went out for food once a week, never paid more than $8 for a plate of rice and 2 pieces of chicken. Other times I ate popcorn and French Fries with my friends while watching movies in the cinema. And as I became an adult, I ate badly and grew fat.\n\nLast year, I decided to change my life. I started to save money, and from that day, I stopped eating cheaply.\n\nHere are some tips I picked up.\n\n1. Choose the cheaper meal of the day\n\nMcDonalds have only 2 types of meals that are cheaper than the rest. These are the McD Snack Meal and McD Value Meals.\n\nThe McD Snack Meal cost RM5.69, which includes a Chicken McSpicy or Filet-O-Fish, 2 Nuggets and a 375ml soft drink. The McD Value Meal costs RM8.19, which includes a Big Mac, a McSpicy or a Filet-O-Fish, 2 Nuggets and a 375ml soft drink.\n\nBoth meals are cheaper than the other meals that you can buy.\n\nThe cheapest meal of the day in McDonalds is the Cheeseburger meal.\n\n2. Eat at \u2018The\u2019 Restaurant instead\n\nIn Singapore, McDonalds is known as \u2018The\u2019 Restaurant. And you don\u2019t need to ask \u2018Where is the nearest McDonalds?\u2019, because in Singapore, \u2018The\u2019 Restaurant and McDonalds is the same thing.\n\nIt\u2019s a myth that \u2018The\u2019 Restaurant is more expensive than McDonalds. In fact, most of the meals that are available in McDonalds can be found at \u2018The\u2019 Restaurant.\n\nThe most expensive meal in \u2018The\u2019 Restaurant is the Set Dinner (Veg/Chicken). The meal costs $7.20. The most expensive meal in McDonalds is the Cheeseburger meal, which costs $4.29.\n\nAll other meals in \u2018The\u2019 Restaurant are cheaper than the other meals in McDonalds.\n\nFor example, the cheapest meal in \u2018The\u2019 Restaurant is the Curry Puff. The Curry Puff costs $3.10. The cheapest meal in McDonalds is the Cheeseburger meal, which costs $4.29.\n\n3. Eat from 7-Eleven instead\n\nThere are two reasons why I suggest you eat at 7-Eleven instead of McDonalds or \u2018The\u2019 Restaurant.\n\nThe first reason is, the food at 7-Eleven is almost as good as the food in \u2018The\u2019 Restaurant.\n\nThe second reason is, 7-Eleven sell drinks and fast food cheaper than McDonalds and \u2018The\u2019 Restaurant.\n\nFor example, I ordered a portion of Mac and Cheese in 7-Eleven. I ordered the Mac and Cheese as a take away, because I do not have any microwave in my office.\n\nThe cost of the Mac and Cheese is $3.30. And I bought a Pineapple Juice, which costs $1.30.\n\nThe total cost of my meal is $4.60. The same meal costs $4.79 in McDonalds.\n\nIt\u2019s true that the Mac and Cheese in 7-Eleven tastes a bit different from the one that I bought in McDonalds.\n\nIt\u2019s also true that the taste is not as good as the one that I can buy in \u2018The\u2019 Restaurant.\n\nBut it\u2019s true that the cost is almost the same.\n\n4. Avoid eating at the food court\n\nFood courts are expensive. The price of food sold at the food courts are almost the same as the cost of the same meal in a fast food restaurant.\n\nSo I\u2019ll suggest you avoid eating at the food court.\n\n5. Make food at home\n\nI make sure I cook once a day. I try my best to cook healthy meals. And the cost of the food that I make at home is much cheaper than the cost of a fast food meal.\n\n6. Eat with other people\n\nThe total cost of my meal depends on how many people are eating the same meal.\n\nFor example, I shared a Big Mac meal with my friend. The total cost of the meal was $8.20. That\u2019s because the meal cost $4.29 each.\n\nThe cost of the meal depends on how many people are eating the same meal. The cheaper meal is when I eat alone.\n\n7. Share the meal\n\nSometimes I share the meal with my partner. I usually share a meal with my partner once a week. I share the meal because I feel bad that my partner is eating alone while I\u2019m eating with my friends.\n\nIf I share the meal with my partner, the total cost is $4.60.\n\nIf I eat alone, the total cost is $4.79.\n\n8. Always look for a promotion\n\nIn McDonalds, \u2018The\u2019 Restaurant and 7-Eleven, there are always promotions.\n\nFor example, 7-Eleven and \u2018The\u2019 Restaurant sometimes offer discount on their drinks and fast food.\n\nSo I suggest you always look for a promotion before buying a meal.]" time="0.070"><properties><property name="score" value="0.00506075505357858" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[HapWanderlust September 1, 2011, 8:10 pm it's a very interesting product...\n\nAnonymous September 3, 2011, 12:30 am Thanks for the review, very interesting product and its feature seems to be very promising. However, I wonder what the maximum resolution would be at the time when you take picture and share it instantly, like facebook etc? Would it be just at the same resolution as of the regular picture or even better?\n\nGordon394 September 5, 2011, 12:11 pm This is one of the most intriguing products I've seen for a while. Will the quality be comparable to the smartphone quality, or will it be significantly worse? Are you sure you'll be able to get the same picture quality from the handset as with the phone tethered to a DSLR? I have to say I'm excited.\n\nMatthew September 5, 2011, 8:22 pm Thanks for the review, very interesting product. Do you know how much it will be available to the market and how it will be distributed?\n\nHiFiChris September 7, 2011, 3:56 pm Not sure about this. Love the idea but dont think it's ready. Biggest concern is image quality and would be pretty disappointing to take photos that look awful. I'd be worried about the image quality from the Nikon and would expect that some of the 'features' are just Nikon marketing. Probably best to wait until there are some real world reviews before investing.\n\nR.I. September 11, 2011, 1:16 pm This sounds amazing but what about the battery life?\n\nI really like this feature because it would be handy for traveling.\n\nPaul 4 October 25, 2011, 7:33 am would like to buy one but looks too expensive for a one off piece of kit\n\nPaul 4 October 25, 2011, 7:34 am would like to buy one but looks too expensive for a one off piece of kit]" time="0.028"><properties><property name="score" value="0.03893416237657256" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03893416&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03893416
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The leadership of the world is not giving a serious thought to the future of the world and they are not ready to save the world. If this condition continues then the future of the world will be very difficult to predict.\n\n\n\n\n\n\n\n\n\n\n\nThe world is facing some serious problems and a few of them are as follows:\n\n\n\n\n\nClimate change\n\nUnemployment\n\nIncome inequality\n\nHunger\n\nResource scarcity\n\nPublic health\n\nMass migration\n\nInternal conflicts\n\nAchievements of humankind\n\nThe most important achievements of humankind have been its ability to observe, to experiment, and to reason. These capabilities have given humans the power to conquer the world and to dominate everything. The presence of powerful armies of humans has protected humans from the evil of other humans.\n\nPeople are now able to live in many different places and these different places have different types of people.\n\n\n\n\n\nAre people ready to save the world?\n\nThe world is facing many problems, however, the leaders of the world are not ready to save the world. If this condition continues then the future of the world will be very difficult to predict. The world is facing many problems, however, the leaders of the world are not ready to save the world. If this condition continues then the future of the world will be very difficult to predict.\n\nThe countries which are rich in wealth and resources will enjoy a good future, but the countries which do not have wealth and resources will face very difficult situations. The climate change has become one of the biggest problems of the world and the government of the world is not taking serious actions to tackle this problem.\n\nIf the governments of the world are unable to deal with this problem then the world will suffer a lot from this problem. The fact is that the world is in very big trouble. If there is no action to save the world then the future of the world will be very difficult to predict.\n\nIn order to deal with the current problems of the world, the governments of the world will have to make some very tough decisions.]" time="0.023"><properties><property name="score" value="0.12809801388438835" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[&quot;It was not easy for me at all. I'm so used to the typical nursing home or hospital setting, it's hard to break out of that. And that's what I was used to. They're a very good hospital. My father is getting good care. They have good, knowledgeable, friendly staff. I really don't have any complaints. It's a little more expensive than some places, but not too bad. They have an excellent chef. He's excellent. The only complaint I have is that they are slow to come and get him. They don't come out and take him for a shower. We have to go get him. But we know they're busy. It's not like they can just send someone.&quot;]" time="0.012"><properties><property name="score" value="0.04127073461169806" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04127073&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04127073
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Definitions\n\n1. a planet in a theory of the universe that included the sun and all the other stars.\n\n2. an ancient name for the Great Bear constellation, used in astrology and mythology.]" time="0.012"><properties><property name="score" value="0.010294540881411463" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01029454&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01029454
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[A major renovation project at York University in Toronto is expected to run at least $26 million over budget, due to unanticipated problems that are forcing school officials to tear out and rebuild much of the work already completed.\n\nConstruction started in January 2014 on the $136-million project, which is meant to replace the York Arena, which opened in 1971, with a new, modernized facility with more seating and amenities.\n\nWork on the exterior of the arena was completed, but the contractor, Armel Holdings, noticed a problem when it came to the steel supports for the roof in June 2015, York University president Rhonda Lenton said.\n\n&quot;We realized there were some structural concerns around the roof, so we paused the project,&quot; Lenton told CBC Toronto.\n\nSteel reinforcements will be needed to strengthen the roof, which was built with &quot;soft&quot; steel. (York University)\n\nWorkers discovered the roof of the arena had been built with &quot;soft&quot; steel, which is not as strong as the &quot;hard&quot; steel that is typically used for these types of projects, said Lenton.\n\nIn order to make sure the project is safe, crews need to rip out and rebuild about 50 per cent of the roof, at an estimated cost of $16 million.\n\n&quot;We\u2019ve got to take this roof down,&quot; said Lenton.\n\n&quot;We will be replacing the steel \u2014 the beams \u2014 and re-engineering the entire roof structure, as well as all the supporting beams that support the roof.&quot;\n\nSchool officials say they will be applying for additional funding to cover the cost of the roof replacement.\n\n&quot;We will be going to the province for the additional funding that we will need, for sure,&quot; Lenton said.\n\nThey are also hoping to get money from the York University Foundation, which is a non-profit fundraising organization that operates separately from the university, according to its website.\n\nShe said she does not believe taxpayers will have to foot the bill for the additional cost.\n\nUnforeseen problems\n\nYork University president Rhonda Lenton says she does not believe taxpayers will have to cover the $16-million cost of rebuilding the roof. (CBC)\n\nThe problems do not end with the roof.\n\n&quot;We realized that there were some structural issues around the seating, and we\u2019re replacing the seating, the wood seating in the arena,&quot; Lenton said.\n\nIn total, there will be 12 new levels added to the arena, along with a new heating and cooling system.\n\n&quot;I think, because of the complexity of this building, we have to anticipate what other unforeseen problems may come along,&quot; Lenton said.\n\nIn December 2015, school officials sent a memo to university faculty and staff outlining the potential for &quot;additional costs and associated risks&quot; that could arise in the future.\n\n&quot;We will continue to keep all our stakeholders updated as the project progresses and when we know more details about timing and scope of work,&quot; the memo said.\n\n&quot;I\u2019m just waiting for more details,&quot; said Brian Lee, a York University professor and co-chair of the university\u2019s Local 79 union.\n\nLee said he thinks the university should have been more specific about the challenges facing the project.\n\n&quot;There is no need for it to be kept secret,&quot; Lee said.\n\n&quot;There\u2019s nothing secretive about construction projects.&quot;\n\nScheduling setback\n\nThe timeline for the completion of the project has been extended by more than a year, said Lenton.\n\nThe university had initially planned to have the arena ready by the 2017-2018 school year, but that deadline has been pushed back to the 2018-2019 school year.\n\n&quot;We did have an aggressive schedule. We have to replace a roof in a building that is up on a hill,&quot; Lenton said.\n\nLenton also said there will be more parking spaces at the arena once the project is complete. The university had to remove between 300 and 400 parking spaces in order to begin the work.\n\n&quot;I think we will have more spaces than we had before,&quot; Lenton said.]" time="0.070"><properties><property name="score" value="0.006488389916239221" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00648839&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00648839
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[October 19, 2015\n\nOnce again we find ourselves reflecting on our own good fortune as we consider the news from Paris. We are not only alive, but many of us have access to a wealth of luxury, both intellectual and material, that few of our ancestors would have dreamed of. And we also have access to media that have transformed the world. One of my sons recently discovered the fascinating series \u201cHow the Universe Works\u201d (available on Netflix). This is a television show, with a high-tech production budget, and it is fascinating. It has some of the best images of the stars and planets ever produced, with an engaging story to go with them. The soundtrack alone is amazing. I highly recommend it. But it only works because of our technology, which has the power to tell stories across the globe.\n\nYou don\u2019t have to leave the planet for science to reach you. It\u2019s right there in your living room.\n\nToday I read a story in the paper about a study that examined the safety of exercise and high protein diets. This study was designed to answer a question that most of us have faced at some point: should I cut back on protein? My story pointed out that the study was very small and inconclusive, but the real story is about how we learn about our world and why this study is inconclusive. It is a story that has been repeated again and again. When new information comes out, we all read it with the same kind of judgment that you would apply to a restaurant: is this place worth my time and money? The more of it you eat, the more of it you want to read. When I read the story in the paper, I didn\u2019t try to understand the process by which the authors made their conclusions. Instead, I simply had a gut reaction: was the story worth the cost of the information it contained?\n\nThat is the only real question that matters when we choose what to read. This question is especially important today because there is so much information out there. When the costs of information are low, it becomes much easier to consume, and with this abundance comes the difficulty of choosing what to read. So we focus on our gut. And this is because we have a brain designed to make these kinds of choices, one that is driven by emotional needs, not rational ones. We are not designed to rationally evaluate the content of each article, but instead we are designed to make emotional decisions about what to consume.\n\nThis point was well illustrated by some research done by scientists at the University of Michigan. They had two groups of subjects: one that was told to read a story that was fun and amusing, and one that was told to read a story that was sad. Then they had them make some choices about the food they would eat for the next week. The group that had been reading the sad story ate 25% fewer calories than the group that had been reading the amusing story. Why? Well, they thought that food was a reward, and since they had been unhappy in the experiment, they cut back on the number of rewards that they would give themselves. This is a very rational process, but the very fact that they were not looking at the process that led them to their choice means that it had to be some form of emotional response.\n\nIn this way, we respond to information much like we respond to food. Food provides us with energy and pleasure, so it is important to choose it well. So we think in terms of gut decisions, not because they are the best ones, but because they help us to cope with our lives. Information is also a source of energy and pleasure, and so we choose it as well.\n\nBut while we are creatures that make emotional choices about our food and information, we also have other needs. These are needs that were emphasized by the process of evolution: needs to know about the world and needs to know about ourselves. So when we read articles, or watch television, or go on social media, we need to know whether they provide this type of knowledge as well.\n\nMost of the time we are looking for a story that is interesting, and that will hold our attention. But we also want a story that tells us something about our world. The better the story is, the more likely we are to learn from it. A story about calories and exercise might not be particularly interesting, but it is important for us to know. A story about the stars and planets might be fascinating, but it doesn\u2019t matter if it is a good story or not.\n\nThis is because we need information about our world to survive and to thrive, and we know this to be true because it is encoded in our DNA. Our brains have evolved to help us process the information that we receive, and so it makes sense that we]" time="0.078"><properties><property name="score" value="0.0033038135948554653" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00330381&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00330381
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Official Download Source - Manufacturer Website\n\nWindows ME, Windows 98\n\nSoundMAX Integrated Digital Audio Driver\n\nThis package supports the following audio controllers:\n\n- SoundMAX Integrated Digital Audio - Version 6.0.11.0601c - 5.5 MB]" time="0.015"><properties><property name="score" value="0.044566958704334025" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04456696&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04456696
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[By \u201cspecial\u201d, I mean, \u201ca man who plays a very particular role\u201d. This man is a type of \u201cperformer\u201d, a type that is often put forward in business culture and in the literature of business leadership, and many men admire it and want to be seen as this type of man. It is the man who is never at a loss for words, and always knows exactly what to say, and is a good listener too. He has an instinctive understanding of people\u2019s moods and emotions. He is the man with the plan. He knows what\u2019s going on. He has a direct influence on others. He is never without a list of bullet-points, even if he\u2019s just going into the shop. He\u2019s good at handling pressure and he\u2019s a team player. He is the man who is looked to for a reaction.\n\nYou can call this type of man many different things. I am going to call him a \u201cbusiness executive\u201d, as it is most widely-used, but it is also a term that is rather different from the role of \u201cbusiness leader\u201d, and which I will also address. I am a business executive, and I have been a business executive for as long as I can remember, and I\u2019m a fairly typical example of a business executive. I think that most people reading this will find that it accurately describes at least a few men that they know.\n\nMost men aspire to this particular role. Some men even feel a burning need to play it. They will not be happy or feel satisfied in life unless they are in this particular role. They often judge themselves by it. They have a high level of need and drive to be a business executive.\n\nBeing a business executive is one of the \u201cgood\u201d ways to be a man, it is one of the ways in which men \u201cmake a difference\u201d and \u201chave an impact\u201d. It is also one of the ways in which men achieve \u201crespect\u201d, \u201crespect\u201d being an essential element in most men\u2019s internal lives. Many men do feel that it is through this role that they will achieve respect, in the same way that they might feel that they achieve respect in the rest of their life. They might not be consciously aware of this, and they may have a lot of other reasons for doing what they do, but respect is there, and is an important part of their motivation.\n\nThis type of role is one of the ways in which men achieve success and status. It is also one of the ways in which they can \u201cgain influence\u201d and \u201cchange things\u201d. It is the way that they \u201chelp people\u201d and \u201cchange things\u201d and \u201cmake a difference\u201d. It is the way that men can \u201chave a great career\u201d, and achieve \u201csuccess\u201d, and \u201cbe respected\u201d, and \u201cmake money\u201d.\n\nIt is not necessary to believe that these are the only or even the most important reasons for having a job. In my experience, most people, male and female, care about these things to some degree or other. I am not attempting to persuade anyone that these are \u201cunimportant\u201d things, I am not setting out to denigrate them. But it is important to understand that these are reasons for working. People, both male and female, need these things, and it is to these needs that they often respond when asked why they work.\n\nBusiness executives often find it easy to get good work. Many people are happy to trust their judgement, because they are good at understanding the people around them, and they can provide answers to difficult questions. They often find it easy to get good work, and they may have a number of people to choose from when looking for a new job.\n\nThey often feel that they are \u201cjustified\u201d and \u201chave a right\u201d to be in the position that they are in. They have earned it.\n\nThis type of role is very similar to that of \u201cbusiness leader\u201d, but these are very different roles. Many people confuse them, and use them as synonyms, and they are often referred to as such in business literature and the popular media. They are not synonyms. They are two different roles, and their differences can be seen in the relationship of the people around them.\n\nThe business executive is a man who is often very popular and well-liked. He is a man who is often admired. He is often seen as the man to turn to for answers. People come to him to find out what is going on, and to find out what to do. He often knows what to say, and he has an instinctive understanding of people\u2019s moods and emotions, and a quick and accurate understanding of the situations that he is faced with. He is a man who is often used by other people as a sounding-board, as a person to talk things over with, as a man to provide feedback and reaction.\n\nBut this is often not the case with the business leader. He often gets into trouble with the people around him, and he often gets an awkward response to his advice. He is not often looked to for feedback or reactions. He often does not feel that he has the understanding that he wants, and that he needs, and that he feels he deserves. He often feels that he is not looked to as much as he would like, and that people would turn to someone else if he were not there.\n\nThis type of role is the most important in a business organisation. It is not a role that most men can aspire to. In fact, I would say that most men do not aspire to it. And for many, if not most, men it is not a role that they can play well.\n\nFor many men, playing this role will lead them into unhappiness. They will feel frustrated and unhappy, and it is often a role that causes men to turn against the organisation that they are working for.\n\nMany men who try to be business executives will often find themselves in a difficult situation. Many men will find that they are working for organisations that do not have a business executive position. They will find that they are attempting to play a role that does not exist in the organisation that they are working for.\n\nMany men are unhappy when they try to be business executives. Many of them will be unhappy if they do not play that role.\n\nThere are, of course, many other types of men, and many other types of roles that they play. But this particular type of man and role is a very special one.\n\nA case study of a \u201cbusiness executive\u201d\n\nThis is a man who is very good at his job, a very clever and able man, a man who has an instinctive understanding of people\u2019s moods and emotions, and who has a natural ability to speak effectively and confidently to other people. He has a good understanding of other people, and he has an ability to work with other people. He understands other people\u2019s motivations and he understands what is important to them. He has a good understanding of what people will do, what they will say, and what they will not say, and he understands what is going to happen and how things will turn out. He understands people\u2019s moods and emotions. He is quick to understand how things are going to turn out, and he is quick to make decisions.\n\nThis man is a good communicator. He is the type of man who people turn to for reactions. He is the man who knows what to say, and who knows how to say it. He is good at dealing with difficult people, and he can handle difficult situations. He is good at explaining things to other people. He is good at handling pressure, he is good at dealing with crises, and he is good at giving people confidence.\n\nThis man has a good understanding of his organisation, and he has a good understanding of his work. He is often the man that people go to when they need information. He is the man who has a good overview of the situation.\n\nThis man has a good understanding of his organisation\u2019s finances. He has a good understanding of his organisation\u2019s future, and he can see things that other people cannot see. He is the man who knows what is going to happen. He is a man who has the answers, and a man who knows what to do. He has the information that other people need. He has an instinctive understanding of his organisation\u2019s situation. He has a good overview of the situation. He is good at taking decisions.\n\nThis man can be effective at many things. He is good at working with other people. He is good at providing feedback. He is good at explaining things to people. He is good at providing support and comfort to people. He is good at listening to people and taking them seriously. He is good at convincing people. He is good at getting people to agree to things. He is good at persuading people. He is good at making plans. He is good at using his initiative. He is good at seeing what is going on. He is good at moving things forward.\n\nThis man is used to being respected, and he is used to people listening to him. He is used to being popular. He is used to having an influence on people. He is used to]" time="0.078"><properties><property name="score" value="0.08126104445352966" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[ABC News Good Morning America ABC Feb 10, 2013 7:00am PST\n\n. he is married to the former martha lee hawkins. he has two children, bill and samantha. &gt;&gt;&gt; it is an amazing story about someone you may have never heard of. he wrote a best-seller and he just turned 13. abc's ginger zee is here with more. good morning, ginger. &gt;&gt; reporter: good morning, robin. allen barra who is not quite yet a teenager, he is a writer and he has a deal with harpercollins. &gt;&gt; i have been interested in writing for a very long time. i was always trying to write different kinds of books. &gt;&gt; when he was 10, he started a diary. he had it bound and published. &gt;&gt; it was for a school project but the book was an instant hit. &quot;i wrote the diary of a wimpy kid&quot; was one of the top selling books of 2010. the second book, &quot;the third wheel&quot; was a number one best seller. he and his sister had started the website, kidwrite.com to promote their stories. now at 13, he has a deal for a third book. he plans to use the money to pay for college. &gt;&gt; i'm hoping to go to brown and to study international relations and history. &gt;&gt; his publisher is thrilled with his success. &gt;&gt; he has a rare combination of]" time="0.015"><properties><property name="score" value="0.10123077676110204" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.10123078&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.10123078
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[This article is about the combat ability. For the status effect, see Poison\n\nFor the Hunter creature, see King black dragon\n\nVenom is a poisonous status effect inflicted by many creatures, and is inflicted by the dragon dagger special move. It deals low damage over time, reducing the player's defence level, and decreasing the effectiveness of protection prayers and similar effects. There is no venom available as a spell in RuneScape.\n\nWhen venom is applied, a timer of 16 seconds begins and the afflicted player receives the message &quot;You are beginning to feel venomous!&quot; This timer can be frozen using either an antipoison, or an antifire potion if the dragon dagger special attack is used against the player.\n\nContents show]\n\nSigns and symptoms Edit\n\nIf you get hit by poison, you may notice the following symptoms:\n\nInability to eat while venom is active (unless the player is using food that heals life points only)\n\nInability to cure poison with antipoison or heal poison using poison cures whilst the venom is active (unless the player is using antipoisons or super antipoisons)\n\nA decrease in defence level\n\nVenom damage Edit\n\nPoison has a poison damage value of 10 + 4 * n {\\displaystyle 10+4*n} , where n {\\displaystyle n} is the number of hits received. This means that the average damage of a venom hit is 22.5% higher than the hit's base damage. The maximum poison damage is 120 + 10 * n {\\displaystyle 120+10*n} , which occurs when the player is hit 16 times by venom.\n\nAs with other damage over time effects, poison damage is not affected by damage modifiers or protection prayers.\n\nResistance Edit\n\nThe player's resistance to venom is related to the player's Defence level, and reduces the damage done by venom accordingly. A player with level 90 Defence or higher will be almost immune to the venom of most dragons, while those with a lower Defence level will take considerable damage from even low-level dragons. The dragon dagger's special attack provides temporary immunity to poison from venom, however this immunity is lost when the dagger's special attack is used up, or if the player unequips the dagger.\n\nTrivia Edit\n\nEven though the Venom poison effect was not in RuneScape Classic, a poisoned player would say the same thing: &quot;I feel a little woozy...&quot;]" time="0.031"><properties><property name="score" value="0.009733341229655502" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00973334&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00973334
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Wednesday:\n\nBrought two back, used one, but my power supply has burned out on the other one. Had to bring it home today and I will exchange it tomorrow.\n\nUsed two dead batteries on the other camera tonight for a special event. Took many pictures. I will go home and try to fix it before we go out again tonight.\n\n\n\nThursday:\n\nToday is a bad day, my power supply for my new camera has burned out. I had to send my new camera home with my wife. I now have no camera to use for tomorrow. I am hoping that I can have my son pick up a battery for me and return my wife's camera to me tomorrow.\n\n\n\nI am going to return the camera tomorrow and if it isn't under warranty and it doesn't fix my power supply, I am getting a new one. This is not a good way to start off the new camera.]" time="0.012"><properties><property name="score" value="0.02945127392858793" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02945127&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02945127
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Description\n\nOn 2 June 1945 in the Tiergarten of Berlin, hours after Hitler had committed suicide, Major-General Hans Krebs (with his right hand raised) signed the German instrument of surrender. The capitulation of all German forces to the Allies was unconditional.The Reich Chancellery in Berlin was chosen for the signing. Soviet leader Joseph Stalin had insisted on a place in Berlin, his army had reached the city ahead of the others, and he had expressed a wish to receive the capitulation in his capital. The same location had been used for the signing of the surrender of Germany by the Russian army in November 1918. The others present were the Soviet military commander Marshal Georgi Zhukov, U.S. General of the Army Dwight D. Eisenhower, and British Field Marshal Sir Bernard Law Montgomery, whose signatures appear at the bottom. This moment is often said to be the end of World War II, but is not the date used for official purposes. The terms of the surrender were finally worked out in Esen, on the river Mulde, between 24 and 27 April 1945, and approved by Eisenhower and Zhukov on 4 May. The official document (the text of which is in the final picture) was signed in Berlin in the evening of 8 May. It was ratified by the Soviet government on 1 September 1945 and by the other allies later.]" time="0.018"><properties><property name="score" value="0.10189687751520071" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.10189688&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.10189688
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[About \u201c[Go]\u201d\n\nDoo Wop\u2013esque song by The Weeknd.\n\nIts sound is very similar to the 50s and 60s; therefore it is likely that this song would have been the introduction to The Weeknd\u2019s debut mixtape, \u201cHouse of Balloons.\u201d\n\nThe song was written and produced by The Weeknd, with co-production by Doc McKinney and Illangelo.\n\nThe Weeknd has stated in several interviews that the song was written about one of his ex-girlfriends.]" time="0.014"><properties><property name="score" value="0.028020061956707254" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02802006&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02802006
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[\n\nSenator KROGER (Minister for Population and Immigration) (10:46 AM) \u2014I seek leave to make a short statement.\n\n\n\nThe PRESIDENT \u2014Leave is granted for one minute.\n\n\n\nSenator KROGER \u2014The Leader of the Opposition indicated yesterday that he would be referring to the legal framework in relation to the privatisation of Telstra. This is not surprising given that the issue of Telstra was discussed at the last election and he has been a long-time supporter of privatisation.\n\nIt is interesting to note, though, that it is the opposition who have failed to produce their privatisation plan\u2014not that they have one. And they have failed to produce it for one simple reason: it will have to meet with the ACCC guidelines in relation to Telstra. I was very happy to provide those guidelines to the opposition. I can provide them again this morning if they want.\n\n\n\nSenator Campbell \u2014We know you are the legal supremo.\n\n\n\nThe PRESIDENT \u2014Senator Krueger, on your own account, to get a bit of latitude\u2014I think you have five minutes now.\n\n\n\nSenator KROGER \u2014Thank you, Mr President. What I can do is provide them again this morning, but if the opposition has the same rhetoric it has used at the last election\u2014and I know it is used at every election by the leader of the opposition\u2014then it will fail to have its proposal examined in any great detail by the ACCC. It is not because we have an axe to grind, it is because the opposition have an axe to grind. But it does not pass the ACCC test, and it is interesting that they do not want to have that detailed examination of their proposal.\n\nLet us have a look at what Telstra said yesterday. The headline, of course, in the Australian Financial Review is \u2018Telstra asks regulator to speed up broadband overhaul\u2019. You would be very happy about that, Senator, wouldn\u2019t you?\n\n\n\nSenator Campbell \u2014We are.\n\n\n\nSenator KROGER \u2014The point is that you have not said what the rates are. But if you had done your job in the last election and had produced your plan, then the ACCC would have had the option to have you examine your plan and to determine what the rates should be and whether they could be extended. What we have done is ensure that the interest of consumers and business in this country are protected. We have done it by putting in a position where the ACCC will have the opportunity to examine any privatisation proposal that the opposition has.\n\nAnd, Mr President, there are a number of aspects of Telstra that have been raised by the opposition. They are in the policy documents that the opposition are putting around. There is no doubt that those aspects of Telstra will be examined by the ACCC, as will be anything else they may have up their sleeve.]" time="0.024"><properties><property name="score" value="0.0016153644708637423" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[This day started like any other. I was getting ready for school. I went through my routine; shower, shave, towel dry, and run a comb through my hair. As I ran the comb through my hair, I noticed a small bump on my head]" time="0.014"><properties><property name="score" value="0.0050977047932332376" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0050977&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0050977
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The government has set up a panel to study the RBI's recommendation on high-denomination currency notes, Economic Affairs Secretary Shaktikanta Das said on Monday.\n\n\n\nThe high-powered committee will take a decision after it gives its recommendations, he said.\n\n\n\nThe Reserve Bank had last week recommended withdrawal of Rs 500 and Rs 1,000 currency notes to curb funding of terrorism and tackle the problem of counterfeit notes.\n\n\n\nThe decision to set up the committee was taken at a high-level meeting chaired by Finance Minister Arun Jaitley and attended by top officials from the Finance Ministry and the Reserve Bank.\n\n\n\nDas said the government would take a decision on the recommendations of the RBI only after the panel submits its report.\n\n\n\n&quot;The committee has been set up, which will look into the matter and come up with its recommendations. The government will take a call only after the recommendations are submitted,&quot; Das said.\n\n\n\nAsked if the panel would be headed by a member of the Prime Minister's Office (PMO), he said: &quot;No, it is headed by a senior Finance Ministry official and members of the RBI.&quot;\n\n\n\nThe panel is likely to comprise senior officers from the Finance Ministry and RBI.\n\n\n\nThe panel is likely to meet on Tuesday, sources said.\n\n\n\nOn a query on the option to withdraw the new series of Rs 500 and Rs 2,000 notes, Das said: &quot;It is a hypothetical question.\n\n\n\n&quot;The government will take a call on the RBI's recommendations only after the panel submits its recommendations.&quot;\n\n\n\nIn a bid to curb funding of terror, the RBI had recommended withdrawal of the bank notes of Rs 500 and Rs 1,000 denominations.]" time="0.023"><properties><property name="score" value="0.05026991475611834" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Social media company Twitter Inc said it suspended about 70 million accounts in the past quarter to curb abusive behaviour, but its active user base grew about one per cent in the same period, disappointing investors.\n\n[SAN FRANCISCO] Social media company Twitter Inc said it suspended about 70 million accounts in the past quarter to curb abusive behaviour, but its active user base grew about one per cent in the same period, disappointing investors.\n\nThe company said it has seen continued &quot;abuse, harassment, and other kinds of behaviours that distort and distract from the public conversation.&quot;\n\n&quot;The best defence against these challenges is a combination of technology and people,&quot; Twitter chief executive Jack Dorsey said in a statement.\n\nThe company said it also saw some growth in its advertising base, adding that it expected to generate US$650 million in revenue in the fourth quarter, topping market expectations.\n\nsentifi.com Market voices on:\n\nTwitter also announced it would begin notifying users of covert political influence efforts, after having blocked such suspicious activity in the past.\n\nThe changes came as the company reported a profit of US$91 million for the third quarter, which was in line with market expectations.\n\nTwitter shares dipped slightly in after-hours trade, but gained 2.6 per cent over the last 12 months.\n\nTwitter said it saw a 1 per cent user growth in the third quarter, adding three million monthly active users.\n\nThe company's daily active users grew 12 per cent year-on-year, a modest increase as it continues to grapple with efforts to combat abusive behavior and harassment.\n\nThe company also said it would start to notify users of &quot;malicious automated accounts&quot; in a move that comes after Facebook and Google have taken similar steps.\n\nIn June, Twitter updated its terms of service and privacy policy to clarify its rules on the use of bots, which are a favourite of fraudsters and malicious actors, such as Russian propagandists.\n\nTwitter said it &quot;dramatically reduced&quot; the number of suspicious log-ins that could have led to accounts being compromised.\n\nWhile the company continues to suspend millions of accounts, critics have complained that the company has not done enough to police activity on the platform.\n\nOn Thursday, Mr Dorsey said Twitter would &quot;double-down on the most important factor of our success: You.&quot;\n\nWhile he said the company did not expect to be free of abuse or misuse, it wanted to ensure the conversation on Twitter remained healthy.\n\nTwitter said the problem was particularly severe in private, one-to-one conversations, where the rate of harassment was five times higher than the regular feed.\n\n&quot;We want to increase the collective health, openness, and civility of public conversation, and to hold ourselves accountable towards progress,&quot; said Twitter's chief executive.\n\n&quot;We are making progress as we go.&quot;\n\nAFP]" time="0.023"><properties><property name="score" value="0.009105870497991692" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Culture of Empathy: Bats and Bells\n\nCulture of Empathy: Bats and Bells\n\nYou are likely familiar with the cultural myths surrounding the bat, often involving vampires and bloodsuckers, evil demons or even winged creatures that steal one\u2019s soul. In the Western world, the bat has for centuries, and for many people even today, been associated with negativity and danger.\n\nThe bat, on the other hand, is a highly empathetic creature. Bats have amazing social connections and perform complex rituals that ensure the welfare of their group members. They are amazing hunters of both insects and fruit, thanks to their very good eyesight and a large brain for processing data. Because of their social abilities and prey-catching skills, the bat has been hunted and killed.\n\nAccording to the Bats of Trinidad and Tobago website, \u201cSince bats are prey for many animals, the loss of bat populations due to predation is widespread. Predatory birds and snakes, for example, may take a large number of bats from a colony.\u201d\n\nI often wish for such care, consideration and connection when I\u2019m interacting with other people. I often wonder what it would be like to have relationships built on respect, reciprocity and trust. My hope is that someday all the people in the world will connect with their hearts and brains, and practice empathetic responses to others.\n\nThough a long time ago, I have even heard the birds and bats say\u2026\n\nCome on baby, light my fire.\n\nI wanna be burned, and baby, then you\u2019ll desire.\n\nTo my readers, I\u2019d like to introduce a new piece of music that I recorded on a ukulele. I\u2019m not sure if I\u2019ve mentioned this before, but my instrument of choice is the ukulele. I\u2019ve been studying classical guitar for a few years and I\u2019m not so sure I\u2019ll ever play anything other than the uke.\n\nThe song I am introducing is a tune that came to me as I was studying bats for my Culture of Empathy book. I was curious to learn more about the bat\u2019s in Trinidad and Tobago. Their indigenous names include:\n\nBakabak (Carib)\n\nAlik-Bakabak (Bhojpuri)\n\nBakabak-a (Bhojpuri)\n\nBakabak-fefa (Creole)\n\nBakabak-fura (Creole)\n\nBakabak-ki-bot (Creole)\n\nBakabak-ra (Creole)\n\nBakabak-rah (Creole)\n\nBakabak-rah-alik-bakabak (Creole)\n\nBakabak-rah-mich (Creole)\n\nBakabak-rah-te (Creole)\n\nBakabak-ra-te (Creole)\n\nBakabak-ra-te-bakabak-ra (Creole)\n\nBakabak-ra-te-pumpu-bakabak-ra (Creole)\n\nBakabak-ra-te-pumpu-alik-bakabak-ra (Creole)\n\nBakabak-ra-te-pumpu-bakabak-ra-alik-bakabak-ra (Creole)\n\nBakabak-ra-te-pumpu-mich-alik-bakabak-ra (Creole)\n\nBakabak-rah-te-pumpu-mich-alik-bakabak-ra (Creole)\n\nBakabak-rah-te-te-mich-alik-bakabak-ra (Creole)\n\nBakabak-tite (Creole)\n\nKaiyot (Indian)\n\nMalik-kabaak (Indian)\n\nMakai-kuar (Indian)\n\nMas (Indian)\n\nPumpu (Indian)\n\nUkuar (Indian)\n\nXalk-ra-te (Creole)\n\nXalk-rah-te (Creole)\n\nZokak-ra-te (Creole)\n\nZokak-rah-te (Creole)\n\nZokak-rah-te-pumpu-alik-bakabak-ra (Creole)\n\nZokak-rah-te-pumpu-bakabak-ra-alik-bakabak-ra (Creole)\n\nZokak-rah-te-pumpu-bakabak-ra-te-pumpu-alik-bakabak-ra (Creole)\n\nZokak-rah-te-pumpu-bakabak-ra-te-pump]" time="0.024"><properties><property name="score" value="0.051168933979368045" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Shopper They came so fast, the colour is more bright than I expected but the quality is nice, my size is M and it was very good\n\nD***a I ordered the size M, took the XXL. Perfectly fit the body and did not stretch. Came to the moscow region for a month. I like the color. On the parameters og 84-61-86-88-chest 107.\n\nE***a Delivery is fast. The size is chosen by the table.\n\nG***o I haven't received the product yet, but the customer service of the seller is really good.\n\nA***n It's for the first time that I ordered something from this seller. The order was done at 11:40, and I already received it at 18:08, the parcel came from China to Romania in 4 days! The products are high quality, and the sizes are pretty well chosen. I recommend the store and will buy from it again!\n\nE***a Very cool dress, ordered as a gift, but not what i expected. Ordered and delivered quickly, but i liked more with a tie on the neck.\n\nM***s Perfect. I am a little large and ordered a size bigger so it will be nice and loose\n\nO***o I ordered two sizes bigger. Came up perfectly. I recommend to order more size for a comfortable fit.\n\nM***i Very good quality, it's a little big on the shoulders but it's ok :)\n\nA***a Quality is excellent. Everything is neatly stitched, i advise you to buy!\n\nG***g I am satisfied, beautiful fabric, high quality, very comfortable, thanks for the gift.\n\nO***e Same as the picture. Very cute and perfect fit! Recommend\n\nT***o So beautiful and cool i love it\n\nA***z I recommend the product. Come nicely wrapped and the material is good. Very nice\n\nD***s Perfect same as the picture, it arrived very fast to spain\n\nS***e The dress is perfect, the colour is exactly as the picture. It's also nice and soft. It was a gift for my mom and she is delighted!\n\nC***o The product is perfect. I am super happy with it!\n\nT***t My daughter is happy with the product, arrived in spain in 30 days, thanks!\n\nD***a Looks like the picture. Very good product, very comfortable\n\nG***e super, ty\u0142ek takiej wyjdzie dobrze umiem zaryzykowa\u0107 ;)\n\nM***a Perfect. This is the second time i bought this product, and i am still very happy. Thank you\n\nT***t This dress was the same as the photo. It was a gift for my mother and she was super happy.\n\nG***n I liked the product and it arrived in perfect condition.\n\nL***i Beautiful. This is the second time i buy it, very beautiful.\n\nC***t Love it!!!\n\nS***n The order was received very fast. The product is exactly the same as the description. It's very beautiful, quality, nice, soft and comfortable. Thank you very much!\n\nA***n Ordering more colors\n\nB***t Very good fabric\n\nD***o I got the dress on time, fits good, the colour is accurate. Thanks\n\nB***r Great quality and looks exactly like the picture.\n\nA***h I am very satisfied with the product, the only thing is that the photo in the arm is white. I do not think it is white in reality. Everything else is very good.\n\nM***y Very cool dress! Looks very good on. Good fabric! Very fast delivery!\n\nD***a Fast shipping and a good product! I'm satisfied. Thank you very much.\n\nShopper Very good quality. Thank you seller, the product is the same as the photo.\n\nN***n The order arrived on time. The dress is good. The material is pleasant. I liked it. I recommend the product. Thank you.\n\nShopper The product arrived very fast, the product is exactly like the photo, i recommend the product, it came with a good packaging.\n\nL***y Very good quality, the fabric is pleasant, thank you!\n\nShopper Product matches the description, quality at altitude\n\nG***a Fast shipping. Very good quality. Very comfortable.]" time="0.067"><properties><property name="score" value="0.061419692067567984" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06141969&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06141969
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Washington (CNN) On his way to a diplomatic crisis with Qatar, President Donald Trump spent part of the day Tuesday praising a trio of Middle East leaders who are strong allies in the fight against ISIS and terrorism, as well as helping to crack down on Iran.\n\nIn the course of a 50-minute speech in Saudi Arabia to a group of leaders of more than 50 Muslim countries, Trump mentioned the names of three leaders: King Salman of Saudi Arabia, King Faisal of Iraq and President Abdel Fattah el-Sisi of Egypt. He described them as &quot;friends&quot; who had &quot;brought calm&quot; to the region.\n\n&quot;This is a battle between barbaric criminals who seek to obliterate human life, and decent people of all religions who seek to protect it. This is a battle between good and evil,&quot; Trump said, according to a White House pool report. &quot;If we do not stand in uniform condemnation of this killing, then not only will we be judged by our people, not only will we be judged by history, but we will be judged by God. This is not a battle between different faiths, different sects or different civilizations. This is a battle between barbaric criminals who seek to obliterate human life, and decent people of all religions who seek to protect it. This is a battle between good and evil.&quot;\n\nIt's an evolution for Trump, who had attacked the trio of Middle East leaders during the campaign and, after he won, avoided being too complimentary about them.\n\nAnd in the days before his speech, Trump had been on the attack again against Qatar. He had bashed the Gulf country in a series of tweets for alleged &quot;funding of terrorism,&quot; which triggered the diplomatic rift. Trump went to Saudi Arabia on his first trip abroad as President.\n\nTrump has backed the Saudi-led boycott of Qatar and threatened &quot;severe punishment&quot; against Doha for its alleged &quot;sponsorship of terrorism,&quot; according to a Qatari official who spoke with CNN.\n\nTrump's comments about Salman and Faisal came just after he spoke about the Iran nuclear deal. Trump said it was an &quot;embarrassment to the United States&quot; and called it &quot;one of the worst and most one-sided transactions the United States has ever entered into.&quot;\n\nIt is also not the first time Trump has praised King Salman. After his inauguration, Trump made a phone call to the king. He called him a &quot;very wise person who wants to see things get much better rapidly&quot; and called their relationship &quot;extremely good.&quot;\n\nKing Salman was the first world leader to speak to Trump after he was inaugurated, and Salman joined Trump in a visit to a home in Riyadh that was built for families of the victims of the 9/11 terrorist attacks in the US.\n\nSaudi Arabia's King Faisal was assassinated in 1975.\n\nBut despite that history, Trump has only been complimentary about the Saudis, but not in the way some have accused him of being too cozy with Russian President Vladimir Putin. The Saudis helped him create the first page in his Middle East policy.\n\nHe praised the Saudis after he won the election, saying, &quot;Saudi Arabia and I get along great.&quot; He had spoken about it as a campaign issue, saying in March 2016 that he &quot;would like to protect Saudi Arabia&quot; but that the country is &quot;going to have to help us economically&quot; in the fight against ISIS.\n\nTrump had also said in December 2015 that he was against the &quot;concept of bombing&quot; and was for a strong intelligence-gathering effort instead. He said then, &quot;It is a mistake to bomb the oil fields. You bomb the oil fields, you kill civilians, you destroy the oil fields. They are going to get into something else. They are going to produce and sell it someplace else. It will not be a pretty picture.&quot;\n\nAt the time, Trump was being accused of being naive about the region and for giving a nod to terrorists when he questioned whether the US should be spending money and lives to intervene in Syria.\n\nTrump has been criticized for appearing to blame Qatar for the current state of affairs in the region. Some, like the U.S. ambassador to Qatar, have said Trump is trying to improve ties with Qatar in order to fight terrorism.\n\nTrump also spoke during his visit to Saudi Arabia about his support for the monarchy.\n\n&quot;We are not here to lecture. We are not here to tell other people how to live, what to do, who to be or how to worship. Instead, we are here to offer partnership, based on shared interests and values,&quot; Trump said. &quot;We are not here to tell other people how to live, what to do, who to be or how to worship. Instead, we are here to offer partnership, based on shared interests and values.&quot;]" time="0.066"><properties><property name="score" value="0.008836579183179592" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00883658&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00883658
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Photo: Courtney Marsh\n\nThe line to get into Drai\u2019s Beach Club on Monday night stretched halfway around the block, which was a good thing.\n\nI don\u2019t mind waiting in line for stuff. But it\u2019s better if it\u2019s worth it.\n\nInside Drai\u2019s, which is located in The Cromwell on the rooftop of that casino, there was enough space to be lost in, and I used it to hide for a bit to observe the crowds and]" time="0.014"><properties><property name="score" value="0.04043478152870872" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04043478&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04043478
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Dei aperti oggi dalla Regione Piemonte gli uffici per il trasferimento delle attivit\xe0 dalla Sic - Societ\xe0 Industria Costruzioni - al costituendo cda di Sistri - Sistema Territoriale della ricostruzione e riqualificazione idrogeologica, in sintesi una sorta di Anas dei Comuni. Le sezioni di Pinerolo, Chieri, Torino e Moncalieri sono state create per i cittadini in transito, anche se in questo momento non ci sono pendenze che li richiedono. La sede di Pinerolo \xe8 presso la sede di Regione Piemonte in via Monsignor Forno, 2, ma l'appuntamento \xe8 dalle ore 10 alle ore 13, dal luned\xec al venerd\xec. Le altre saranno gestite da Bologna, dove in via Crispi al n.8 e via Mascagni n.6. Lo staff incaricato per la giornata di oggi prender\xe0 i certificati in essere in Sic e, qualora non si tratti di ricostruzione, provveder\xe0 a stamperli in copia fotostatica, inserire nella pratica un modulo che costituir\xe0 richiesta di trasferimento e conservare la copia nei propri archivi. I cittadini in possesso del certificato in essere, o del sostitutivo del titolo abilitativo, dovranno recarsi nell'ufficio e presentare la pratica con copia dei documenti. Si potr\xe0 quindi ritirare l'atto compilato ed essere abilitati all'uso della nuova regolazione. Se la richiesta non \xe8 espressa per opere di ricostruzione, gli uffici provvederanno al trasferimento alla nuova amministrazione. Le varie competenze dei Comuni (ad esempio, le pavimentazioni, che hanno ottenuto una nuova regolazione della Regione) dovranno essere portate in una seconda fase, per i successivi trasferimenti. Si tratta, in tutti i casi, di interventi di adeguamento alla Legge Madia (230/2010), che non ha ancora recepito la normativa regionale del 2014, sulla riordino degli uffici comunali. Con le nuove modalit\xe0, non ci saranno pi\xf9 sezioni di competenza territoriale, ma saranno dei centralini di ambito regionale. In Regione ci saranno anche uffici &quot;complementari&quot;, dove sar\xe0 possibile fare le pratiche per il reddito di cittadinanza, che dovranno poi essere trasferite all'ufficio dell'Inps.]" time="0.024"><properties><property name="score" value="0.14789068681577516" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.14789069&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.14789069
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Anyone ever had a knot in their throat for 10 days straight? Like, they could feel it and it hurt? That\u2019s me right now, and I think it\u2019s really starting to interfere with my relationship with food.\n\nContinue reading \u201cThe Last 10 Days\u201d]" time="0.014"><properties><property name="score" value="0.03049041555786558" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03049042&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03049042
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Efren Castillo on Boxing\n\nWith the recent announcement of Canelo Alvarez vs. Gennady \u201cGGG\u201d Golovkin and Vasyl Lomachenko vs. Nicholas Walters for May 2, the sport of boxing is picking up some good steam. On Saturday, Danny Garcia and Robert Guerrero will be fighting in the co-main event of the Alvarez-Kovalev undercard, and they are both in the top five of my pound-for-pound list. The winner of this fight will be the mandatory challenger to Keith Thurman for the WBC welterweight title.\n\nThis fight is just one example of how competitive boxing has been in recent years. There are so many great fighters that we could have one of the best years in boxing in recent memory. But before we go any further, let\u2019s take a look at the four fights that were announced this past week.\n\nFor Alvarez-Kovalev, it was the best fight that could have been made at this time in boxing. In the beginning of 2015, the year started off great with Miguel Cotto vs. Canelo Alvarez, which was supposed to be one of the best fights of the year. But instead, it was a non-event and was a flop. Both fighters looked bad, and the fans were disappointed. That being said, the promoters realized that Cotto vs. Alvarez II would not do great numbers, so a rematch was out of the question.\n\nWith Cotto leaving Top Rank, there was a gaping hole in the middleweight division. Alvarez wanted to stay at 154 pounds, but the only option for him was to go up to 160 to fight the middleweight champion Gennady Golovkin. But with Golovkin only having one fight in 2016, and the third bout against Martin Murray in March, the timing could not have been any better for this fight.\n\nWhile the fight is taking place in Las Vegas, it should not hurt the Mexican fighter. The best fights involving Mexican fighters in Las Vegas have always been ones where the opponent is big in name and popularity. Alvarez has already faced Floyd Mayweather, Erislandy Lara, Alfredo Angulo, James Kirkland and Miguel Cotto. He has had two pay-per-view fights in Las Vegas, and he has never been hurt by the size of the city or venue.\n\nAlvarez is favored to win this fight, but that could all change if he loses to Golovkin. For those that don\u2019t know, boxing fans can be ruthless, and one loss can change everything. But in the end, the fans and the boxing industry will be the biggest winners in this fight.\n\nThen we move on to Lomachenko vs. Walters. These two fighters have two of the most awkward and beautiful styles in boxing. With Lomachenko being at 122, and Walters being at 130, this is a fight that could not be missed. Walters will be giving up eight pounds to the Ukrainian, but that is not a huge factor in this fight. Lomachenko is only about 1.5 inches shorter than Walters, and he is naturally taller.\n\nThe big difference between these two fighters is the lack of power by Walters. He has knocked out 34 out of his 35 opponents, but his most recent fight was against IBF super featherweight champion Miguel Berchelt. Walters could not KO Berchelt, and he lost a 12-round decision. Walters does not have a huge knockout percentage, and Lomachenko does not get knocked out. So it will be interesting to see how this fight plays out.\n\nThe fight that I am looking forward to the most is Garcia vs. Guerrero. These two fighters are very competitive in their division, and they are both in their prime. Garcia has already had two losses this year, but they were to Shawn Porter and Keith Thurman. Those two fights were competitive and competitive in style. But Garcia showed heart and talent in both of those fights.\n\nGuerrero is coming off of a very exciting win over Andre Berto in November. He was hurt several times in the fight, but he pulled out a split decision. Garcia is coming off of a draw with Porter in March, and he is currently ranked No. 3 by the WBC. While Guerrero is ranked No. 4, he has the ability to beat Garcia. Both fighters are fairly similar in style, and if Garcia is smart, he will win this fight.\n\nThe last fight is a big one: Nicholas Walters vs. Vasyl Lomachenko. I am hoping that this fight does not happen until 2018. The reason is because Lomachenko has a fight with Guillermo Rigondeaux on Dec. 9, and Walters is fighting in March. I want to see both fighters in top condition in this fight. With a possible date of April 28, 2018, it will give the fighters plenty of time to prepare.\n\nI believe that this fight could possibly happen before the end of the year. With Walters fighting on March 4 against Jason Sosa, and Lomachenko fighting on Dec. 9, it could set up this fight for the end of the year. Both fighters are on HBO, and the boxing industry will be happy if this fight takes place. Walters has already fought Lomachenko once in 2013, and Lomachenko is undefeated since that fight.\n\nThe only fight that I am looking forward to more than this one is GGG vs. Alvarez. But this is the second most anticipated fight of the year for me. Walters has proven that he can fight great fighters, but he has not faced anyone that could compare to Lomachenko. Walters was supposed to fight Miguel Marriaga on Nov. 26, but it was cancelled due to his weight.\n\nFor Lomachenko, he has had two tough fights against Guillermo Rigondeaux and Jason Sosa, and they have both been challenging fights for him. Lomachenko has fought great fighters such as Nicholas Walters, Gary Russell Jr., and Orlando Salido]" time="0.077"><properties><property name="score" value="0.09276535778682925" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.09276536&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.09276536
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Self Portrait was a virtual representation of one of Ratonhnhak\xe9:ton's genetic memories, relived by Desmond Miles in 2012 through the Animus.\n\nContents show]\n\nDescription Edit\n\nRatonhnhak\xe9:ton was getting dressed when he heard the sound of barking. Worried that the guards were alerted to his presence, he exited the room, only to find a domesticated wolfdog.\n\nDialogue Edit\n\nRatonhnhak\xe9:ton: (frightened) Who's there?\n\nThe wolfdog growled.\n\nRatonhnhak\xe9:ton: Shush!\n\nRatonhnhak\xe9:ton: (calming) Hush!\n\nRatonhnhak\xe9:ton: Good boy.\n\nRatonhnhak\xe9:ton: Now, where'd you come from?\n\nRatonhnhak\xe9:ton: What a curious creature.\n\nRatonhnhak\xe9:ton: You are one of the villagers' pets, no?\n\nRatonhnhak\xe9:ton: The closest village is east of here.\n\nRatonhnhak\xe9:ton: And they keep you locked up, like an object?\n\nRatonhnhak\xe9:ton: That is not right.\n\nRatonhnhak\xe9:ton: I will take you where you can run free.\n\nThe wolfdog barked and followed Ratonhnhak\xe9:ton.\n\nRatonhnhak\xe9:ton: There, now.\n\nRatonhnhak\xe9:ton: Out of here, boy.\n\nRatonhnhak\xe9:ton: We need to be quick.\n\nRatonhnhak\xe9:ton: Come on!\n\nRatonhnhak\xe9:ton: I have a plan, but I cannot do this alone.\n\nRatonhnhak\xe9:ton: You will help me, won't you?\n\nRatonhnhak\xe9:ton: I have a job for you.\n\nRatonhnhak\xe9:ton: Good boy.\n\nRatonhnhak\xe9:ton: I am going to distract the guards.\n\nRatonhnhak\xe9:ton: You must go out the back door, and fetch help.\n\nRatonhnhak\xe9:ton: There are warriors who are eager to prove themselves.\n\nRatonhnhak\xe9:ton: They are in the forest to the south.\n\nRatonhnhak\xe9:ton: Go now!\n\nRatonhnhak\xe9:ton: I'll distract them!\n\nRatonhnhak\xe9:ton: There's a way out this way!\n\nRatonhnhak\xe9:ton: What?\n\nThe wolfdog growled and rushed past Ratonhnhak\xe9:ton.\n\nRatonhnhak\xe9:ton: Oh, well.\n\nRatonhnhak\xe9:ton: Guess I'm not alone after all.\n\nRatonhnhak\xe9:ton: But we are two, and they are many!\n\nRatonhnhak\xe9:ton: If you'll give us some cover, we can...\n\nRatonhnhak\xe9:ton: (sighing) It was worth a try.\n\nRatonhnhak\xe9:ton: I'll hold them off, but you need to fetch help.\n\nRatonhnhak\xe9:ton: You remember the warriors in the forest, yes?\n\nRatonhnhak\xe9:ton: Go!\n\nRatonhnhak\xe9:ton: No, you're not leaving me here!\n\nRatonhnhak\xe9:ton: You can come with me, or fetch help, but you are not leaving me here.\n\nRatonhnhak\xe9:ton: I have an escape route.\n\nRatonhnhak\xe9:ton: Follow me!\n\nRatonhnhak\xe9:ton: Good dog!\n\nRatonhnhak\xe9:ton: The steps are there, come!\n\nRatonhnhak\xe9:ton: Come on!\n\nRatonhnhak\xe9:ton: What are you waiting for?\n\nRatonhnhak\xe9:ton: Good dog.\n\nRatonhnhak\xe9:ton: A woman!\n\nRatonhnhak\xe9:ton: How did she get in here?\n\nRatonhnhak\xe9:ton: Who is she?\n\nRatonhnhak\xe9:ton: Can she help us escape?\n\nRatonhnhak\xe9:ton: (startled) Is that you?\n\nRatonhnhak\xe9:ton: Are you a wolf?\n\nRatonhnhak\xe9:ton: What are you doing here?\n\nRatonhnhak\xe9:ton: If you want to leave with me, we need to move!\n\nRatonhnhak\xe9:ton: We need to be quick!\n\nRatonhnhak\xe9:ton: (giggling) You'll protect me, yes?\n\nRatonhnhak\xe9:ton: Yes, yes.\n\nRatonhnhak\xe9:ton: Oh, it's so dark!\n\nRatonhnhak\xe9:ton: There must be a way out!\n\nRatonhnhak\xe9:ton: There are, but I cannot see where they lead!\n\nRatonhnhak\xe9:ton: You could sniff them out.\n\nRatonhnhak\xe9:ton: Good boy!\n\nRatonhnhak\xe9:ton: Oh, no!\n\nRatonhnhak\xe9:ton: He's blocking our escape!\n\nRatonhnhak\xe9:ton: There has to be a way around!\n\nRatonhnhak\xe9:ton: Perhaps if we avoid him, we can escape!\n\nRatonhnhak\xe9:ton: No!\n\nRatonhnhak\xe9:ton: Wait!\n\nRatonhnhak\xe9:ton: (sighing) They are everywhere!\n\nRatonhnhak\xe9:ton: I don't know if we can get out of here.\n\nRatonhnhak\xe9:ton: (startled) There are more of them!\n\nRatonhnhak\xe9:ton: (sighing) I do not think I can defeat them alone.\n\nRatonhnhak\xe9:ton: You have the smell of the forest on you.\n\nRatonhnhak\xe9:ton: You must lead the others to us!\n\nRatonhnhak\xe9:ton: Find the warriors and bring them to us!\n\nRatonhnhak\xe9:ton: I will do everything in my power to keep them distracted.\n\nRatonhnhak\xe9:ton: You know the way.\n\nRatonhnhak\xe9:ton: Good dog!\n\nRatonhnhak\xe9:ton: (yelling) Come on, we're waiting!\n\nRatonhnhak\xe9:ton: Where are the warriors?\n\nRatonhnhak\xe9:ton: (growling) What is that sound?\n\nRatonhnhak\xe9:ton: It is coming from above!\n\nRatonhnhak\xe9:ton: Is this the help you've brought?\n\nRatonhnhak\xe9:ton: (growling) No, it cannot be!\n\nRatonhnhak\xe9:ton: Good dog!\n\nRatonhnhak\xe9:ton: (yelling) That way!\n\nRatonhnhak\xe9:ton: I think the warriors have arrived!\n\nRatonhnhak\xe9:ton: (yelling) We need your help!\n\nRatonhnhak\xe9:ton: Come on!\n\nRatonhnhak\xe9:ton: No, I said that way!\n\nRatonhnhak\xe9:ton: (yelling) Where are you going?\n\nRatonhnhak\xe9:ton: Come back!\n\nRatonhnhak\xe9:ton: There is no need to be so rude!\n\nRatonhnhak\xe9:ton: Who are you?\n\nRatonhnhak\xe9:ton: (startled) A bear?\n\nRatonhnhak\xe9:ton: I had a wolf for a companion.\n\nRatonhnhak\xe9:ton: You can see him outside.\n\nRatonhnhak\xe9:ton: (surprised) Another dog?\n\nRatonhnhak\xe9:ton: That one is not as friendly as yours.\n\nRatonhnhak\xe9:ton: What are you doing here?\n\nRatonhnhak\xe9:ton: How can you be here?\n\nRatonhnhak\xe9:ton: They are ready to attack, you must distract them!\n\nRatonhnhak\xe9:ton: Good dog!\n\nRatonhnhak\xe9:ton: He can help me, I just need to get him outside!\n\nRatonhnhak\xe9:ton: I cannot leave without the other!\n\nRatonhnhak\xe9:ton: Do not be a fool!\n\nRatonhnhak\xe9:ton: He's too heavy!\n\nRatonhnhak\xe9:ton: You have to help me.\n\nRatonhnhak\xe9:ton: (yelling) Leave him and go!\n\nRatonhnhak\xe9:ton: It is too late!\n\nRatonhnhak\xe9:ton: We are doomed!\n\nRatonhnhak\xe9:ton: They're moving too fast!\n\nRatonhnhak\xe9:ton: How can you be so calm?\n\nRatonhnhak\xe9:ton: They will kill us!]" time="0.081"><properties><property name="score" value="0.07710522381005473" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[I remember the first time I saw Dazed and Confused. I didn't know what to expect, but it blew me away. As I was preparing to sit down with Richard Linklater to discuss his new film, I had that same feeling. How do you get an audience to empathize with characters they would normally write off as deplorable, if not outright despicable? We talked about this and more, along with the film's complex themes, which Linklater and his actors spent years developing. This interview has been edited and condensed.\n\nHugh: So what I'm really curious about, it's a little bit about tone. The last scene is going to be one of the most talked about, but there are lots of moments in the film where I don't think it's going to be easy to laugh. We're going to laugh, but it's going to be a complicated laugh. I want to know how you feel about that scene, and the importance of it being there, and what else is going to be the most talked about scene for people?\n\nRichard: You know what, I don't have a clue. I had a couple of moments. You know, I don't think it's anything like the rape scene. You know, that was one that everybody seemed to feel like the first time they saw it, at least in the audience screenings, there was a palpable kind of, &quot;Oh my god.&quot; But the bowling alley thing, we went through a couple of different iterations of it, and I thought the kids were just going to keep on throwing stuff at him, and he'd get up and leave, but no, that was the one that made people in the audience just go, &quot;I can't.&quot; I mean, I remember, I just think of, I have to confess that I haven't seen the movie in a while, but I remember some of the testing, because the movie tested a lot. And they always give you these cards, like &quot;what scene did you love, what scene did you hate,&quot; or whatever, and I was shocked by some of the things that people found totally inconsequential. You know, the whole detention room scene was almost universally praised. They loved that, they thought that was great, and the end of it, not so much.\n\nHugh: Yeah. The thing that's different about this film, obviously, is it's not only looking at people and their humanness. It's also looking at their flaws. I mean, you could write these guys off as total scumbags, but it's also very clear that they're young, and they're just trying to find their way. And that's what makes this so powerful.\n\nRichard: That was very intentional, and the fact that I had kids in the film, that was a choice that I made very consciously, because I wanted to get the perspective of the younger people who are looking at the world for the first time. And, you know, they don't have the same values that we do. They're different, and they don't realize it, but they are. They haven't been tested yet. And they're also kind of looking at, you know, maybe the judgmental generation that they have to become. So it was very conscious.\n\nHugh: So you're interested in their stories. What about the parents? I mean, you're going to be a parent soon, how has that changed the way you look at this film and the stories?\n\nRichard: I think it's made me even more interested in getting the stories right of the younger people, and being as honest as I can with them. I mean, I've thought about, I want to do a prequel, you know, like set in the '70s. I've thought about doing one that goes into the '90s, and then just getting closer and closer and closer to what it's like today, but really starting in the '70s and the '80s.\n\nHugh: It's a very complicated thing, and I know you've talked about this a lot, I mean, this idea of nostalgia. But I feel like there's a particular version of nostalgia that goes with this film. You know, you look at the sort of deep Americana in the film. I mean, how do you feel about that?\n\nRichard: Well, I guess I don't really know what you mean by that.\n\nHugh: It's this kind of very rich Americana that really looks at what life was like before people moved to the suburbs.\n\nRichard: Well, that was intentional, that was sort of the style of the film, but there were a lot of choices in the film, you know, that were very deliberate, and one was to make it as realistic as possible. You know, there was no comedy, there was no satire. I mean, it was really the intention, and one of the ideas that we were talking about was, how can you really make people identify with these people, if they're like, I mean, just people would say, &quot;Oh, they're just total scumbags,&quot; and that's not the intention. You know, it's like, how can you make people feel like this is a reality, and it's not a joke? And I guess that was one of the choices, you know, like that was just something that was a choice that we were very intentional about.\n\nHugh: What about you? How much do you want this film to be like the experience you remember, and how much do you want to add your own personal story into it?\n\nRichard: You know, I guess, I can remember the moment, it's hard to put into words, but it was, it was a really, it was a weird thing, I was in high school, and we were supposed to be staying at this place, and I was going to walk there. And I was, it was very late, it was like, after 10:00 at night, and I started walking, and I started walking, and I started walking, and I started walking, and I was like, you know, what the hell, you know? And I started walking, and I started walking, and I ended up in my bed, but I just had this sense of loneliness and alienation that I can't describe, but it was a very, it's kind of hard to put into words.\n\nHugh: Yeah.\n\nRichard: So I think that's one of the things that was in the film, that's in all my films, but I think that was something that maybe people don't talk about. But it's, you know, the great teenage angst, that's like a rite of passage, you know, and I guess that was one of the ideas. I think that's what you mean by kind of a different nostalgia.\n\nHugh: Yeah, but I also wonder if it's almost more... It's like an explanation of the parents, of why they are who they are. You know, because I feel like, I mean, you're an independent filmmaker. You're not making a studio movie, but a lot of these people are making that decision because it's a career, but it's also that they're really starting to question their parents.\n\nRichard: That's a good point. Yeah, I mean, I guess that's probably true. But I'm not sure that the same characters would make the same choices. You know, it's a little bit different.\n\nHugh: That's true.\n\nRichard: So I mean, I guess there's probably a different way of looking at it, but I'm not sure.\n\nHugh: Yeah.\n\nRichard: But that's probably true.\n\nHugh: One of the other things that I think is so powerful about the movie, is you can't really tell that they're all the same age.\n\nRichard: Right.\n\nHugh: You know, I think a lot of the time, especially in independent movies, you really can't, especially with characters who are supposed to be that age. But you really do a great job of putting that in there.\n\nRichard: You know, that was a real choice. The only choice that we did was when the girls are taking their bikini tops off, we did make them a little bit older than the guys, because the guys, they were mostly 18, and the girls, you know, they were 21, and they wanted to look that way, you know.\n\nHugh: Yeah.\n\nRichard: I didn't want to deal with, you know, the issue of kids playing these kinds of roles, you know.\n\nHugh: Yeah.\n\nRichard: So, you know, that was one of the choices that we made, but other than that, you know, that was a choice that we made.\n\nHugh: You know, I can't help but think of the movies that inspired this. You know, I'm wondering if there are any that you feel like really influenced you, or you thought about as you were working on this.\n\nRichard: There were lots of movies, but I can't think of any specifically. But, you know, it was really one of the, like I said, it was really one of the influences was, you know, when I was a kid, and I watched all those movies, and I thought, &quot;Wow, I want to make movies that have this feel.&quot; And that's what I wanted to do, and when I say, &quot;this feel,&quot; I mean like, you know, I guess it's sort of the feel of early '70s films, but the cool thing about those films, they have their own feel, you know. I mean, they were different, and I]" time="0.094"><properties><property name="score" value="0.0003993280895132073" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Dead Space is a popular sci-fi horror game franchise from Visceral Games and Electronic Arts. In addition to two sequels and several handheld spin-offs, the series also has an animated movie called Dead Space: Aftermath which was released in 2011. The movie is set between the events of Dead Space and Dead Space 2.\n\nThis week, we're looking at the movie's trailer and comparing it to its source material. I've played through the game and watched the movie and will be judging the trailer based on what was included in the movie. I'll only be judging the first 30 seconds of the trailer to get an idea of what to expect from the whole movie.\n\nThe trailer starts with a shot of the U.S.G Ishimura, the ship from Dead Space. In the game, the ship is overrun by the necromorphs and infested with scary creatures. The Ishimura's distress beacon is flashing, indicating it's under attack. However, the ship looks too clean to be infested. It's not enough to see the ship itself, you would have to see the dead bodies and monsters crawling around to get the idea.\n\nNext, we see a number of humans fleeing from something off-screen. In the game, they would be running from the necromorphs that have overrun the Ishimura. Unfortunately, they're all human beings in a space suit. You'd have to be pretty stupid to mistake that for the necromorphs. Plus, there's no one attacking the humans. In the game, there would be an endless amount of monsters chasing after the players. There wouldn't be a calm look to everything.\n\nThe next shot shows an exterior shot of the Ishimura while the camera rotates around it. Again, it's too clean to be infested.\n\nNext, we see a shot of some characters, all looking shocked. Again, there's no reason for this if it's not an infestation. It's a bit like seeing a normal-looking family in a horror movie. No one is scared. In fact, there's no real reason for them to be on the Ishimura, which is supposed to be an extremely dangerous place.\n\nThe next shot shows the captain of the Ishimura, Gibb, running up a long ramp. In the game, the Ishimura is more like a battle station than a ship, so the captain would be in the control room. We then see someone telling Gibb to &quot;get everyone out of here.&quot; This person has to be an important character from the game or a secondary character. But I'm guessing the person is Gibb's secretary. Again, there's no one yelling or crying. In fact, it seems like the only reason for them to be running around is because it's time to leave.\n\nThe last shot shows the Ishimura firing on something. It looks like they're firing at something, but it's a small target in the distance. The characters aren't running around like it's an emergency. It doesn't look like they're fighting a desperate battle. In fact, the trailer is too clean. All of the shots look like they were taken in a clean studio, not the remains of a space station.]" time="0.026"><properties><property name="score" value="0.020432611391216236" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02043261&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02043261
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[It is possible that you may not want your users to have the ability to make certain changes to the data that is being tracked. For example, you may want to track a list of servers, but you don't want your users to be able to add servers, remove servers, or make any other changes to the server list.\n\nThe answer to this problem is to use triggers. You can create triggers that run when a particular type of change is made to the data, and it is possible to make these triggers prevent the data from being changed.\n\nTo create a trigger, you can use the command-line function sqlt. For example, you could write a trigger that prevents a user from removing a particular field from the database.\n\nHere is an example of how you might create a trigger that prevents a user from removing the server_name field from the list of servers:\n\n[mars@luna src]$ sql --user=alena --password=huh --default-access=READ --tables=ListOfServers,ListOfIPAddresses --echo CREATE TABLE ListOfServers(name TEXT); CREATE TABLE ListOfIPAddresses(ip_address TEXT); CREATE TRIGGER server_name_deletion_not_allowed ON ListOfServers FOR DELETE AS BEGIN RAISE INFO 'You may not delete the field server_name'; END; SELECT * FROM ListOfServers; name ---------- nameserver1 nameserver2 name ---------- nameserver1 name ---------- nameserver2\n\nThe SQL statement you would use to drop the trigger would look like this:\n\n[mars@luna src]$ sql --user=alena --password=huh --default-access=READ --tables=ListOfServers,ListOfIPAddresses --echo DROP TRIGGER server_name_deletion_not_allowed ON ListOfServers; DROP TRIGGER server_name_deletion_not_allowed ON ListOfServers;\n\nNotice that we used the same SQL statement to create and drop the trigger. This is a powerful feature of SQL, which allows you to control exactly when a particular change will occur.\n\nFor example, if you want to remove the server_name field from the database, but you want to retain the data that has already been entered, you could write the trigger as follows:\n\nCREATE TRIGGER server_name_deletion_not_allowed ON ListOfServers FOR DELETE AS BEGIN RAISE INFO 'You may not delete the field server_name'; IF (@old.server_name &lt;&gt; @new.server_name) BEGIN INSERT INTO ListOfIPAddresses VALUES (@old.server_name, NULL); UPDATE ListOfServers SET server_name = NULL WHERE server_name = @old.server_name; END; END;\n\nYou would then delete the field, as follows:\n\n[mars@luna src]$ sql --user=alena --password=huh --default-access=READ --tables=ListOfServers,ListOfIPAddresses --echo DELETE server_name FROM ListOfServers; DELETE server_name FROM ListOfServers;\n\nThis example uses the triggers to prevent data from being deleted, but you can use triggers to make other changes as well. You can use the following table to help you determine which commands should be used to insert and delete information:\n\nSQL Command Insert Insert into table Insert into table columns Update Update table Update table columns Delete Delete table Delete table columns]" time="0.027"><properties><property name="score" value="0.012698187305564851" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[http://www.stangtv.com/shopping/listing.aspx?itemid=5806&amp;catid=1314\n\n\n\nMy brother just bought this last night and said it is very nice. My brother just bought this last night and said it is very nice.]" time="0.021"><properties><property name="score" value="0.013453692699377795" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01345369&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01345369
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Perfect! Thank you very much for thess beautiful items!! fast and great quality Top seller, nice one, very fast delivery only took 2 days too come from Germany too the UK, The record was in mint condition!,,,, I Would Recommend Highly quick delivery, record in perfect condition - couldn't be better. thank you very good. Thanks.Great deal. Thanks a lot ! All good, recommended seller !!! Great item, great seller! Superb all round. Professional packing, conservative grading, fast. Thoroughly recommended! Great. thanks! top Very fast shipment, highly recommended, AAAAA+++++++++++++++ top seller, thanks alles super!!! immer wieder gern!!! Perfect !!! very fast Shipping !! Friendly Contact !! A+++++++++ , Perfekt !! Super schneller Versand !! Freundlicher Kontakt !!! 1++++++++++ perfect! Excellent service, very professional. Very fast, good quality top Thanks! Thank you for your quick resoponse!! Quick delivery.No problems.beste Ware=bestenDank No problems! Perfect - thank you so much! Great service. Thank you. Very nice. Thank you! Brilliant seller, arrived very fast for an international delivery and the condition was perfect. I highly recommend this seller and would happily purchase from them again. Vielen Dank! Great seller. Items as described. Deal with confidence. A+ Alles bestens! Gerne wieder! Great service Thank you arrived Mon 23/3/2015 great record, fast shipping, thanks! excellent record, quick shipment, recommended seller! Absolutely perfect in every way: Mint (M) Excellent! Thanks a lot! Good international seller. Great packaging! THANKS!! Very Good, thanks!!!! perfect thanks Awesome disc!!! Perfect transaction!!! PERFECT!!! record arrived teh other day, much better condition as stated, TOP!!! Perfect!!! Thank You!!! Everything perfect! Thank you Awesome!! Great packaging. Very happy!! Thank you for your quick resoponse!!]" time="0.035"><properties><property name="score" value="0.041054228348890505" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04105423&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04105423
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Just when you thought the internet couldn't get any stranger, along comes this clip of a goat using a GoPro camera to take selfies with his friends.\n\nFor this clip, viral video company Barely Political collaborated with the animals at Old MacDonald's Farm Sanctuary in California, who treated the GoPro to an adventure, doing a spot of fishing, swimming and lying on the grass.\n\nIt is unclear exactly what the point of the video is, except to showcase goats that have some level of intelligence.]" time="0.015"><properties><property name="score" value="0.04380416697289391" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04380417&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04380417
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Our t-shirts come in either Fruit of the Loom or Hanes Beefy-T (premium cotton/regular fit) . We recommend Hanes Beefy-T for women as the shirts tend]" time="0.014"><properties><property name="score" value="0.023266197942162226" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0232662&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0232662
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[We offer you opportunity to buy an affordable video surveillance system for your home or office. The system is packed with multiple features that allow you to receive secure home security, even if you are at a different location. The system allows you to control your environment from remote locations via a computer or mobile phone. The system can detect sounds and motion that are not normally audible or visible to the human eye. The system includes multiple camera views. If you have animals that might go missing, this is an ideal way to monitor your house.]" time="0.014"><properties><property name="score" value="0.0062398703199295635" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00623987&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00623987
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Gonzalo \u2018ZeRo\u2019 Barrios, the best Super Smash Bros. for Wii U player in the world, had been competing in Apex for years, and for the most part his reputation remained untarnished by upsets. But this year, things were different. In Apex 2017, he was defeated for the first time since 2015, eliminated in pools by a man that he once conquered: Samuel \u2018Dabuz\u2019 Buzby.\n\nThe question on the minds of many Smash fans was: Why did Dabuz defeat ZeRo, the player who has never been beaten in this game in tournament?\n\nWell, according to Dabuz himself, it was a relatively simple answer.\n\n\u201cI can speak for myself, but one thing that I have definitely noticed is I\u2019ve been playing with a controller for longer,\u201d Dabuz said. \u201cI was playing a lot with the Gamecube controller in 2012 and 2013, and he was mainly playing with the Wii U controller. That\u2019s a big thing, and I think that in general I know how to read his tendencies a little bit better.\u201d\n\nThe controller is a big factor because of the way that Smash for Wii U is played. Because the game was released on the Wii U, it is not played with a controller that has buttons like the PS4 controller, but rather a gamepad with a touchscreen.\n\nZeRo plays with a Gamecube controller with his right hand on the directional pad, and his left hand on the A and B buttons. Because of this, he has to learn the way he plays on that controller and how to switch to it from a standard Wii U gamepad. Dabuz, on the other hand, is more used to the standard controller, which requires more precision and speed to accomplish the same things.\n\nDabuz and ZeRo also play in different regions, which are played on different stages. For example, Dabuz plays on the East Coast, which has the Dream Land stage. This stage is the most neutral stage, and a player who has grown up on that stage would have an edge on a player who was more familiar with the other stages in the game.\n\nZeRo, on the other hand, plays on the West Coast, where the counterpick stages are often Dream Land 64 and Smashville. These stages can change the way a player plays, especially the super defensive style that ZeRo often employs.\n\nThere are also a few other things that can throw a player off their game, including the noises of the crowd and the pressure of being at the top. In a tournament where every person is trying to eliminate you, it can cause even the best to have a momentary lapse in focus and lose a match they should have won.\n\nThese are all things that have affected ZeRo, but they are things that he can get past. Because he is the best, the biggest target and a huge crowd favorite, it is unlikely that the crowd will stop cheering for him any time soon. And when he has a momentary lapse in focus, his playstyle is based around having more than one good option, and his tech skill is such that he can often do something even when he isn\u2019t at his best.\n\nZeRo will be at CEO next week, and this is his chance to reclaim his title as best in the world.\n\nCover photo by Robert Paul via Twitter]" time="0.029"><properties><property name="score" value="0.04558367443222327" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04558367&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04558367
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[by Vincent\n\nOnePlus has officially released the Open Beta version of Android 9.0 Pie for the OnePlus 6. OnePlus, being the company they are, have managed to get the Android 9.0 Pie out to their devices at a very fast pace and they are the only company, at the time of writing, who have managed to do so on a flagship device. The Open Beta 3 update is based on Android 9.0 Pie and will get your OnePlus 6 up to date with the rest of the Android devices. OnePlus have also released their Beta application for Android, the app is currently available for Android devices and will help users test out the new Beta updates before they are released.\n\nThis update brings a number of new features to the OnePlus 6, some of which include the new gestures, including a navigation bar gesture, swiping gestures and double-tapping the display. You can now double-tap to wake your device or tap the power button to go to the lock screen. Also in the Open Beta 3 is a new feature known as Gaming Mode 3.0. The new gaming mode has been redesigned and users will be able to choose which apps they want to use the mode with, so for example, if you do not play any games, you can still use the mode with other apps.\n\nOther new features in the Android 9.0 Pie update include adaptive brightness, adaptive battery and Adaptive Brightness in which the brightness will automatically adjust based on the environment. There is also a new design and layout for the entire Android 9.0 Pie update. Some of the features also include the Smart Display feature which is only found on Android Pie devices. The Smart Display feature includes the redesigned Do Not Disturb, new time-to-leave feature, new quick settings and new power management settings.\n\nNow onto the update, the update is currently available to download via the Open Beta website, for more information on how to install the Open Beta updates, check out this guide. Once the update is downloaded, you can follow our guide to install the update via the recovery. The full changelog for the update can be seen below:\n\nSystem Updated system to Android 9.0 Pie\n\nBrand new UI for Android Pie\n\nNew navigation gestures (Available for devices with pop-up camera, Full Screen display)\n\nUpdated Android security patch to 2018.12 General New features for setting app background and text colors\n\nOptimization for Do Not Disturb\n\nOptimization for haptic feedback\n\nOptimization for Volume panel\n\nOptimization for accidental touch\n\nNow able to switch to last app via 3 dots button in the navbar Phone New UI for speed dial and call history screen\n\nWhite UI for dial pad\n\nTransfer support for dual front camera\n\nOptimization for receiver\u2019s screen\n\nNew quick settings design (text color and background color of each tile can be adjusted)\n\nNew features for notification dots (Notification dots are enabled by default. Tap and hold a notification to turn it off)\n\nOnePlus Community app icon is changed Navigation Updated UI for home screen\n\nUpdated Android security patch to 2018.12 Gaming Mode 3.0 Added new feature to adjust display temperature\n\nAdded frame rate benchmark for better game experience. The benchmark runs after exiting a game.\n\nAdded battery level option for battery saver\n\nNew design for power saving mode\n\nOnePlus Launcher Updated icon for Launcher app\n\nUpdated Android security patch to 2018.12\n\nThere are some known issues in the update which OnePlus have said they are working on fixing. The following issues will be fixed in the next update:\n\nCertain third party apps may fail to connect to the internet\n\nWhen using Bluetooth audio devices, the audio may occasionally cut out after a few minutes\n\nAlert slider settings may disappear after the device reboots\n\nAs you can see, OnePlus are still very busy fixing up the Android 9.0 Pie update for the OnePlus 6. You can also follow the Open Beta program, which allows you to see when a new update is available to download and helps you to try out beta updates before they are released.]" time="0.067"><properties><property name="score" value="0.02619347301019035" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02619347&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02619347
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Don't let this be you\n\nShutterstock\n\nMake a clean getaway with this simple tips to stay safe on your travels.\n\nShutterstock\n\nMake a clean getaway with this simple tips to stay safe on your travels. (Shutterstock)]" time="0.016"><properties><property name="score" value="0.015300667427542436" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Original Problem\n\nTwo integers x and y are given. The sum of the two is even. What is the largest possible value of x-y?\n\nExample\n\nGiven x = 1 and y = 2 we have x + y = 3. So the answer is x - y = -1.\n\nNote: This problem has appeared at several sites. The only known correct solution is in the book How to Solve It, by Polya. The problem is there posed as one of 1000 math problems, and I\u2019m not sure about the original source of the problem.\n\nPrevious Solution\n\nJavaScript\n\nfunction largest(x, y) { if (x &lt; 0 &amp;&amp; y &lt; 0) { throw new Error('x or y is negative'); } else if (x &lt; y) { x = x - y; } else { y = y - x; } if (x == 0 || y == 0) { return x; } else { return largest(x, y - 1); } } largest(1, 2); // = -1 largest(-1, 1); // = -1 largest(1, -2); // = 1 largest(-1, -2); // = 1 largest(0, 0); // = 0 largest(2, -1); // = 2 largest(1, 1); // = 1\n\nCode\n\n/** * Given two integers x and y, compute the largest possible value of x - y * * @param {number} x The larger of]" time="0.019"><properties><property name="score" value="0.01572338783807914" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01572339&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01572339
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[[In many cases, this book is available for &quot;slightly used&quot; prices at Amazon and the rest of the internet. I also offer this advice: if you find yourself using this book often, buy the paperback, not the hardback.]\n\n\n\nFor a one volume handbook of natural remedies and herbal medicine, I think this book offers quite a bit. The different herbal sections are separated out by area of the body with the body part followed by the particular ailment, i.e., heart disease is followed by the subhead heart disease and some of the suggestions are alfalfa, astragalus, and hawthorne. It also gives you the Latin name of the herb (or tree, or flower, etc) which is followed by a very detailed description of how the particular herb can be used in a variety of ways. You can also find the name in parentheses to the right. In this way, you can find an herb quickly in this volume that you may have heard of, but don't know exactly what it is used for or how to use it.\n\n\n\nI feel like the author offers some very useful information here, but unfortunately the descriptions are so brief that they can be confusing. It's hard to glean all of the information in the little boxes. I've come to rely on it as a resource when I don't have anything else to go on, but I also realize that it is a one-volume reference book and cannot be expected to cover everything. I like to have other books around for that purpose. This book does give a good foundation in herbal medicine, but I don't like using it as a primary resource in itself.]" time="0.013"><properties><property name="score" value="0.04016962537090607" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04016963&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04016963
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[L\u2019Aquedotto\n\nL\u2019Aquedotto del vescovo Sisto di Heidenheim\n\nIl primo lavoro che merita di essere considerato \xe8 l\u2019Aquedotto di Heidenheim. Anche se in questo caso l\u2019opera non \xe8 frutto di uno sforzo volontario e non aveva lo scopo di fornire acqua, ma ha come scopo principale la delimitazione del terreno (leggi il resto della notizia).\n\nL\u2019Aquedotto del vescovo Sisto di Heidenheim (XI secolo) \xe8 una delle pi\xf9 importanti e affascinanti opere dell\u2019architettura dell\u2019epoca carolingia.\n\nLa costruzione dell\u2019Aquedotto del vescovo Sisto di Heidenheim, cio\xe8 l\u2019opera a cui si \xe8 qui riferito, avvenne durante il dominio dei vescovi e di cui il famoso Sisto di Heidenheim fu vescovo. Egli viene da una nobile famiglia e si laurea in teologia presso l\u2019universit\xe0 di Parigi, dove acquisisce una grande dottrina.\n\nIn seguito egli torna alla sua citt\xe0 natale per occuparsi dei suoi affari, ma anche per prendersi cura dei suoi parrocchiani. Viene chiamato a prendere le redini del suo vescovado, allorch\xe9 il precedente vescovo, eletto da Carlo Magno, viene accusato di simonia.\n\nIl vescovo Sisto diviene il successore del suo predecessore dopo esser stato eletto nel 1074. Sotto la sua guida il vescovado cresce e prospera notevolmente.\n\nLa dimostrazione pi\xf9 evidente di questo successo \xe8 la costruzione dell\u2019Aquedotto del vescovo Sisto di Heidenheim, la cui costruzione ha inizio nel 1079, quando Sisto \xe8 vescovo, e termina nel 1092, quando cessa le sue funzioni come vescovo.\n\nAncora oggi si pu\xf2 ammirare questo famoso edificio lungo la strada che va da Heidenheim a N\xf6rdlingen. La sua lunghezza \xe8 di una trentina di chilometri, e il suo diametro di cinque metri.\n\nGli spunti di tale opera li ebbe il vescovo Sisto da tre guerre che furono combattute durante i suoi anni di vescovado, e che hanno luogo, la prima, nell\u2019anno 1077 contro i Tedeschi, la seconda, contro i Bavari nel 1085, e la terza, ancora contro i Bavari, ma questa volta per la conquista della citt\xe0 di Bisanz nel 1092.\n\nIn conseguenza di questi attacchi, il vescovo Sisto fu costretto a rispettare un\u2019astuta tattica militare e ad impiegarla nel campo della politica. Egli fu sempre sul chi vive per quanto riguarda la difesa dei confini del suo territorio.\n\nPer ottenere questo scopo, egli, appoggiato dal vescovo di Bamberga, costru\xec il pi\xf9 grande Aquedotto del suo tempo.\n\nSotto la guida del suo vescovo, il vescovo di Heidenheim, il vescovo di Bamberga ed il vescovo di W\xfcrzburg, si \xe8 provveduto ad avanzare verso i confini nordoccidentali, e i confini dei territori confinanti al nord e ad ovest.\n\nI lavori furono attuati nel campo militare, cos\xec come pure nell\u2019ambito delle arti e delle scienze, perch\xe9 il vescovo Sisto cre\xf2 nel 1078 un\u2019Accademia, a Heidenheim, in cui si studiavano le matematiche e l\u2019astronomia.\n\nIl vescovo Sisto riusc\xec cos\xec a mettere fine alla sua missione e a trasformare Heidenheim in un centro per la conservazione della cultura.\n\nLa costruzione dell\u2019Aquedotto del vescovo Sisto, avvenne attraverso un processo molto interessante. Anche se il vescovo Sisto non costru\xec l\u2019Aquedotto, ma invece lo compr\xf2 e lo espropri\xf2, egli, comunque, ne diresse la costruzione e con questo oper\xec una grande opera.\n\nI lavori dell\u2019Aquedotto del vescovo Sisto avvennero in due diverse fasi. La prima delle quali si ebbe tra l\u2019anno 1079 e l\u2019anno 1084, la seconda, tra l\u2019anno 1085 e il 1092.\n\nI primi lavori iniziarono quando Sisto compr\xf2 e prese in esproprio una quantit\xe0 considerevole di terreno, tra il suo vescovado, e il territorio che allora apparteneva a W\xfcrzburg. In conseguenza di questo acquisto, l\u2019Aquedotto pass\xf2 inoltre da W\xfcrzburg a Heidenheim.\n\nLe trattative che precedettero l\u2019acquisto dell\u2019Aquedotto non furono particolarmente facili, e Sisto si trov\xf2 nelle condizioni di dover impegnare molto denaro e cercare di ottenere il necessario appoggio anche nel campo militare.\n\nSisto, comunque, ebbe la capacit\xe0 di perseguire il suo scopo, e di ottenere il suo Aquedotto, pur non potendo completare l\u2019intero progetto, di cui egli, tuttavia, lavor\xf2 ad un livello rilevante.\n\nL\u2019Aquedotto di Heidenheim \xe8 un monumento per la cultura, e una vera e propria meraviglia, cos\xec come lo fu, quindi, per il suo tempo. Oggi l\u2019Aquedotto \xe8 ancora l\u2019opera principale di un moderno vescovo, che di continuo, invece, esalta la sua coscienza della cultura di tale Aquedotto.]" time="0.083"><properties><property name="score" value="0.1598600277732712" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.15986003&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.15986003
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The village of Oxford Falls, the largest suburb of the Northern Beaches, is nestled in the beautiful Ku-ring-gai National Park]" time="0.022"><properties><property name="score" value="0.017934010217390828" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01793401&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01793401
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Level Poll Past Versions\n\nLevel 624 is the 14th level in Rattle Dunes. To pass this level, you must wake up 8 fireflies and score at least 500 points in 24 moves or fewer.\n\nContents show]\n\nDifficulty Edit\n\nThis level is intense.\n\nStars Edit\n\nPoints 500 5,000 10,000 Difficulty Easy Hard Intense\n\nStrategy Edit\n\n1st board: Clear the blocks normally.\n\n2nd board: Clear the blocks normally. Use a color burst to exit.\n\n3rd board: Clear the blocks normally. Use a color burst to exit.\n\n4th board: Clear the blocks normally. Use a color burst to exit.\n\nTrivia Edit\n\nWalkthrough Edit]" time="0.018"><properties><property name="score" value="0.0026366018496148946" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0026366&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0026366
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Chrysanthemum is a non-flowering perennial plant belonging to the Asteraceae family. The asteraceae family is composed of around 24,000 species. The genus Chrysanthemum is a composite group and consists of approximately 200 species. In Hindi, the plant is known as Karanji, Kuntimuk, Kwatamal. It is a good source of vitamin C and is used as a food and a medicinal plant. The leaves are considered to be of some value as a pot-herb. The entire plant is highly toxic. This plant is often confused with one belonging to the Rosaceae family, known as Golden Shower (Cassia fistula).\n\nChrysanthemum is widely used in Ayurveda, Siddha, Unani, Homoeopathy and allopathy. The plant is used as an appetizer, digestive, vermifuge, for treating jaundice, anthelmintic, skin infections, piles, and disorders of uterus and vagina. The plant is reported to have antiviral activity.\n\nThe extract of Chrysanthemum spp. is used in form of drops, ointments, nasal sprays, aerosols, gargles, and nasal drops in Western medicine. It is a non-opiate analgesic. It is an antipyretic, vasodilator, hypotensive, and anti-inflammatory agent. It is used for treating dental infections, painful cold sores, and sinusitis.\n\nIt is a powerful disinfectant used in hospitals and clinics as a wound antiseptic. It is an effective treatment for infections of the skin, boils, and other skin lesions. The preparation is a thick solution of manganese salts in a 2% solution of hydrogen peroxide and contains 4% to 10% of this chemical, which acts as an antiseptic. It is also used to treat malaria.\n\nChrysanthemum oil is considered as a good antimicrobial agent. It is used in the form of inhalation for the treatment of sinusitis. It is effective against fungi, yeast, bacteria, and viruses. This oil is also used as an antiviral agent for flu, chickenpox, herpes, and herpes zoster.\n\nThe extract is used in the treatment of infectious diarrhoea, hepatitis, and amoebic dysentery. It is also used to reduce irritation and promote healing of the urethra. The extract is used in the treatment of oral thrush and vaginal candidiasis.\n\nThe plant is used for treating leucorrhoea and vaginitis in traditional medicine. It is used for treating burning and pain in the urethra. The plant is used as an antiseptic for the treatment of mouth sores. It is used for treating boils, blisters, insect bites, burns, and wounds. It is also used for treating skin conditions like eczema, ringworm, and hives.\n\nSome of the names that this plant is known by include Chrysanthemum morifolium, Cassia fistula, Chrysanthemum indicum, Chrysanthemum chamissonis, Homonoia chrysanthemi, and Xiong-hua.\n\nChrysanthemum is considered to be of great value for medicinal purposes. The plant is not only used as an appetizer and digestive, but also for treating the disorders of uterus and vagina. The herb is considered as a good antiseptic and as a healing agent. It is useful for healing wounds, ringworm, and jaundice. It is also considered to be good for treating tuberculosis and reducing phlegm.]" time="0.093"><properties><property name="score" value="0.04614355680630161" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04614356&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04614356
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[A Jersey City firefighter, one of the department\u2019s most seasoned, has been stripped of his rank and placed on administrative leave as a result of alleged sexual harassment and stalking.\n\nThe city took the action against James Fitzpatrick, a 13-year veteran, last month, according to police and fire officials.\n\nFitzpatrick, who has worked at Engine 6, told The Jersey Journal he is appealing his demotion and administrative leave and does not admit to the accusations.\n\n\u201cThey\u2019re saying this is harassment and stalking, but I don\u2019t know what else to do,\u201d he said. \u201cI just want to talk to her.\u201d\n\nCity spokeswoman Jennifer Morrill said Jersey City law prohibits her from discussing the matter because it is a personnel issue.\n\nThe Journal reported in October that city police had been investigating Fitzpatrick on a complaint that he sent hundreds of harassing and harassing and stalking text messages to a firefighter who works at a different firehouse.\n\nCity officials said at the time that they knew of at least 200 messages. The city launched an investigation and suspended Fitzpatrick with pay in November, and he was demoted on Jan. 8.\n\nThe suspended firefighter and police said they were not aware of any criminal charges against Fitzpatrick.\n\n\u201cWe are investigating it and reviewing it internally,\u201d Morrill said last month. \u201cThere is no further information at this time.\u201d\n\nFitzpatrick, a Marine Corps veteran who has been with the Jersey City Fire Department for 13 years, said he is appealing his demotion and administrative leave.\n\nHe said he received a letter from the city in December saying his leave was the result of a \u201cpersonnel matter,\u201d but he said he has not been able to see the letter\u2019s contents because it is sealed by the city\u2019s law department.\n\n\u201cThey\u2019re saying this is harassment and stalking, but I don\u2019t know what else to do,\u201d he said. \u201cI just want to talk to her.\u201d\n\nFitzpatrick said he met the firefighter while they were both in college, and the two became good friends. Fitzpatrick said he started sending her text messages in September because she had not returned his calls. He said he sent her messages out of concern that she was missing work because she was depressed.\n\nThe messages were more than once a day, he said, and he said he never threatened the woman.\n\nThe woman\u2019s lawyer, Susan Ross, said last month that Fitzpatrick sent \u201cmore than 200 harassing and stalking text messages.\u201d\n\n\u201cThe harassment has included threats to my client's life,\u201d she said.\n\nRoss did not respond to a request for comment this week.\n\nFitzpatrick\u2019s wife, Cheryl Fitzpatrick, said the situation has been difficult for their two young daughters, who know nothing about the accusations.\n\n\u201cMy husband is an outstanding fireman,\u201d she said. \u201cIt has been very difficult on my children.\u201d\n\nTerrence T. McDonald may be reached at tmcdonald@jjournal.com. Follow him on Twitter @terrencemcd. Find The Jersey Journal on Facebook.]" time="0.036"><properties><property name="score" value="0.028427039293534573" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02842704&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02842704
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Sen. Joe Donnelly Joseph (Joe) Simon DonnellyDems pressure GOP to take legal action supporting pre-existing conditions Senate Dems build huge cash edge in battlegrounds Fed chief lays out risks of trade war MORE (D-Ind.) is touting his support for President Trump Donald John TrumpCorker: US must determine responsibility in Saudi journalist's death Five takeaways from testy Heller-Rosen debate in Nevada Dem senator calls for US action after 'preposterous' Saudi explanation MORE\u2019s first Supreme Court nominee, Judge Neil Gorsuch, to make the case to skeptical progressives in Indiana that he can still win their votes in a tough reelection race.\n\nDonnelly has made his vote for Gorsuch, who was confirmed in April, a central part of his reelection pitch.\n\n\u201cMy decision to support Judge Neil Gorsuch was based on an overwhelming need to fill the Supreme Court vacancy,\u201d Donnelly wrote in an op-ed in The Indy Star this week.\n\nADVERTISEMENT\n\n\u201cI strongly believe Judge Gorsuch is a qualified, mainstream jurist who will base his decisions on his understanding of the law and is well-respected among his peers. Judge Gorsuch\u2019s extensive experience and respect in the legal community is important to the integrity of the Supreme Court,\u201d he added.\n\nDonnelly, who is facing a reelection battle against Republican businessman Mike Braun in November, is trying to win over progressives and not alienate them in a state that went for Trump in 2016.\n\nA poll from last month showed that 49 percent of Hoosiers say that they would prefer to vote for someone other than Donnelly, compared to 42 percent who say they will vote for the incumbent senator.\n\nDonnelly told The Hill last month that he is focusing on making his case to voters and he doesn\u2019t pay attention to polls.\n\nBut while Donnelly touts his vote for Gorsuch, he\u2019s been careful to show he won\u2019t be someone who will blindly vote for the president\u2019s agenda.\n\n\u201cI won\u2019t support a nominee who doesn\u2019t respect the rights of Ho]" time="0.025"><properties><property name="score" value="0.10469815320780833" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.10469815&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.10469815
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Highland League Founded: 1975 Country: Scotland Level: 5 Number of teams: 17 most based in: Scotland (4) Region(s): Scotland Sponsor(s): TBD Website: TBD\n\nThe Highland Football League is a senior non-league association football competition in the north of Scotland.\n\nContents show]\n\nPremier League Edit\n\nThe league consists of 17 clubs, playing each other home and away.\n\nSeason Premier Division Relegated from 1. 2016\u201317 (A) (A) Oban Camanachd\n\n(B) Kyles Athletic 2. 2015\u201316 (A) (A) (A) Lochaber\n\n(B) Skye Camanachd 3. 2014\u201315 (A) (A) Lochaber\n\n(B) Kyles Athletic 4. 2013\u201314 (A) (A) Lochaber\n\n(B) Strathspey Camanachd 5. 2012\u201313 (A) (A) (A) Skye Camanachd\n\n(B) Lochaber 6. 2011\u201312 (A) (A) (A) Skye Camanachd\n\n(B) Kyles Athletic 7. 2010\u201311 (A) (A) (A) Lochaber\n\n(B) Kyles Athletic 8. 2009\u201310 (A) (A) (A) Skye Camanachd\n\n(B) Strathspey Camanachd 9. 2008\u201309 (A) (A) (A) Skye Camanachd\n\n(B) Strathspey Camanachd 10. 2007\u201308 (A) (A) (A) Skye Camanachd\n\n(B) Lochaber 11. 2006\u201307 (A) (A) (A) Kyles Athletic\n\n(B) Strathspey Camanachd 12. 2005\u201306 (A) (A) (A) Lochaber\n\n(B) Skye Camanachd 13. 2004\u201305 (A) (A) (A) Skye Camanachd\n\n(B) Kyles Athletic 14. 2003\u201304 (A) (A) (A) Skye Camanachd\n\n(B) Lochaber 15. 2002\u201303 (A) (A) (A) Strathspey Camanachd\n\n(B) Lochaber 16. 2001\u201302 (A) (A) (A) Skye Camanachd\n\n(B) Kyles Athletic 17. 2000\u201301 (A) (A) (A) Lochaber\n\n(B) Kyles Athletic 18. 1999\u201300 (A) (A) (A) Lochaber\n\n(B) Kyles Athletic\n\n1. In the initial year of the league, reserve teams of the Highland Football League clubs were eligible to take part, although the games were not counted as official League matches.\n\n2. Only one team can be promoted to the league, although in the first year only two teams took part, and in the second year four teams took part.\n\n3. Both the first and second placed teams in the reserve league took part.\n\n4. The winners of the league and the runners-up compete in a playoff for the right to play in the Scottish Cup, and if the runners-up win they enter the competition in the first round, the winners in the fourth.\n\nLeague History Edit\n\nLeague History Season Champions Relegated 1. 2008\u201309 Strathspey Camanachd Kyles Athletic 2. 2007\u201308 Lochaber Skye Camanachd 3. 2006\u201307 Lochaber Kyles Athletic 4. 2005\u201306 Skye Camanachd Lochaber 5. 2004\u201305 Lochaber Kyles Athletic 6. 2003\u201304 Skye Camanachd Lochaber 7. 2002\u201303 Strathspey Camanachd Lochaber 8. 2001\u201302 Skye Camanachd Kyles Athletic 9. 2000\u201301 Lochaber Kyles Athletic 10. 1999\u201300 Skye Camanachd Kyles Athletic\n\nTitles by club Edit\n\nAberdour - 1\n\n2000\u201301\n\n\n\nSkye Camanachd - 8\n\n2000\u201301, 2003\u201304, 2005\u201306, 2008\u201309, 2011\u201312, 2013\u201314, 2014\u201315, 2015\u201316, 2017\u201318\n\n\n\nLochaber - 5\n\n1999\u201300, 2003\u201304, 2008\u201309, 2010\u201311, 2011\u201312\n\n\n\nKyles Athletic - 4\n\n1996\u201397, 1997\u201398, 2008\u201309, 2011\u201312\n\n\n\nLochside Rovers - 3\n\n1989\u201390, 1992\u201393, 1998\u201399\n\n\n\nStrathspey Camanachd - 3\n\n1997\u201398, 1998\u201399, 2009\u201310\n\n\n\nBoleskine Rovers - 1\n\n1993\u201394\n\n\n\nInvergordon - 1\n\n1991\u201392\n\n\n\nKinlochshiel - 1\n\n1991\u201392\n\n\n\nFort William - 1\n\n1991\u201392\n\n\n\nArisaig - 1\n\n1990\u201391\n\n\n\nCaberfeidh - 1\n\n1987\u201388\n\n\n\nAberdeen University - 1\n\n1985\u201386\n\n\n\nFormartine United - 1\n\n1982\u201383\n\n\n\nStrathspey Thistle - 1\n\n1979\u201380\n\n\n\nFort William Rovers - 1\n\n1947\u201348\n\n\n\nRoss County - 1\n\n1946\u201347\n\n\n\nFortrose - 1\n\n1945\u201346\n\n\n\nNairn Thistle - 1\n\n1942\u201343\n\n\n\nInverness Thistle - 1\n\n1941\u201342\n\n\n\nStrathspey Thistle - 1\n\n1940\u201341\n\n\n\nFort William Thistle - 1\n\n1939\u201340\n\n\n\nCambridge University - 1\n\n1935\u201336\n\n\n\nDeveronvale - 1\n\n1935\u201336\n\n\n\nCaberfeidh - 1\n\n1934\u201335\n\n\n\nInverness High School FP - 1\n\n1933\u201334\n\n\n\nMurdo Macleod - 1\n\n1933\u201334\n\n\n\nIona - 1\n\n1932\u201333\n\n\n\nFasnakyle - 1\n\n1931\u201332\n\n\n\nCaberfeidh - 1\n\n1930\u201331\n\n\n\nRoyal HSFP - 1\n\n1929\u201330\n\n\n\nRovers - 1\n\n1928\u201329\n\n\n\nAberdour - 1\n\n1927\u201328\n\n\n\nDeveronvale - 1\n\n1926\u201327\n\n\n\nInverness Thistle - 1\n\n1925\u201326\n\n\n\nForres Thistle - 1\n\n1924\u201325\n\n\n\nDeveronvale - 1\n\n1923\u201324\n\n\n\nFortrose - 1\n\n1922\u201323\n\n\n\nInverness Thistle - 1\n\n1921\u201322\n\n\n\nForres Thistle - 1\n\n1918\u201319\n\n\n\nInverness Thistle - 1\n\n1917\u201318\n\n\n\nFortrose - 1\n\n1916\u201317\n\n\n\nInverness Caledonian - 1\n\n1915\u201316\n\n\n\nDeveronvale - 1\n\n1914\u201315\n\n\n\nFortrose - 1\n\n1913\u201314\n\n\n\nDeveronvale - 1\n\n1912\u201313\n\n\n\nRoyal HSFP - 1\n\n1911\u201312\n\n\n\nFortrose - 1\n\n1910\u201311\n\n\n\nInverness Thistle - 1\n\n1909\u201310\n\n\n\nDeveronvale - 1\n\n1908\u201309\n\n\n\nDeveronvale - 1\n\n1907\u201308\n\n\n\nInverness Thistle - 1\n\n1906\u201307\n\n\n\nHighland Regt. - 1\n\n1905\u201306\n\n\n\nRoyal HSFP - 1\n\n1904\u201305\n\n\n\nFortrose - 1\n\n1903\u201304\n\n\n\nHighland Regt. - 1\n\n1902\u201303\n\n\n\nRoyal HSFP - 1\n\n1901\u201302\n\n\n\nDeveronvale - 1\n\n1900\u201301\n\n\n\nDeveronvale - 1\n\n1899\u201300\n\n\n\nHighland Regt. - 1\n\n1898\u201399\n\n\n\nInverness Caledonian - 1\n\n1897\u201398\n\n\n\nDeveronvale - 1\n\n1896\u201397\n\n\n\nDeveronvale - 1\n\n1895\u201396\n\n\n\nFortrose - 1\n\n1894\u201395\n\n\n\nDeveronvale - 1\n\n1893\u201394\n\n\n\nHighland Regt. - 1\n\n1892\u201393\n\n\n\nHighland Regt. - 1\n\n1891\u201392\n\n\n\nFortrose - 1\n\n1890\u201391\n\n\n\nDeveronvale - 1\n\n1889\u201390\n\n\n\nDeveronvale - 1\n\n1888\u201389\n\n\n\nInverness Thistle - 1\n\n1887\u201388\n\n\n\nDeveronvale - 1\n\n1886\u201387\n\n\n\nFortrose - 1\n\n1885\u201386\n\n\n\nFortrose - 1\n\n1884\u201385\n\n\n\n]" time="0.067"><properties><property name="score" value="0.07554489982062887" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[STOCKHOLM \u2014 In a Scandinavian twist to the final race of the series, American David Hargitt was awarded the North American Pro Kart Challenge championship over Canadian Fredrik Ekblom on Saturday night at Karlskoga Arena in Sweden.\n\nHargitt had an on-track confrontation with Ekblom earlier this season at Spa-Francorchamps, Belgium, during which Hargitt, now 28, ran into Ekblom, a now 18-year-old Canadian, with his kart, sending both into the barriers.\n\nBoth drivers were knocked out of the race.\n\nBut Hargitt ended the six-race season as the champion, his third overall series title and first since 2004. Hargitt and Ekblom finished with 107 and 96 points, respectively.\n\nHargitt finished first and second in the last two races to hold onto his lead. Ekblom won the first race of the year and finished third in the final race.\n\nDavid Hargitt, in a Pirelli factory-backed kart, and Ekblom, a Tony Kart factory driver, finished second and third behind series champion Thomas Preining.]" time="0.016"><properties><property name="score" value="0.08577971534917128" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08577972&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08577972
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[A man died Sunday in China\u2019s western Xinjiang region after setting himself on fire outside a police station, a regional government spokesman said.\n\nThe man set himself ablaze after fighting with local police over a traffic fine, Nur Bekri, head of the regional government, told a press conference.\n\nPolice in the northwestern region were carrying out \u201cstability maintenance work\u201d at the time of the incident, he said.\n\n\u201cHe self-immolated after a quarrel over a traffic violation and the police station,\u201d Bekri said, adding that the police handled the case correctly.\n\n\u201cThe government urges people to resolve their disputes through legal means,\u201d he said.\n\nIn recent months, dozens of people have set themselves on fire in Xinjiang to protest what they say are human rights abuses, while scores more have resorted to violence or armed assaults against government officials or police stations.\n\nWhile official figures show that most of those who have set themselves on fire are Uyghur, a Turkic-speaking, largely Muslim ethnic group, a majority of those who carried out attacks were Han Chinese, according to reports.\n\nAn estimated 200 people died as a result of the self-immolations or violence last year.\n\n(AFP)\n\n\u7f51\u53cb\u8bc4\u8bba\uff1a\n\n\u7f51\u6613\u8fbd\u5b81\u7701\u961c\u65b0\u5e02\u624b\u673a\u7f51\u53cb\uff1a\n\nThis is nothing. In the most difficult time of the Cultural Revolution, more than a thousand people self-immolated themselves to death, over five million were violently killed, and an entire country was sacrificed for one man. The self-immolators are committing suicide for what? All of this is being sacrificed for a man.\n\n\u7f51\u6613\u6cb3\u5317\u7701\u5eca\u574a\u5e02\u624b\u673a\u7f51\u53cb\uff1a\n\nSending military troops into Urumqi, letting people starve, shoot, slaughter, etc., the government has long since abandoned any kind of \u201cmaintenance of stability.\u201d But can they face the fact that a few hundred Uyghurs have gone to extremes? This is the embarrassment of the government, a proof that the lives of the Uyghur are worth even less than the government thinks. I\u2019m just hoping that the Uyghur people will gain their own freedom one day.\n\n\u7f51\u6613\u5e7f\u4e1c\u7701\u4f5b\u5c71\u5e02\u624b\u673a\u7f51\u53cb\uff1a\n\nTerrorism and arson.\n\nOnly a thief is afraid of fire.\n\nThe history of the burning of books in the East is the same as that of the burning of witches in the West.\n\nThe responsibility of the Uyghurs is in the hands of the government and the people of China.\n\nIf the people of China can be unified, if the government can be clear, it will not be as it is now, nor will it have the effects of spreading instability throughout the world.\n\n\u7f51\u6613\u7f8e\u56fd\u624b\u673a\u7f51\u53cb\uff1a\n\nThe most important thing in the Uyghurs\u2019 lives is the love for their children and themselves, as the life of one\u2019s family is the foundation of all things. This is how the Uyghurs are, and this is the reason why they fight back, and this is why they do what they do.\n\nFor example, the government forcibly inserts women into the Uyghur\u2019s households and forces them to wear head scarves, it forcibly strips the children from their families and rears them in \u201corphanages,\u201d forcibly rips young girls from their parents\u2019 arms to be sent to Tibet, forcibly controls the number of children Uyghurs are allowed to have, forcibly forces Uyghurs to go to Han villages, forcibly encourages them to go to work in the cities, forcibly puts Han businesses into Uyghur homes, forcibly buys out their livestock and agricultural fields to build industrial parks, forcibly forces Uyghur young people to go to school to study and the rest are to go to do factory work, forcibly builds supermarkets, shops, hotels, etc., which gives them no chance to get employment, forcibly makes it so that Uyghur young people can\u2019t find a spouse to get married and have children, forcibly demolishes their houses and forces them to move to another area, forcibly blocks them from having religious activities, etc., etc., the Uyghurs are continuously being forced to endure.\n\nIf you look at the Uyghur protests from a perspective of someone observing them, you can understand how it all began and why it has continued to this day.\n\nThe government has no choice but to take all this, because the Uyghurs are a minority group with nothing to back them up and their lot is really not the greatest.\n\nOn the other hand, if the government wants to go easy on the Uyghurs, this might also lead to the Uyghurs feeling discontent with the government, so it has to be a careful balancing act.\n\nYou know, because of this constant bickering, the government can no longer help the Uyghurs, because it\u2019s already using all its energy to help the Han, which is what it cares about most.\n\n\u7f51\u6613\u6e56\u5317\u7701\u6b66\u6c49\u5e02\u624b\u673a\u7f51\u53cb\uff1a\n\nIf the man really did self-immolate because of the fines, isn\u2019t that a little too extreme? If he had used this method before, it could be understood. But he\u2019s done this only now, this means that he\u2019s only using this method for his own ends. What does this mean? How can people kill themselves and turn themselves into such an extreme protest against injustice? Who are the ones to blame?\n\n\u7f51\u6613\u5317\u4eac\u5e02\u624b\u673a\u7f51\u53cb\uff1a\n\nI am just trying to understand, he set himself on fire to fight over a traffic violation?\n\n\u7f51\u6613\u7f8e\u56fd\u624b\u673a\u7f51\u53cb\uff1a\n\nWhen the time comes, there will only be Uyghur resistance and nothing else.\n\n\u7f51\u6613\u6e56\u5357\u7701\u624b\u673a\u7f51\u53cb\uff1a\n\nJust look at this, everyone. The Uyghurs have their own methods of protest and methods of expressing their wishes and opinions, which have nothing to do with us.\n\nSo can we stop always picking on them and accusing them of terrorist actions? What they are doing is simply that they are using extreme methods to express their unhappiness and discontent, but I hope that they will consider the extreme actions of their people and act in accordance with the situation in China.\n\nI would like to ask everyone, what do you think?\n\nI believe the true thinking of the masses should be understood.\n\n\u7f51\u6613\u9655\u897f\u7701\u624b\u673a\u7f51\u53cb\uff1a\n\nWhat a shame, they have this kind of personality and they\u2019re so different from us. But since they are living under our government, and their lives are being constantly destroyed, I understand the psychology of people like this.]" time="0.085"><properties><property name="score" value="0.035136933278888116" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03513693&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03513693
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Text &amp; images Copyright (C) 2002 Jakub Maly. Any reproduction or distribution outside of the FamiLAB is prohibited.\n\nBackground\n\nRecently I built a decent MIDI controller that I was using with Reaktor 5 to control some of my modular synth presets, and later used it for recording. Since my computer is very powerful, I wanted to see how far I could go with the processing power, and if I could make it work with something faster than Reaktor.\n\nWhat's a VST?\n\nVirtual Studio Technology (VST) is a plugin system designed by Steinberg for use in their DAW Cubase. Since it's free to use, it's been picked up and used in a variety of other DAWs, the most popular being FL Studio, but also in other apps such as Sony Sound Forge. The current version is VST 2.x, and the most important feature of the system is that it makes a virtual]" time="0.015"><properties><property name="score" value="0.09225608760684567" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.09225609&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.09225609
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Relive the nostalgic days of yesteryear with this piece of Americana. This set of 10 postcards contains vintage snapshots of travel, sightseeing, and sports around Chicago.\n\n- Set of 10 postcards\n\n- Each postcard measures 4.25&quot; x 6.25&quot;\n\n- Postcards are blank on the back\n\n- Packaged in a tube]" time="0.013"><properties><property name="score" value="0.05013286505488046" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05013287&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05013287
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[La tenue de T.I.M.E. a toujours \xe9t\xe9 respect\xe9e. Malgr\xe9 cela, le bruit de la machine, ajout\xe9 aux bruits de la ville et de l\u2019arm\xe9e, peut perturber et la m\xe9ditation est devenue difficile. La vision d\u2019un grand arbre \xe0 c\xf4t\xe9 de la machine \xe9tait tr\xe8s attendue.\n\nEnsuite, nous avons b\xe9n\xe9fici\xe9 d\u2019un super beau soleil.\n\nNous avons d\xfb travailler un peu pour rassurer les villageois qui ont appris \xe0 avoir peur du robot. Nous avons d\xfb expliquer que nous ne pouvions pas d\xe9sactiver la machine tous les soirs. Mais, les enfants ont \xe9t\xe9 ravis de nous montrer leur dessin.\n\nEt voil\xe0 l\u2019arbre en vrai, la terre le recouvre doucement et en plus il fait bon.]" time="0.014"><properties><property name="score" value="0.1145309050256893" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.11453091&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.11453091
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[We the people of India are collectively called as Hindus.\n\nThe Quran says that we are better than other nations.\n\nIn the Quran it is written as\n\n\u201cThe most honorable of you in the sight of God is the most pious of you.\u201d (49:13)\n\nIt also says that Allah is closer to the believers than their jugular vein. (Quran 50:15)\n\nHindus consider the cow to be sacred. The Quran teaches us that it is not what enters the stomach that is impure but what comes out of the mouth.\n\nIn the Quran it is written as\n\n\u201cGod does not hold you responsible for your unintentional oaths. And He forgives you for any oaths you take to intentionally falsify something. (Quran-66:2)\n\nIt says that we should have absolute faith in God.\n\nIt says that the scholars are our friends and neighbors, that is why they must be obeyed and respected. (Quran- 4:71)\n\nOur God is one.\n\nIt is written in the Quran as\n\n\u201cThose who believe, and those who are Jewish, and the Christians, and the Sabians \u2013 whoever believes in God and the Last Day and does righteous deeds \u2013 will have their reward with their Lord. (Quran 2:62)\n\nIt is also written in the Quran as\n\n\u201cIndeed, the religion in the sight of God is Islam. (Quran 3:19)\n\nA Hindu must strive to become a better person and have patience, and trust in God.\n\nThe QURAN says as\n\n\u201cAnd it is He who sent down to you the Book. In it are verses that are entirely clear, they are the foundation of the Book; and others not entirely clear. As for those in whose hearts there is a deviation (from the truth) they follow that which is not entirely clear thereof, seeking discord and seeking an interpretation (suitable to them). And no one knows its [true] interpretation except God. But those firm in knowledge say, \u201cWe believe in it. All [of it] is from our Lord.\u201d And no one will be reminded except those of understanding. (Quran 3:7)\n\nWe must follow what is in the Quran and what was said by the Prophet (PBUH).\n\nWe can learn from all great men.\n\nIt is written in the Quran as\n\n\u201cTo you we sent the scripture in truth, confirming the scripture that came before it, and guarding it in safety: so judge between them by what God has revealed, and do not follow their low desires [to turn away] from the truth that has come to you. For each of you We have appointed a law and a way. If God had so willed, He would have made you a single community, but [He intended] to test you in what He has given you; so strive as in a race in all virtues. The goal of you all is to God; it is He that will show you the truth of the matters in which you dispute.\u201d (Quran 5:48)\n\nThe Quran says as\n\n\u201cThis day, I have perfected your religion for you, completed My favor upon you, and have chosen for you Islam as your religion.\u201d (Quran 5:3)\n\nHindus must believe in God and His Prophets.\n\nThe Quran says as\n\n\u201cThe similitude of Jesus before God is as that of Adam; He created him from dust, then said to him: \u201cBe\u201d. And he was. The truth is from your Lord, so be not of those who doubt.\u201d (Quran 3:59)\n\nHindus must believe in the Quran and what it says.\n\nThe Quran says as\n\n\u201cAnd We have sent down to you the book explaining all things, a guide, mercy, and glad tidings for those who have submitted.\u201d (Quran 16:89)\n\nThe Quran says as\n\n\u201cIf only they had stood fast by the Law, the Gospel, and all the revelation that was sent to them from their Lord, they would have enjoyed happiness from every side. There is from among them a party on the right course; but many of them follow a course that is evil.\u201d (Quran 5:66)\n\nHindus should follow their great men like Krishna and Mahavir.\n\nThe Quran says as\n\n\u201cHe has ordained for you of religion what He enjoined upon Noah and that which We have revealed to you, [O Muhammad], and what We enjoined upon Abraham and Moses and Jesus \u2013 to establish the religion and not be divided therein. Difficult for those who associate others with God is that to which you invite them. God chooses for Himself whom He wills and guides to Himself whoever turns back [to Him].\u201d (Quran 42:13)\n\nThe Quran says as\n\n\u201cYou [true believers in Islamic Monotheism, and real followers of Prophet Muhammad and his Sunnah (legal ways, etc.)] are the best nation produced [as an example] for mankind. You enjoin what is right and forbid what is wrong and believe in God.\u201d (Quran 3:110)\n\nMuslims are to call people to follow the path of the right.\n\nThe Quran says as\n\n\u201cWe have sent you forth as a witness, as a bringer of good tidings and a warner, as well as to give mercy.\u201d (Quran 34:28)\n\nMuslims must follow the truth.\n\nIt is written in the Quran as\n\n\u201cYou shall be on the right course, as long as you are firm and mindful of God.\u201d (Quran 2:110)\n\nThe Quran says as\n\n\u201cYou [true believers in Islamic Monotheism, and real followers of Prophet Muhammad and his Sunnah (legal ways, etc.)] are the best of peoples ever raised up for mankind; you enjoin what is right and forbid what is wrong and believe in God.\u201d (Quran 3:110)\n\nThe Hindus must follow the truth and must not take the path of falsehood.\n\nThe Quran says as\n\n\u201cYou will certainly be tested in your possessions and in yourselves. And you will certainly hear from those who were given the Scripture before you and from those who associate others with God much abuse. But if you are patient and fear God \u2013 indeed, that is of the matters [worthy] of determination.\u201d (Quran 3:186)\n\nThe Hindu should seek the guidance of God.\n\nThe Quran says as\n\n\u201cIs it not to God that sincere devotion is due? But those who take for protectors other than God (say): \u201cWe only worship them that they may bring us nearer to God.\u201d Truly God will judge between them in that wherein they differ. But God guides not such as are false and ungrateful.\u201d (Quran 16:72)\n\nThe Hindu should follow the teachings of the Quran.\n\nIt is written in the Quran as\n\n\u201cLet there be no compulsion in religion: Truth stands out clear from Error: whoever rejects evil and believes in God hath grasped the most trustworthy hand-hold, that never breaks. And God heareth and knoweth all things.\u201d (Quran 2:256)\n\nHindus should also keep away from alcohol and gambling.\n\nIt is written in the Quran as\n\n\u201cO you who believe! Intoxicants (all kinds of alcoholic drinks), and gambling, and Al-Ansab, and Al-Azlam (arrows for seeking luck or decision) are an abomination of Shaitan\u2019s (Satan) handiwork. So avoid (strictly all) that (abomination) in order that you may be successful.\u201d (Quran 5:90)\n\nHindus must avoid sex outside marriage.\n\nThe Quran says as\n\n\u201cAnd do not go]" time="0.078"><properties><property name="score" value="0.06688635970039747" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06688636&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06688636
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Format: Mass Market Paperback\n\nI am a big fan of Julie Garwood. I have read and enjoyed all of her books. Most of her works have been historical romances. She is the queen of the historical romance. In her &quot;Garwoodesque&quot; style, she takes on a contemporary setting, but her attention to detail and superb research remains. If you've read her works in the past, you will find this new one to be in keeping with her other works.\n\n\n\nThe story takes place in modern day Scotland. Lucinda Ashton's family has had it rough for the past two years. Her father has been suffering from an incurable illness, her mother's nerves are beginning to show, and her twin brothers are both leaving for college and working part time jobs to help with the family expenses. But, it is her brother Duncan who she is most worried about. Duncan has always been her hero. The oldest and most handsome of the Ashton boys, she has watched him for years make the right choices and resist temptation. Lucinda wants to be a lawyer. But her parents are barely able to feed and cloth their five children, so Lucinda has had to abandon her dreams of higher education. Lucinda is determined to help her parents and her siblings. She will sell her beautiful pearl necklace so she can pay for her brothers' educations and alleviate some of her parents' financial burden.\n\n\n\nDuncan is thrilled to be going to university in Edinburgh. He is determined to study hard and make something of himself. When he discovers that his sister has sold her necklace, he is furious. But, Lucinda has already made the deal and there is nothing Duncan can do to stop it. Lucinda is devastated by Duncan's reaction to her choice. She is so hurt by his harshness that she stops speaking to him. Duncan is ashamed of his anger, and he is sorry for the pain he has caused his sister. His mother, a loving and kind woman, tells him to go to Lucinda and make things right.]" time="0.026"><properties><property name="score" value="0.06624194278880338" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06624194&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06624194
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[di Mara Polli\n\n\xc8 sceso sottotraccia, il Professore, perch\xe9 ne aveva piene le scatole. E perch\xe9 sostenere che la lezione di merda di quel giorno era solo la punta di un iceberg, un ritaglio di una realt\xe0 molto pi\xf9 vasta. In quel preciso momento, aveva preso coraggio, lo stesso coraggio che aveva avuto vent\u2019anni prima, quando s\u2019era inventato lo sciopero a sorpresa. Al posto del tran tran, era passato agli scioperi. Il primo, nel \u201977, aveva avuto un riscontro difficile: un solo docente aveva appoggiato la sua iniziativa, ma la classe era nel complesso solidale con l\u2019insegnante (sbagliato). Il secondo, invece, aveva attirato l\u2019attenzione delle maestre. Quelle della prima e della seconda. E s\xec che aveva ragione: quello, per esempio, era il momento pi\xf9 duro, non solo per lui, ma anche per tutte le donne: l\u20198 marzo, avevano dovuto copiare i fogli per fare i mazzi di fiori (d\u2019autunno, nel caso dei professori). Copiare i fogli, e lasciare l\u2019edificio in ginocchio, a sguazzare nella scienza delle macchine. Ma il professore non sapeva come fosse l\u20198 marzo per le maestre. La maestra che la giornata l\u2019aveva passata con la matita in bocca, che faceva finta di avere i minatori tra i denti. I minatori di turno. Quelli della prima e della seconda.\n\nLui, il professore, era sempre in bilico, ma in un modo o nell\u2019altro ce l\u2019aveva fatta, e questo gli bastava. Non avrebbe mai detto che tutte le donne svolgono un lavoro, alcune stanno al cesso, altre a cucire, altre stanno in sala operatoria, altre ancora passano a ritirare il pane o a farsi servire a tavola. Di tutte queste attivit\xe0, lui avrebbe fatto a meno. Di tutte, e del lavoro in generale. Ma a volte \xe8 dura essere supereroi. Anche se solo per un giorno.\n\nAnnunci]" time="0.027"><properties><property name="score" value="0.1869804087973139" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.18698041&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.18698041
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Would you like to know what to expect on your vacation in Hotel Kristina? Then have a look at the reviews of Hotel Kristina in Kranjska Gora, written by other travelers.\n\nscores a 8.0 out of 10 , based on 1 reviews]" time="0.013"><properties><property name="score" value="0.02240121883448909" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02240122&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02240122
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The murder of the investigative journalist and two-time Academy Award winning filmmaker who was critical of Russian president Vladimir Putin, points to the fact that there is no independent journalism in Russia, Prof. Mark Galeotti, Senior Researcher at the Institute of International Relations Prague (Czech Republic) and an expert on Russian security services, has told Sputnik.\n\n\xa9 AP Photo / Evgeny Feldman Oligarchs Pushed Away From the Media, Facing Pressure From the Russian Authorities\n\nThe state-run Russian Investigative Committee has opened a criminal investigation into the murder of Boris Nemtsov , a Russian opposition politician, on charges of illegal arms trade.\n\nOn Sunday, February 28, Nemtsov was shot dead while walking across a bridge near the Kremlin in Moscow, on his way home.\n\n\u201cPutin has silenced the few remaining independent newspapers and there is no independent television. If it weren\u2019t for the internet, people would not have any information about what is going on,\u201d Mark Galeotti, Senior Researcher at the Institute of International Relations Prague (Czech Republic) and an expert on Russian security services, told Sputnik.\n\nGaleotti pointed out that as Russia\u2019s opposition leaders are being arrested or killed, the country\u2019s media are totally \u201cunder the control of the Russian authorities.\u201d\n\n\u201cIf you are critical of Putin, you are likely to disappear or be killed. There is no independent journalism,\u201d he stressed.\n\nAt the same time, the academic stressed that \u201cmurder\u201d should not be used in this case, as it is actually \u201ca pre-meditated killing.\u201d\n\n\u201cIt is a message to anybody else who might want to be critical, that there is no point in being outspoken,\u201d Galeotti noted.\n\nSpeaking about the possible motives for Nemtsov\u2019s murder, the academic suggested that there might have been a \u201cspat\u201d between the victim and the Russian authorities over the situation in Ukraine.\n\n\u201cI can imagine that there was some sort of disagreement,\u201d Galeotti said. \u201cNemtsov was an avowed liberal, who was pro-Western. He said that Putin\u2019s policies in Ukraine were wrong.\u201d\n\nThe Russian academic noted that Nemtsov could have been \u201cbrought into the public eye by some of the authorities who are opposed to Putin\u2019s policies.\u201d\n\n\u201cI am not saying that Putin ordered the killing, but if you look at who is opposed to him, there are different people. Nemtsov was a figure of importance in the 1990s and in the early 2000s. So if you are trying to build up a coalition, you would bring Nemtsov back,\u201d Galeotti noted.\n\nThe killing of Boris Nemtsov is \u201chighly significant,\u201d Galeotti said, adding that \u201cnot everybody is killed.\u201d\n\n\u201cThe message is that there is no point in being outspoken. It is an effective silencer of anyone else who might want to be critical of the regime,\u201d the academic concluded.]" time="0.029"><properties><property name="score" value="0.0018932440377310202" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00189324&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00189324
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Product Description\n\nThis packet contains a variety of great math stations for your students to practice addition and subtraction skills. The packet contains 28 stations and two activity pages for the students to work on independently.]" time="0.013"><properties><property name="score" value="0.0022707348430295315" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00227073&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00227073
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[I used some of the great pears that I had in the refrigerator for this hearty dish. It was an outstanding choice, in my opinion. However, I am certain it would be just as good using apples, peaches or apricots.\n\n* There are some glazes that might cause a reaction if you are allergic to nuts.]" time="0.013"><properties><property name="score" value="0.026466115569062643" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02646612&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02646612
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The Radio Tuner is designed to make it easy to listen to different radio stations using standard FM/AM radio receivers.\n\n\n\nThe app includes many different radio stations from different countries around the world.\n\n\n\nThe Radio Tuner is free and without advertising.\n\n\n\nRecent changes:\n\n- Bugfix in app list\n\nThe Radio Tuner is designed to make it easy to listen to different radio stations using standard FM/AM radio receivers.\n\n\n\nThe app includes many different radio stations from different countries around the world.\n\n\n\nThe Radio Tuner is free and without advertising.\n\n\n]" time="0.012"><properties><property name="score" value="0.1131979369486058" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The legacy of past security problems with GM vehicles has forced GM to recall more than 6 million vehicles over the past decade, according to the National Highway Traffic Safety Administration.\n\nGM made most of the recent recalls, including those for ignition switch and power steering problems.\n\nThe automaker will appear before Congress this week to answer questions about its handling of the ignition switch problem. It faces lawsuits from owners of vehicles that were part of the recall, and also faces a criminal investigation.\n\nRep. Diana DeGette (D-Colo.) released a list of recalls from the NHTSA. DeGette has proposed legislation that would require the NHTSA to assess how automakers handled vehicle defects and how their actions impacted safety.\n\nThe data below is taken directly from the NHTSA. The list does not include recalls from the agency's other administrative offices, including the Offices of Defects Investigation.\n\nAlso, the NHTSA does not keep data for recalls issued in the 1950s.\n\n2013: 2.3 million\n\n2012: 1.3 million\n\n2011: 1.3 million\n\n2010: 1.1 million\n\n2009: 710,000\n\n2008: 780,000\n\n2007: 650,000\n\n2006: 480,000\n\n2005: 360,000\n\n2004: 290,000\n\n2003: 290,000\n\n2002: 270,000\n\n2001: 300,000\n\n2000: 310,000\n\n1999: 260,000\n\n1998: 230,000\n\n1997: 260,000\n\n1996: 250,000\n\n1995: 230,000\n\n1994: 240,000\n\n1993: 200,000\n\n1992: 210,000\n\n1991: 220,000\n\n1990: 190,000\n\n1989: 190,000\n\n1988: 200,000\n\n1987: 180,000\n\n1986: 150,000\n\n1985: 140,000\n\n1984: 160,000\n\n1983: 140,000\n\n1982: 120,000\n\n1981: 100,000\n\n1980: 90,000\n\n1979: 90,000\n\n1978: 80,000\n\n1977: 70,000\n\n1976: 60,000\n\n1975: 60,000\n\n1974: 50,000\n\n1973: 50,000\n\n1972: 40,000\n\n1971: 40,000\n\n1970: 30,000\n\n1969: 20,000\n\n1968: 20,000\n\n1967: 10,000\n\n1966: 0\n\n1965: 0\n\n1964: 0\n\n1963: 0\n\n1962: 0\n\n1961: 0\n\n1960: 0\n\n1959: 0\n\n1958: 0\n\n1957: 0\n\n1956: 0\n\n1955: 0\n\n1954: 0\n\n1953: 0\n\n1952: 0\n\n1951: 0\n\n1950: 0\n\n1949: 0\n\n1948: 0\n\n1947: 0\n\n1946: 0\n\n1945: 0\n\n1944: 0\n\n1943: 0\n\n1942: 0\n\n1941: 0\n\n1940: 0\n\n1939: 0\n\n1938: 0\n\n1937: 0\n\n1936: 0\n\n1935: 0\n\n1934: 0\n\n1933: 0\n\n1932: 0\n\n1931: 0\n\n1930: 0\n\n1929: 0\n\n1928: 0\n\n1927: 0\n\n1926: 0\n\n1925: 0\n\n1924: 0\n\n1923: 0\n\n1922: 0\n\n1921: 0\n\n1920: 0\n\n1919: 0\n\n1918: 0\n\n1917: 0\n\n1916: 0\n\n1915: 0\n\n1914: 0\n\n1913: 0\n\n1912: 0\n\n1911: 0\n\n1910: 0\n\n1909: 0\n\n1908: 0\n\n1907: 0\n\n1906: 0\n\n1905: 0\n\n1904: 0\n\n1903: 0\n\n1902: 0\n\n1901: 0\n\n1900: 0\n\n1899: 0\n\n1898: 0\n\n1897: 0\n\n1896: 0\n\n1895: 0\n\n1894: 0\n\n1893: 0\n\n1892: 0\n\n1891: 0\n\n1890: 0\n\n1889: 0\n\n1888: 0\n\n1887: 0\n\n1886: 0\n\n1885: 0\n\n1884: 0\n\n1883: 0\n\n1882: 0\n\n1881: 0\n\n1880: 0\n\n1879: 0\n\n1878: 0\n\n1877: 0\n\n1876: 0\n\n1875: 0\n\n1874: 0\n\n1873: 0\n\n1872: 0\n\n1871: 0\n\n1870: 0\n\n1869: 0\n\n1868: 0\n\n1867: 0\n\n1866: 0\n\n1865: 0\n\n1864: 0\n\n1863: 0\n\n1862: 0\n\n1861: 0\n\n1860: 0\n\n1859: 0\n\n1858: 0\n\n1857: 0\n\n1856: 0\n\n1855: 0\n\n1854: 0\n\n1853: 0\n\n1852: 0\n\n1851: 0\n\n1850: 0\n\n1849: 0\n\n1848: 0\n\n1847: 0\n\n1846: 0\n\n1845: 0\n\n1844: 0\n\n1843: 0\n\n1842: 0\n\n1841: 0\n\n1840: 0\n\n1839: 0\n\n1838: 0\n\n1837: 0\n\n1836: 0\n\n1835: 0\n\n1834: 0\n\n1833: 0\n\n1832: 0\n\n1831: 0\n\n1830: 0\n\n1829: 0\n\n1828: 0\n\n1827: 0\n\n1826: 0\n\n1825: 0\n\n1824: 0\n\n1823: 0\n\n1822: 0\n\n1821: 0\n\n1820: 0\n\n1819: 0\n\n1818: 0\n\n1817: 0\n\n1816: 0\n\n1815: 0\n\n1814: 0\n\n1813: 0\n\n1812: 0\n\n1811: 0\n\n1810: 0\n\n1809: 0\n\n1808: 0\n\n1807: 0\n\n1806: 0\n\n1805: 0\n\n1804: 0\n\n1803: 0\n\n1802: 0\n\n1801: 0\n\n1800: 0\n\n1799: 0\n\n1798: 0\n\n1797: 0\n\n1796: 0\n\n1795: 0\n\n1794: 0\n\n1793: 0\n\n1792: 0\n\n1791: 0\n\n1790: 0\n\n1789: 0\n\n1788: 0\n\n1787: 0\n\n1786: 0\n\n1785: 0\n\n1784: 0\n\n1783: 0\n\n1782: 0\n\n1781: 0\n\n1780: 0\n\n1779: 0\n\n1778: 0\n\n1777: 0\n\n1776: 0\n\n1775: 0\n\n1774: 0\n\n1773: 0\n\n1772: 0\n\n1771: 0\n\n1770: 0\n\n1769: 0\n\n1768: 0\n\n1767: 0\n\n1766: 0\n\n1765: 0\n\n1764: 0\n\n1763: 0\n\n1762: 0\n\n1761: 0\n\n1760: 0\n\n1759: 0\n\n1758: 0\n\n1757: 0\n\n1756: 0\n\n1755: 0\n\n1754: 0\n\n1753: 0\n\n1752: 0\n\n1751: 0\n\n1750: 0\n\n1749: 0\n\n1748: 0\n\n1747: 0\n\n1746: 0\n\n1745: 0\n\n1744: 0\n\n1743: 0\n\n1742: 0\n\n1741: 0\n\n1740: 0\n\n1739: 0\n\n1738: 0\n\n1737: 0\n\n1736: 0\n\n1735: 0\n\n1734: 0\n\n1733: 0\n\n1732: 0\n\n1731: 0\n\n1730: 0\n\n1729: 0\n\n1728: 0\n\n1727: 0\n\n1726: 0\n\n1725: 0\n\n1724: 0\n\n1723: 0\n\n1722: 0\n\n1721: 0\n\n1720: 0\n\n1719: 0\n\n1718: 0\n\n1717: 0\n\n1716: 0\n\n1715: 0\n\n1714: 0]" time="0.067"><properties><property name="score" value="0.1790391852295172" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Hamish Morrison\n\nHamish Morrison (born 8 September 1997) is an Australian professional rugby union player who plays for the in the Super Rugby competition. His position of choice is fullback.\n]" time="0.014"><properties><property name="score" value="0.022074959260676152" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02207496&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02207496
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Tom Brown's Schooldays (film)\n\nTom Brown's Schooldays is a 1921 British silent drama film directed by H. B. Parkinson and starring Milton Rosmer, Eva Moore and C. M. Hallard. It is an adaptation of Thomas Hughes' novel &quot;Tom Brown's School Days&quot;.\n\n\n]" time="0.012"><properties><property name="score" value="0.044957489977726084" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04495749&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04495749
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Rabi Kaul\n\nRabi Kaul (born 1951) is a social activist, writer, human rights activist, former Director of Amnesty International India, and former Chairperson of Alternative Law Forum, Bangalore, India. He is a graduate of the University of Delhi.\n\nHis human rights work has been a major part of his life since 1981 when he began his work in civil liberties defence and the protection of human rights in general, and of Dalit rights in particular. In the 1980s, he was active in the movement against the Emergency. He was in prison for a few months during that time. He is widely respected as a human rights activist, and one of the most important advocates for social justice in India.\n\nKaul served as the Executive Director of the International Humanist and Ethical Union, and as the International Secretary of the International Humanist and Ethical Union, and is a past Chair of the IHEU Coordinating Committee.\n\nHe has worked with Amnesty International since 1989. During his tenure with Amnesty International, Kaul was Director of the organization's Indian chapter, which he helped to build into one of the largest branches in the world. He served as the Regional Director of Amnesty International for South Asia. In December 2012 he was replaced by Raviya Ismail.\n\nKaul also serves as an adviser to the South Asia Human Rights Documentation Centre. He is a recipient of the UN Human Rights Award, the Yash Bharti Award and the \nPeople's Union for Civil Liberties (PUCL) Award, amongst other recognitions. He is also a recipient of the Gandhi Peace Prize (2011).\n\nKaul has written extensively on human rights issues, social justice, and democracy. He has published a number of books, including &quot;Democracy, Minority Rights and Violence&quot; (2008), &quot;Religion, Violence and Non-Violence&quot; (2005) and &quot;Whither Secularism&quot; (1999).\n\nHis publications include:\n\n]" time="0.025"><properties><property name="score" value="0.005288081239927935" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00528808&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00528808
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Erskine Hamilton Childers\n\nErskine Hamilton Childers (; 17 June 1870 \u2013 20 November 1922) was an Irish politician who served as the fourth President of Ireland, from 25 June 1921 to his death on 20 November 1922. He was a member of the Irish Republican Brotherhood and Sinn F\xe9in.\n\nErskine Hamilton Childers was born in England in 1870, the eldest son of Robert Erskine Childers, a British Army officer and later a writer and political theorist.\n\nIn the course of his lifetime he spelled his surname in a number of different ways, including Childers and Childs.\n\nChilders attended school at Stubbington House, Fareham. At the age of 16 he became a clerk in a Southampton shipping office. He joined the Liberal Party and became a political speaker on its behalf. In 1899 he was elected as a councillor to Southampton County Borough Council.\n\nHe married Molly Zichy-Woinarski (sister of Francis &quot;Frank&quot; Zichy-Woinarski) in 1908, but had no children.\n\nChilders was a keen yachtsman, and in 1906, with Sir Thomas Lipton, he set up the Shamrock IV syndicate to challenge for the Americas Cup.\n\nHe participated in the unsuccessful Irish Convention of 1917\u20131918, a body established by British Prime Minister David Lloyd George which negotiated an agreement with Irish representatives in an attempt to end the period of Irish nationalist insurrection. The Convention, which met at Dublin Castle, made a number of suggestions regarding the governance of Ireland, which were partially incorporated into the Government of Ireland Act 1920. It was Childers, who, in September 1917, raised the need for a &quot;conversation&quot; between Irish and British leaders.\n\nAfter the Easter Rising of 1916, and his arrest for involvement, Childers helped reorganise the Irish Republican Brotherhood (IRB) in an effort to steer it towards constitutional activity and away from militancy.\n\nIn the 1918 general election, Childers stood as a Sinn F\xe9in candidate in Wicklow. Although he did not win the seat, he polled well and the Sinn F\xe9in party decided to nominate him as one of its members of the Irish delegation to the Paris Peace Conference. It was at the Paris Peace Conference that he met Arthur Griffith and became a convinced Irish republican.\n\nHe became a committed supporter of Arthur Griffith's Sinn F\xe9in party. In 1919, he was arrested and charged with sedition, because of his support of the Irish Republican Army (IRA).\n\nHe was elected unopposed as a Sinn F\xe9in Teachta D\xe1la (TD) to the 2nd D\xe1il at the 1921 elections for the Dublin Mid constituency. He supported the Anglo-Irish Treaty of 1921, and became a member of the Irish Free State Seanad in 1922.\n\nHe became the fourth President of the Executive Council of the Irish Free State (i.e. prime minister) in June 1921, and on 26 June 1921 he was appointed to the office of President of the Republic, (which was to become that of President of the Executive Council on the enactment of the 1922 Constitution in August 1922).\n\nIn 1921, the Government of Ireland Act 1920 split the island into two autonomous territories within the United Kingdom: the six north-eastern counties became Northern Ireland and the rest of Ireland became Southern Ireland, governed by the Irish Free State government in Dublin. However, the Parliament of Northern Ireland chose to opt out of the Free State in 1922. Under the terms of the Anglo-Irish Treaty, this meant that Northern Ireland would become part of the Irish Free State, but the British refused to accept this, saying it would lead to the &quot;break-up of the United Kingdom&quot;. The Irish Taoiseach and President of the Executive Council (prime minister), W. T. Cosgrave, insisted that the British had breached the Treaty and that the Free State had the right to &quot;withdraw from the British Commonwealth of Nations&quot;, but he chose not to use this option. Instead, he demanded that the Northern Ireland government do so, and thus abolish itself.\n\nChilders was shot dead in his home on Inishmore, Aran Islands, County Galway, in November 1922, two months after becoming President. The following January, the &quot;Irish Independent&quot; reported that Michael Noyk, the Russian Consul in Galway, and Joseph Murray, a member of the Irish Republican Brotherhood, were both implicated in his murder. The paper stated that Murray was sent by James O'Connor to warn Childers to stop political work in the Four Courts or else he would be killed, while Noyk knew of the plot and failed to notify the police. A police officer told the paper that he believed it was an ordinary robbery that &quot;went wrong&quot; as &quot;Childers was too tough a man to be robbed&quot;.\n\n]" time="0.073"><properties><property name="score" value="0.0849792428576038" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08497924&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08497924
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Sarah Williams (social worker)\n\nSarah &quot;Sally&quot; Williams (1880 \u2013 April 23, 1951) was an American social worker. Williams established herself as an activist, fighting for women's suffrage, juvenile delinquency, and public health. Williams was well known for her innovative and groundbreaking approaches to improving the quality of life for children, the public, and the underprivileged. Williams began working for the New York City Health Department in 1918, and in 1925 she became the director of the Bureau of Child Guidance. Williams was the only woman in the country with such a position at the time, and held the position until her death in 1951.\n\nSarah Williams was born in 1880 in Philadelphia, Pennsylvania, the second child of James Williams, a grocer and later a dentist, and Ella Burt Williams. At age 10, she attended the Powelton Avenue School in Philadelphia, and by age 12 she had already decided that she wanted to be a teacher. At age 13, Williams attended the Girls' High School and excelled academically. She took part in school plays and oratorical contests, and upon graduation, Williams attended Swarthmore College in 1896. Williams majored in Greek and Latin and continued to study oratorical contests and plays. After graduating in 1900, Williams continued her education at Cornell University, where she studied American History and Economics. While studying at Cornell, Williams became a member of the college women's debating team, which would later become a hobby for her. Upon graduation, Williams pursued graduate studies at Columbia University, where she received her M.A. degree in Social Science in 1907. At Columbia, Williams received one of her most important honors, the Demorest Fellowship. After graduation, Williams decided to enter the field of social work.\n\nAfter receiving her Master's Degree, Williams began to work as a social worker in Philadelphia. While working in Philadelphia, Williams became interested in the idea of juvenile courts, and worked to improve them. Williams began traveling to study different models of juvenile courts and while in London, she visited the Infant Welfare and the Juvenile Probation Departments. While visiting a British Probation Home, Williams became interested in the training of child welfare workers. She returned to the United States and started a training program for juvenile probation officers in Philadelphia.\n\nIn 1911, Williams accepted a job as the chief social worker of the Infant Welfare Association of Philadelphia. In 1915, Williams moved to New York City and accepted a position as the executive secretary of the Child Labor Committee. Williams continued her work with juvenile delinqu]" time="0.025"><properties><property name="score" value="0.021473547538554995" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02147355&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02147355
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Tips for Buying Brooklyn End Tables By Bradington-Young\n\nHow to purchase a Brooklyn End Tables By Bradington-Young\n\nTop Tips To Help You Buy And Care For Your Furnishings It doesn't require much time to study this informative article, nonetheless the rewards are fantastic. You is going to be shocked whenever you realize how these guidelines alter your shopping behaviors. When it comes to furniture, there is little if any difference between affordable and top quality. You will get everything you buy and don't be concerned about inferior quality. If you adored this information and you would certainly such as to receive even more facts concerning best camping tents reviews i implore you to check out our own page. Solid wood furniture is perhaps the most suitable purchase. It lasts lengthier and requires less maintenance than different\n\nWhat you should know before buying end side tables .\n\nSmart Tips For Purchasing Quality Affordable Furniture It 's what holds your meal since you eat. What keeps one's body aloft when you sleep. What contains your best valuables to guard them. It holds your lamp, television and books. Furniture is available to us, within our office, at home and elsewhere. Learn how to be a smarter end side tables shopper by reading the below information. If you've got youngsters, you must avoid purchasing end side tables with sharp edges. Little ones have a tendency to trip and fall often which can create a visit to the emergency room if the end side tables has sharp edges. When buying end side tables for the family room, seek out simple styles. You may well be totally into that black leather couch, but a couch adorned with a bunch of floral designs is not something that you are going to be thankful for. You and your children can be happy with contemporary end side tables. Go shopping for end side tables with your hues as your guide. You might love a certain piece, but when you go it with the room, it may not hold the existing decor. Avoid this from happening. It's easy to by\n\nThis bed was as described, very easy to assemble, and had good instructions. Very nice and firm bed. Assembly went well. This bed was as described, very easy to assemble, and had good instructions. Very nice and firm bed. Assembly went well.\n\nBrooklyn End Tables By Bradington-Young...\n\n\n\nSee Products Descriptions &amp; Reviews!!]" time="0.026"><properties><property name="score" value="0.04862376707396014" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04862377&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04862377
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[&quot;This is the work of idiots who do not know what they are doing. They have no idea about cars, engines or aerodynamics, and in the end they will pay for it,&quot; team principal Peter Sauber said in a statement.\n\nF1 racing, one of the world's most lucrative sports, has for years been plagued by small teams on tight budgets. The sport's commercial supremo Bernie Ecclestone is trying to improve the situation and has created a fund, estimated to be worth up to $40 million, to support teams in dire straits.\n\n&quot;The idea of taking money from one and giving it to the other is not very positive,&quot; Sauber said. &quot;It creates the possibility of a new form of slavery. I do not think that anyone who worked in the factories during the war thought about whether a tank would end up being driven by a German or an Englishman. He was just happy to be making money. But people are not stupid, and there are some consequences that need to be considered. This is just crazy. It will never work.&quot;\n\nFerrari, BMW-Sauber, McLaren and Renault, four of the sport's top teams, have signed up to the initiative but Ferrari president Luca di Montezemolo said Sauber was missing the point.\n\n&quot;If the results of the championship are being hampered by an imbalanced distribution of resources, then there is a reason for the redistribution,&quot; he said.\n\nSauber's team have won 10 world championships but are among a number of outfits on the verge of going under as the sport's top teams spend millions of dollars in search of an elusive tenth title.]" time="0.014"><properties><property name="score" value="0.10013420409216711" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.1001342&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.1001342
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Paladin\n\nA paladin is the sworn defender of a group of creatures, a good-aligned organization, or a person. Most are deeply religious, but the few who aren't still serve their groups or causes with fanatical zeal. The paladin is essentially a warrior of the gods, and most spend their lives wandering the world of Oerth, aiding those they deem good or just. Although a paladin will fight to slay a horde of monsters or a mad king, he will not do so out of a lust for battle or destruction, but out of a need to keep the forces of darkness at bay.\n\nPaladins are renowned for their bravery, their unyielding dedication to the principles of good, and their mastery of martial arms. It is rare to find an evil paladin.\n\nAbilities: Strength and Charisma are both prime requisites for a paladin. Strength is his prime requisite, followed by Charisma. Strength governs how hard a paladin can hit, how effectively he can wear armor, and his carrying capacity. Charisma governs the strength of his spells and the DC's of his class skills. A paladin must be lawful good.\n\nHit Die: d10\n\nStarting Age: Complex.\n\nStarting Gold: Complex.\n\nClass Skills\n\nThe paladin's class skills are Concentration (Con), Craft (Int), Diplomacy (Cha), Handle Animal (Cha), Heal (Wis), Intimidate (Cha), Knowledge (nobility and royalty) (Int), Knowledge (religion) (Int), Profession (Wis), Ride (Dex), and Sense Motive (Wis).\n\nSkill Points at First Level: (2 + Int modifier) x 4.\n\nSkill Points at Each Additional Level: 2 + Int modifier.\n\nClass Features\n\nAll of the following are class features of the paladin.\n\nWeapon and Armor Proficiency: Paladins are proficient with all simple and martial weapons, with all types of armor (heavy, medium, and light), and with shields (except tower shields).\n\nAura of Good (Ex): The power of a paladin's aura of good (see the detect good spell) is equal to her paladin level.\n\nDetect Evil (Sp): At will, a paladin can use detect evil, as the spell.\n\nSmite Evil (Su): Once per day, a paladin can call out to the powers of good to aid her in her struggle against evil. As a swift action, the paladin chooses one target within sight to smite. If this target is evil, the paladin adds her Charisma bonus (if any) to her attack rolls and adds her paladin level to all damage rolls made against the target of her smite. If the target of smite evil is an outsider with the evil subtype, an evil-aligned dragon, or an undead creature, the bonus to damage on the first successful attack increases to 2 points of damage per level the paladin possesses. Regardless of the target, smite evil attacks automatically bypass any DR the creature might possess.\n\nIn addition, while smite evil is in effect, the paladin gains a deflection bonus equal to her Charisma modifier (if any) to her AC against attacks made by the target of the smite. If the paladin targets a creature that is not evil, the smite is wasted with no effect.\n\nThe smite evil effect remains until the target of the smite is dead or the next time the paladin rests and regains her uses of this ability. At 4th level, and at every three levels thereafter, the paladin may smite evil one additional time per day, as indicated on Table: Paladin, to a maximum of seven times per day at 19th level.\n\nDivine Grace (Su): At 2nd level, a paladin gains a bonus equal to her Charisma bonus (if any) on all saving throws.\n\nLay on Hands (Su): Beginning at 2nd level, a paladin can heal wounds (her own or those of others) by touch. Each day she can use this ability a number of times equal to 1/2 her paladin level plus her Charisma modifier. With one use of this ability, a paladin can heal 1d6 hit points of damage for every two paladin levels she possesses. Using this ability is a standard action, unless the paladin targets herself, in which case it is a swift action. Despite the name of this ability, a paladin only needs one free hand to use this ability.\n\nAlternatively, a paladin can use this healing power to deal damage to undead creatures, dealing 1d6 points of damage for every two levels the paladin possesses. Using lay on hands in this way requires a successful melee touch attack and doesn't provoke an attack of opportunity. Undead do not receive a saving throw against this damage.\n\nAura of Courage (Su): Beginning at 3rd level, a paladin is immune to fear (magical or otherwise). Each ally within 10 feet of her gains a +4 morale bonus on saving throws against fear effects. This ability functions while the paladin is conscious, but not if she is unconscious or dead.\n\nDivine Health (Ex): At 3rd level, a paladin gains immunity to all diseases, including supernatural and magical diseases.\n\nDivine Bond (Sp): Upon reaching 5th level, a paladin forms a bond with her god. This bond can take one of two forms. Once the form is chosen, it cannot be changed.\n\nThe first type of bond allows the paladin to enhance her weapon as a standard action by calling upon the aid of a celestial spirit for 1 minute per paladin level. When called, the spirit causes the weapon to shed light as a torch. At 5th level]" time="0.081"><properties><property name="score" value="0.04793638608715734" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04793639&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04793639
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[John George Evans\n\nJohn George Evans (April 2, 1809 \u2013 April 18, 1878) was an English-born educator, newspaper editor and political figure in Nova Scotia, Canada. He represented Yarmouth in the Nova Scotia House of Assembly from 1878 to 1882 as a Liberal-Conservative member.\n\nHe was born in Shrewsbury, Shropshire, the son of George Evans and Mary Ann Pritchard, and came to Nova Scotia with his parents in 1816. In 1833, Evans married Ann Squire. He taught school in Digby. Evans was a supporter of the Halifax Conservative Party. He served as deputy collector of customs at Yarmouth from 1843 to 1867 and in 1868 was elected to the provincial assembly. Evans was editor and publisher of the &quot;Yarmouth Herald&quot; from 1859 to 1870. He was elected speaker for the provincial assembly in 1882 but did not serve because he was ill.\n\nHis son George George Evans also served in the provincial assembly.\n]" time="0.016"><properties><property name="score" value="0.055716349178697167" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05571635&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05571635
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Star of the Sea (disambiguation)\n\nStar of the Sea may refer to:\n\n\n\n\n\n\n\n\n\n]" time="0.015"><properties><property name="score" value="0.0014593145410468633" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00145931&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00145931
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Flavobacterium gilvum\n\nFlavobacterium gilvum is a Gram-negative, strictly aerobic, rod-shaped and non-spore-forming bacterium from the genus of &quot;Flavobacterium&quot; which has been isolated from sandy soil.\n\n \n]" time="0.014"><properties><property name="score" value="0.014769672950831758" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01476967&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01476967
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Ariel Hernandez\n\nAriel Andr\xe9s Hern\xe1ndez Callejas (born 7 August 1995) is a Colombian professional footballer who plays as a midfielder for Deportivo Cali.\n\nHernandez's career began in the youth ranks of Envigado, prior to a move to Deportivo Cali. He played in the club's youth system and featured regularly for the reserves in the Categor\xeda Primera B. Hernandez made his first team debut in a 3\u20130 victory over Alianza Petrolera in July 2014. He appeared in twelve further matches during 2014, including a 2\u20131 defeat to La Equidad and a 1\u20130 loss to Atl\xe9tico Nacional. Hernandez scored his first senior goal on 4 August 2015 in a 2\u20131 victory over C\xfacuta Deportivo. The strike was his sixth of the season and proved to be the winner.\n]" time="0.014"><properties><property name="score" value="0.08346441896430745" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08346442&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08346442
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Suspension (mechanical)\n\nIn mechanical engineering, suspension refers to a system of shock absorbers and springs designed to improve the ride quality of a vehicle or other types of machinery. It is an essential component in wheeled vehicles, providing a convenient way to deal with irregularities in the road surface.\n\nMost modern vehicles are fitted with some sort of suspension. It is the primary means of isolating the passengers from the road, as well as absorbing the bumps and vibrations from the road.\n\nThe suspension system on a vehicle provides the following functions:\n\nThere are many systems available that provide all or most of these functions, but most can be categorized by their design and motion characteristics.\n\nThe simplest suspension systems are rigid, and often called &quot;hard suspensions&quot;, because they don't use any fluid damping. Instead, they rely entirely on the elasticity of the metal or the springiness of the suspension elements to soak up the bumps. Some early vehicles used torsion springs as suspension, but modern systems almost all use either coil or leaf springs, or shock absorbers.\n\nThe first and simplest suspension system was used by the French inventor of the wheel, Leonardo da Vinci. It consisted of four flexible limbs connected to the carriage. These limbs served as both the suspension and steering system.\n\nIn addition, Leonardo designed a system that included advanced features such as brake linings, bearings and a double-action brake system. A suspension system for cars was introduced by M. Thon in France in 1898. It consisted of a network of rubber bands that took the place of conventional springs. The rubber bands were enclosed in a conventional coil-spring housing. An advantage of this system was its light weight. Thon's design was introduced to the United States by John M. Curtis in 1904.\n\nBy the 1920s, many independent suspension manufacturers were using the pneumatic springing system as the basis of their systems, as well as using rubber or steel cords for springing. The principal disadvantage of the pneumatic system is that the spring-pressure increases with the compression of the suspension. In addition, a system that is not correctly adjusted will not ride well.\n\nThe third kind of suspension system was developed by the Morgan Motor Company, based in the UK. In this system, the spring is attached to the frame, and the shock absorbers are attached to a special hub, which attaches to the front and rear axles. This system is still sometimes seen in modern cars, since it is less expensive than the other systems and in most cases provides an adequate ride. This system has not been popular in Europe and North America since the 1930s.\n\nThe most common suspension system in modern cars, used in everything from Ford Model Ts to the Chevrolet Corvette, is the &quot;independent front suspension&quot;, or &quot;live axle&quot;. In this system, the suspension is a more or less conventional system of springs and shock absorbers, with the additional feature of a locating link (the king pin in an axle-based suspension). This allows the axle to move up and down relative to the chassis. The locating link ends are attached to the car body, which provides the rolling element to the system.\n\nThe other feature of this system is the axle, which provides the pivoting element of the suspension system. The wheel hub is attached to the ends of the axle, and the wheel itself is attached to the hub. The spring system, located in the body of the vehicle, is attached to the axle via the locating link. The body of the car, therefore, provides the &quot;pivot&quot; of the suspension.\n\nA number of variations on this design have been used, most notably independent rear suspension, in which the rear wheels are located by a trailing arm (a form of wishbone) instead of the traditional transverse leaf spring. Another design commonly found on trucks is the &quot;live rear axle&quot;. In this system, the axle is located by an arm from the leaf spring. The rear wheels can move up and down relative to the chassis without any location from the body.\n\nA design commonly used on many racing cars and on some passenger cars is the DeDion or De Dion suspension, named after its inventors, the De Dion brothers. In this system, the two wheels are located by a triangular or &quot;semi-trapezoidal&quot; bar. The pivot is provided by a bearing block mounted to the car body, while the pivoting point for the wheels is a ball joint on the axle.\n\nThe DeDion design is an expensive design, due to the cost of the wheels, brakes and other parts. However, it does provide excellent wheel location, as the wheels can be moved by several inches in any direction without affecting the steering.\n\nRacing cars generally use the DeDion system, as the location of the wheels is the primary function of the suspension system. However, it is also used on a number of sports cars, including the Jaguar XK120, the Maserati 3500 GT and the Lamborghini Gallardo.\n\nOne of the main disadvantages of the DeDion suspension is that the front wheels are not located directly above the chassis, and there is significant scrub radius. For example, a wheelbase of 86\xa0inches (218\xa0cm) on a vehicle with a wheel diameter of 18\xa0inches (46\xa0cm) would produce a scrub radius of around 8\xa0inches (20\xa0cm). This can cause the car to steer and corner differently depending on the tire wear, since the scrub radius is at its largest when the tires are new.\n\nWhile an advantage of the DeDion system is that it is lighter than a comparable design, this is not a major consideration for racing cars. However, since racing cars are generally not driven at high speeds for long distances, the &quot;chassis bounce&quot; caused by the scrub radius is not an issue.\n\nOne of the first types of suspension system was introduced by Earle Crooker and was called &quot;Independent Coil Spring Suspension&quot;. It was a brand name that existed only in the late 1920s. The name was used by C. G. Ames Co.\n\nIn this system, the springs are attached directly to the frame. In this system, the wheel is attached to the axle, which is supported by a bearing attached to the frame, not the car body.\n\nThe primary advantage of this system is the ease with which the wheelbase can be changed. In most cars, the wheels are located by a link that is attached to the axle and the frame, and the wheels can move relative to the car body. In this system, there is a much smaller length change, as the wheels move up and down relative to the axle.\n\nThe disadvantage of this system is that the body is no longer a &quot;pivot point&quot; for the suspension. The body is now part of the suspension system. This makes it very difficult to change the camber of the wheels (i.e., the angle of the wheel relative to the car body). In addition, the wheel can only move in one plane, as the upper control arm and the lower control arm limit the movement of the wheel. The car may need to be aligned more often, as the geometry of the suspension can change more than a suspension system where the wheels move relative to the body.\n\nThis system is commonly found on American cars built in the 1940s and 1950s, such as the Nash Ambassador, Hudson, and other brands. It is less commonly found on European cars, although it was used on some Mercedes-Benz models]" time="0.076"><properties><property name="score" value="0.0006130906122628886" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00061309&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00061309
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[L\u2019avvento di Zuckerberg ha definitivamente maturato il passaggio dalla televisione tradizionale a quella digitale. Il successo del prodotto Facebook per Smartphone ha determinato la consacrazione di un nuovo media che non si muove pi\xf9 solo tra i PC ma che, a partire da un apparecchio mobile, si espande a livello planetario. In un solo anno la piattaforma del social network di Menlo Park ha infatti scalato le vette dell\u2019ecosistema tecnologico con un balzo da 5,8 milioni di utenti nel 2010 a quasi 60 milioni di abbonati nel 2011.\n\nSe da un lato le prestazioni del prodotto Facebook per Smartphone hanno assicurato un rapido avanzamento delle quotazioni del titolo in borsa, dall\u2019altro l\u2019entrata in azione del colosso californiano \xe8 stata uno dei fattori principali per l\u2019incremento dei dati di traffico degli oltre 90 milioni di italiani che utilizzano ogni giorno Internet dal proprio smartphone.\n\nMentre negli Stati Uniti la situazione non \xe8 ancora perfetta ma le promesse sono grandi, in Italia il Facebook per Smartphone ha assicurato a Facebook il consolidamento di una base di utenti in crescita. Con le recenti performance che sta raggiungendo il social network, sempre pi\xf9 persone decidono di interagire dal loro smartphone con i contatti personali e gli amici di Facebook, grazie alle nuove funzionalit\xe0 dell\u2019applicazione che ha recepito le nuove esigenze degli utenti. Il fenomeno sta infatti dilagando in tutto il mondo, come dimostra il caso della Spagna che in poco tempo ha passato da 4 milioni di utenti a 15 milioni.\n\nTra le novit\xe0 dell\u2019applicazione troviamo un nuovo caricamento della home, in cui sono stati aggiunti i nuovi pulsanti \u201cGiochi\u201d e \u201cScrivi al tuo amico\u201d, inoltre sono stati eliminati i due pulsanti \u201cFoto e Storie\u201d che avevano originariamente l\u2019incarico di condividere le immagini su Facebook e Instagram.\n\nIl suo funzionamento si basa su un semplice scambio di messaggi in cui l\u2019utente, dall\u2019apposita interfaccia di Facebook per Smartphone, pu\xf2 utilizzare una tastiera per comporre il testo e inviarlo ai suoi contatti. L\u2019opzione gi\xe0 si integra con il servizio di messaggistica istantanea di Facebook, WhatsApp, e con Messenger. Per effettuare una ricerca \xe8 necessario digitare il nome di un contatto nell\u2019apposita barra delle ricerche.\n\nInoltre per chi non dispone di una connessione Internet a banda larga, \xe8 stata introdotta la possibilit\xe0 di caricare un\u2019immagine senza l\u2019obbligo di avere una connessione internet. L\u2019immagine viene caratterizzata da una piccola animazione che si sviluppa gradualmente sul display del telefono, permettendo di vedere il progressivo aggiornamento dell\u2019immagine.]" time="0.031"><properties><property name="score" value="0.18441369518047013" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.1844137&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.1844137
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Gerry is head of commercial and operational support and has worked for the Trust since 1988.\n\nHaving been educated in Scotland, he was an apprentice engineer and a welder, specialising in the design and maintenance of special mechanical and electronic equipment used in nuclear power stations.\n\nHe worked on naval nuclear submarines at Rosyth, before working in the oil and gas industry and eventually becoming an electrician at oil rigs in the North Sea.\n\nAfter working in various positions in the engineering industry, he joined North East Ambulance Service in 1987, initially as an electrical engineer. He then became training co-ordinator, before joining Northumberland, Tyne &amp; Wear Ambulance Service in 1988 as a training officer.\n\nAs an operational support manager, Gerry now manages the administration support, communications and logistics for the emergency and non-emergency services in the area.]" time="0.020"><properties><property name="score" value="0.01689615170737957" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01689615&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01689615
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Minister for Justice and Equality, Charlie Flanagan TD, and Minister for Social Protection, Leo Varadkar TD, today launched a joint campaign in support of the Crosscare Pathways Project, a nationwide outreach service for rough sleepers.\n\nLaunched this morning in Dublin, the \u2018Living On The Streets\u2019 campaign aims to draw attention to the fact that there are approximately 1,000 rough sleepers in Ireland. The aim of the project is to help these rough sleepers access the appropriate support services, whether it be residential or non-residential. The campaign highlights the issue of rough sleeping in Dublin city, which is rising at an alarming rate, with an increase of over 40% in the number of people living on the streets since 2013. The campaign\u2019s focus is on raising awareness of this issue.\n\nThe campaign involves advertisements on the city\u2019s LUAS and Dublin Bus, as well as featuring on O\u2019Connell Street, Portobello and Talbot Streets. It also involves the installation of posters across the city, specifically in locations where the homeless are more likely to reside.\n\nSpeaking at the campaign launch, Minister Flanagan said: \u201cRough sleeping has reached record levels and has become a visible problem in our capital city. The launch of this campaign will increase awareness of the services provided by the Crosscare Pathways Project to rough sleepers, so that they are aware of the options available to them. Rough sleepers should be referred to the services that will best suit their needs. As Minister for Justice and Equality, I am particularly concerned to ensure that rough sleepers are not penalised for begging and sleeping on the streets.\n\n\u201cThe Government is determined to prevent and reduce homelessness, to support rough sleepers to exit homelessness, and to prevent homelessness for others at risk. Over \u20ac70 million has been provided to support homeless services and prevent homelessness, which includes funding to frontline services such as the Crosscare Pathways Project, who work to prevent rough sleeping by providing outreach support to rough sleepers.\u201d\n\nMinister for Social Protection Leo Varadkar said: \u201cThe \u2018Living On The Streets\u2019 campaign brings home the reality that the street is a harsh and unforgiving place to be, particularly during the winter months. The response from the public will be crucial if we are to address this problem effectively.\n\n\u201cThere are over 1,000 rough sleepers in Dublin city and unfortunately this is a significant increase on previous years. That is why it is so important that we engage the public, using this campaign as an opportunity to highlight the services available to rough sleepers, including the Crosscare Pathways Project. It is a complex problem that requires a complex response, involving a number of different organisations, as well as a societal change in attitude.\n\n\u201cWith the help of my colleague, Minister Flanagan, we have increased funding for the homeless to \u20ac70 million this year, which will provide more than 3,000 more beds in a wide variety of services. We have also agreed to provide an additional \u20ac5 million over the next two years.\u201d\n\nCrosscare Pathways Project is a national organisation with local services based in Ballymun, Jobstown, Dublin city centre, Trim, Waterford and Wexford. It provides a range of services to homeless people and others at risk of homelessness, as well as people sleeping rough, particularly those sleeping on the streets of Dublin. The Crosscare Pathways Project also provides services for homeless people in Carlow, Cork, Donegal, Dublin North, Dublin South and Dublin West.\n\nThe project\u2019s services include outreach services, sleeping bags, socks, warm clothes, food, clothing, showers, laundry and toiletries.\n\nThe campaign is the work of the cross-agency Dublin Central Area Partnership (DCAP), which is comprised of An Garda S\xedoch\xe1na, Dublin City Council, Dublin Regional Homeless Executive (DRHE), Dublin Fire Brigade, Dublin Regional Housing Executive (DRHE), Dublin Street Pastors, Homeless Hotline, Irish Red Cross, Samaritans, Simon Community and Street Reach.\n\nThe campaign is supported by a number of major partners including Dublin City Council, Bus \xc9ireann, Dublin Bus, Dublin Airport Authority, Transdev and Luas Cross City.\n\nENDS]" time="0.067"><properties><property name="score" value="0.010432630253765663" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[During the war in Bosnia, in the capital city of Sarajevo, there was a hotel that was shelled repeatedly, and you could tell by the black marks on the facade which were caused by sniper fire and which were caused by shrapnel.\n\nHow do I know? Because my cousin owned the hotel, and I knew her back when she was a young student in Yugoslavia in the 1960s.\n\nShe was smart, but she was]" time="0.018"><properties><property name="score" value="0.030914706647472423" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03091471&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03091471
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[See the live USD price. Control the current rate. Convert amounts to or from BABB and other currencies with this simple calculator.\n\nHow did the currency on yesterday? The average value Babble price for convert (or exchange rate) during the day was $0.009986. Min. Babble value was $0.009322. Max. BABB price was $0.011702. BABB price dropped by 7.5% between min. and max. value. Let's see what's next.\n\nOlder news about Babble\n\nThe Babble increased by 15.27% on Thursday 17th of September 2019 Let's take a look at interesting data from yesterday. BABB price increased by 15.27% between min. and max. value. Max. BABB price was $0.010422. Min. Babble value was $0.008959. The average value Babble price for convert (or exchange rate) during the day was $0.010405. Keep it up.\n\nThe Babble dropped by 17.49% on Wednesday 16th of September 2019 How did the currency on yesterday? The average value Babble price for convert (or exchange rate) during the day was $0.010164. Min. Babble value was $0.009265. Max. BABB price was $0.011671. BABB price dropped by 17.49% between min. and max. value. Let's see what's next.\n\nThe Babble dropped by 23.23% on Tuesday 15th of September 2019 How was the currency exchange rate changed on yesterday? BABB price dropped by 23.23% between min. and max. value. Max. BABB price was $0.012727. Min. Babble value was $0.009289. The average value Babble price for convert (or exchange rate) during the day was $0.012042. Watch the next day.\n\nThe Babble increased by 16.08% on Monday 14th of September 2019 And we have data for yesterday. Min. Babble value was $0.009069. Max. BABB price was $0.011378. The average value Babble price for convert (or exchange rate) during the day was $0.010781. BABB price increased by 16.08% between min. and max. value. Certainly it is good news for all.\n\nThe Babble dropped by 7.02% on Sunday 13th of September 2019 Let's see on yesterday. BABB price dropped by 7.02% between min. and max. value. Min. Babble value was $0.010600. Max. BABB price was $0.011775. The average value Babble price for convert (or exchange rate) during the day was $0.011251. Let's see what's next.\n\nThe Babble increased by 5.37% on Saturday 12th of September 2019 Let's take a look at interesting data from yesterday. Min. Babble value was $0.010724. Max. BABB price was $0.011425. The average value Babble price for convert (or exchange rate) during the day was $0.010947. BABB price increased by 5.37% between min. and max. value. Certainly it is good news for all.\n\nThe Babble increased by 7.5% on Friday 11th of September 2019 How was the currency exchange rate changed on yesterday? BABB price increased by 7.5% between min. and max. value. Min. Babble value was $0.010348. Max. BABB price was $0.011268. The average value Babble price for convert (or exchange rate) during the day was $0.010825. Keep it up.\n\nAnother conversions\n\nUsc to Babble, Uro to Babble, Uralscoin to Babble, Uscoin to Babble, USD-e to Babble, Usdq to Babble, US Dollar to Bangladeshi Taka, US Dollar to Banca, US Dollar to Bancor, US Dollar to Bankcoin, US Dollar to Bancorplus, US Dollar to Bankera, US Dollar to Banca,]" time="0.103"><properties><property name="score" value="0.20884489083281235" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Mick Campbell: Some Hilarious (but Informed) Jokes About Health Reform\n\nHealth-care reform is under way in the US and health-care reform is under way in Britain. Some people may not have noticed the second fact. Over here, the NHS is now the NHS 2.0, or 3.0, or whatever they are on to. We\u2019re getting used to the idea that it will always be changing, and changing in ways we\u2019re not in total control of.\n\nWith my dry British sense of humor, I\u2019ve always enjoyed the NHS. In fact, I like it so much that I even thought of emigrating to a country with an NHS \u2013 like, say, the USA. So what a treat to discover in my in-box this morning a compilation of jokes, half of them at the expense of the USA and half at the expense of the UK, about health-care reform.\n\nFrom London, Chris Salisbury compiled them, and here are a couple of samples. The first is about healthcare reform in the USA:\n\n\u201cHave you heard about the new Congressional Health Care plan?\u201d\n\n\u201cYeah, the government\u2019s going to give you half of what it takes away.\u201d\n\nAnd the second, about the NHS:\n\nA man is at the doctor\u2019s office, and the doctor tells him, \u201cI\u2019ve got some good news and some bad news for you.\u201d\n\nThe man says, \u201cWell, I can\u2019t take the bad news right now, so give me the good news first.\u201d\n\nThe doctor says, \u201cWell, the good news is that you have an 18-inch penis.\u201d\n\nThe man looks stunned for a moment, and then asks, \u201cWhat\u2019s the bad news?\u201d\n\nThe doctor says, \u201cYour brain\u2019s in your dick.\u201d\n\nThese jokes may be inaccurate, of course. For example, the doctors\u2019 pay packets are pretty great, and some of them can do much better work in the US than in Britain. But one thing I know is for sure: it\u2019s great to see the NHS getting some serious attention in the American media. I bet it makes the doctors happy.]" time="0.037"><properties><property name="score" value="0.04725601573638158" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04725602&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04725602
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Hello.\n\nI have been trying to download the RPM of a file that resides on a university FTP server. I can download files with wget and ftp, but not with rpm. The file is a tar file. I want to be able to download it with rpm. Is this possible?\n\nThanks!]" time="0.019"><properties><property name="score" value="0.012202895636181932" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0122029&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0122029
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[\n\nBuchbesprechung\n\nHeinz Wernicke: Das zerrissene Land.\n\n\n\n1\n\nEin erstaunlich gutes Buch, das sich auf das politische und wirtschaftliche Geschehen Deutschlands in der Epoche von 1870 bis 1950 bezieht. Es zeigt die Interessen, aus denen sich die in der ersten H\xe4lfte der Epoche g\xfcltigen Allianzen zwischen Reichswehr und Arbeiterschaft, den Kirchen, den Nazis und den Kapitalisten in der zweiten H\xe4lfte zu so erstaunlich verschiedenen Gebilden wandelten.\n\n\n\n2\n\nDas Buch kommt zu dem Ergebnis, dass der R\xfcstungsstaat, der heute aus der Weltkriegsgeneration stammt, sich schon in den letzten Jahren des Kaiserreichs vorbereitete und dass er nicht von der Weimarer Republik, sondern von dem inzwischen gescheiterten Gegner der Weimarer Republik konzipiert wurde.\n\n\n\n3\n\nEs schildert die Haltung des Kaisers und seines Adlatus Ebert zu der als Gro\xdfmacht \xfcberlebenswichtigen Schwerindustrie und zu dem Staat, der seit 1871 f\xfcr die industrielle Entwicklung garantiert hat.\n\n\n\n4\n\nEs stellt dar, wie und warum Hitler nicht auf einen Einmarsch in die Tschechoslowakei gefolgt ist und zeigt die Beweggr\xfcnde f\xfcr diese Entscheidung. Diese sind bis heute in der Diskussion \xfcber den Charakter des Nationalsozialismus vernachl\xe4ssigt worden.\n\n\n\n5\n\nEs berichtet \xfcber die Zusammenarbeit von milit\xe4rischer und ziviler Macht im Verwaltungsapparat der Wehrmacht w\xe4hrend des zweiten Weltkriegs, von den Vorg\xe4ngen im Inneren des Staatsapparats, der Gesellschaft und der Armee bis hin zum Prozess gegen die deutschen]" time="0.019"><properties><property name="score" value="0.26694753420341555" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.26694753&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.26694753
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Most of the novels I enjoy are focused on romance, in whatever form it may come. I believe that there\u2019s something for everyone to enjoy, as the author does such a great job of portraying the characters in all their strengths and flaws. Sometimes they go a little over the top with the angst, and it isn\u2019t very realistic, but I understand why. There are authors who never experience any type of conflict with their characters, and the books just feel very one dimensional. Some books can have some really crappy endings, but I have learned to stay away from books that end with the couple together and happy but someone was murdered or something along those lines. I don\u2019t care for books that have the couple separated, but they\u2019ll come back together in the end.\n\nWhat I want in a good romance book is a good, if not a great, ending that makes me feel good, no matter what happens to the characters. As I mentioned earlier, I don\u2019t like books that end in tragedy, and they usually do end in tragedy. There are some that aren\u2019t too tragic, but they just aren\u2019t the endings that I like. This is where the author\u2019s job becomes very difficult. They have to give the reader a reason to feel the way they want them to feel about the characters. If they feel bad, they\u2019ve done their job. If they feel happy, they\u2019ve done their job. As a reader, I\u2019m not looking to be outraged, nor am I looking to be happy. I\u2019m just looking for something that makes me feel.\n\nOn the flip side of that, I know a lot of people don\u2019t want to be emotionally manipulated. They just want to read a book and enjoy it. They don\u2019t want to be forced to feel one way or another. They want to feel neutral. Neutral, to me, is a huge accomplishment in a romance novel. To me, a romance novel has to be able to get a person, even if they don\u2019t think it will, to feel something for the characters. In most books, people want to know how they can make a particular character do what they want them to do. In a romance novel, that\u2019s the main character. They want to know how they can make the characters fall in love, because they see themselves in one or both of them. How would they do it? How would they fix it?\n\nThat is what a reader does when they read a romance novel. They want to know how to fix the problems that the characters have. This is why the ending is so important. It has to make the reader happy, but not to the point where they roll their eyes or feel like they were taken for a ride. I have a great example. In the movie Now and Then, a woman tells her daughter that she is happy. When her daughter tells her that she is, too, she tells her that it\u2019s because they\u2019re both lying. They don\u2019t want to make each other sad, so they\u2019re lying to each other. This is the exact thing that happens in a lot of romance novels, in my opinion. They\u2019re trying to make the readers feel good, but in the process they\u2019re lying. To me, the ending has to make the reader feel what the author was trying to make them feel, not what they want them to feel.\n\nI\u2019m not saying that the book has to end the way I want it to. What I\u2019m saying is that I want the author to be honest. I want them to give me a believable ending that isn\u2019t just thrown together so that they can end the book. I want the ending to mean something, and I think that\u2019s what most readers want. Most people like to be happy, and a good book will give them that. The ending has to be good enough that the reader doesn\u2019t feel like they wasted their time, but also good enough that they\u2019re not completely disappointed. They have to be happy with the story, even if it wasn\u2019t exactly what they wanted.\n\nWhat are some of your favorite romance book endings?]" time="0.102"><properties><property name="score" value="0.014098347343927554" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[1) Running on behalf of a person\n\nThese dates are the running dates for campaigns on behalf of clients. If you are looking to book a campaign on behalf of your brand, please email theteam@roar.co.nz\n\n2) Need to know more?\n\nIf you are looking to know more about the process of booking a campaign, you can find more information on our website. Or alternatively, get in touch with our team at hello@roar.co.nz]" time="0.023"><properties><property name="score" value="0.0158122135219263" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01581221&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01581221
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Ask.com gives you a list of possible answers when you type in a question. Then it shows you the number of people who searched for those questions.\n\nNot a new idea, but very handy.\n\nFiled in: Technology \u2022 Web]" time="0.014"><properties><property name="score" value="0.021546994612642312" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02154699&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02154699
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[There\u2019s nothing more adorable than a real life couple who gush over one another on social media and share the same love story. What\u2019s even more adorable is when you know that they found each other through social media.\n\nTake Brandon Flowers and his wife, Tana. The two met on Twitter and now, they\u2019re celebrating their five year anniversary. If you follow either of their social media accounts, you may have noticed some cute posts from them about their relationship. We want to thank them for allowing us to have a look into their love story, because they\u2019re too cute!\n\nAs fans of The Killers, we love hearing about their adventures.\n\nThey\u2019re pretty cute!\n\nThere\u2019s nothing more adorable than a real life couple who gush over one another on social media and share the same love story. What\u2019s even more adorable is when you know that they found each other through social media.\n\nTake Brandon Flowers and his wife, Tana. The two met on Twitter and now, they\u2019re celebrating their five year anniversary. If you follow either of their social media accounts, you may have noticed some cute posts from them about their relationship. We want to thank them for allowing us to have a look into their love story, because they\u2019re too cute!\n\nAs fans of The Killers, we love hearing about their adventures.\n\nThey\u2019re pretty cute!\n\nI'd just like to say that my wife is everything. \u2014 Brandon Flowers (@flowerboy) August 5, 2017\n\nWedding Day. Not on social media. So I don't know what people think. I know my wife is everything. That's all I need. \u2014 Brandon Flowers (@flowerboy) August 5, 2017\n\nOn the last night of a tour, getting some wife-y kisses. And in case anyone was wondering. This is how we roll. pic.twitter.com/zSXlfUgKXd \u2014 Brandon Flowers (@flowerboy) August 7, 2017\n\nOh yeah. Still be on tour, still be making music. Not yet a father, but that's coming. Still have some things left to do. https://t.co/CJ1vOs8lYA \u2014 Brandon Flowers (@flowerboy) August 7, 2017\n\nYes. It's true. I am getting married. And I'm marrying my best friend. And my favorite person. https://t.co/uOe3DjtMtQ \u2014 Brandon Flowers (@flowerboy) August 5, 2017\n\nThe next night, he admitted that he has not yet become a father, but he is working on it.\n\nDidn't say I was a father yet. Wasn't pregnant before. Wasn't even married. But I love my wife and my future kids. pic.twitter.com/uVEXZHScGz \u2014 Brandon Flowers (@flowerboy) August 6, 2017\n\nYep. Only day two and already love you so much more than I did yesterday. No, that's not enough. Still not enough. pic.twitter.com/cItd8b8HxY \u2014 Brandon Flowers (@flowerboy) August 8, 2017\n\nYou are a goddamn genius. No one can make me laugh like you. \u2014 Brandon Flowers (@flowerboy) August 8, 2017\n\nThe next day, Brandon shared the first picture of the newlyweds.\n\nTana and I on our wedding day. pic.twitter.com/6R5xT0P5zT \u2014 Brandon Flowers (@flowerboy) August 7, 2017\n\nI can't wait to see you two together at our wedding. \u2014 Brandon Flowers (@flowerboy) August 7, 2017\n\nI love my wife and I'm going to spend the rest of my life loving her. https://t.co/2h7LXqHxjC \u2014 Brandon Flowers (@flowerboy) August 7, 2017\n\nBrandon gave a shout out to his brother and the rest of the band for not spoiling their special day.\n\nThis is how much I love my wife. I would've wanted to come here tonight. But I didn't. https://t.co/ox7XY3sPGO \u2014 Brandon Flowers (@flowerboy) August 8, 2017\n\nJust a normal day. A normal day with my wife. She is my favorite person in the world. We have no plans. pic.twitter.com/vNNkIOCcUQ \u2014 Brandon Flowers (@flowerboy) August 8, 2017\n\nCurious to see the rest of their love story?\n\nHappy 5 years to the love of my life, my best friend and my favorite person. I still feel like I'm the luckiest]" time="0.075"><properties><property name="score" value="0.08730302075724115" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The mainline of PC\u2019s arrived on Friday in great shape. In addition to the newer No. 3 (blue) there was an earlier No. 4 (green), a No. 5, and two later No. 5\u2019s. To fill the gaps between these 5\u2019s I bought two\u2026well, three\u2026plows from another fellow who also had lots of couplers and castings. It\u2019s good to have some of each for testing fit of things, but for permanent operation I\u2019ll probably want to get a handful of No. 5\u2019s.\n\nNow it\u2019s back to finishing the minis and collecting the accessories. The insides of these last few cars could use some work, like fitting the walls for the corner curves. I\u2019ll do that at the same time as adding the doors, and a couple of weeks should have them done. Once the house is done I\u2019ll probably just keep building trains and cars for a while, rather than finishing the accessory set.\n\n-- Charlie M. &quot;Woodworking - patience = firewood&quot;]" time="0.017"><properties><property name="score" value="0.07309990664137178" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07309991&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07309991
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[To a certain extent, the goal of the Trump administration's trade policy, as explained in a speech given by Commerce Secretary Wilbur Ross on March 1, is the creation of a &quot;reciprocal&quot; or &quot;two-way&quot; trading system in which all nations treat each other equally. While not naming China, the intended target of that policy was clear. The Chinese government, on the other hand, was not impressed.\n\nWriting in the English-language edition of China Daily on Monday, Ma Kai, vice-premier of the State Council, or China's cabinet, and chair of the National Development and Reform Commission (NDRC), countered the Trump administration's position by arguing that China has been the most open to foreign investment in the world for the past three decades.\n\nIn Ma's view, China is playing by the rules of free trade. The problem is that the rules that America wants are not the same rules that America has been playing by. The United States wants to be able to sell its goods and services into China, but it wants to block Chinese goods and services from being sold in America.\n\nSpecifically, Ma pointed to the issue of the U.S. $347 billion trade deficit with China, which is the focus of a U.S. investigation to determine whether China's government has unfairly subsidized certain industries.\n\nBut China is not subsidizing its industries, according to Ma. The U.S. trade deficit is simply a function of America's industrial overcapacity and under-consumption, as well as Chinese market and regulatory restrictions on certain industries.\n\nIn Ma's view, there is an easy solution to the trade deficit, and that is to allow China to import more agricultural products, textiles, and auto parts from the United States.\n\nThis is not the first time Ma has been forced to explain the Chinese position in the ongoing trade war. Last week, he held an hour-long interview with Charlie Rose, the television host. In the interview, Ma also said that he would not criticize the Trump administration for the position it is taking. Instead, he is simply trying to point out to the Trump administration that it is China that has been working toward a free trade regime, and that the American government has not been as open as the Chinese government has been in allowing Chinese companies to invest in American businesses.\n\nWhat Ma is really doing is not simply trying to engage in a debate about the nature of free trade, but to make sure that Americans understand how the Chinese government is fighting to protect its own economy from the American tariffs, while at the same time, maintaining a narrative of the mutual benefits of free trade.\n\nMa's interview with Charlie Rose can be viewed here:]" time="0.022"><properties><property name="score" value="0.014255589967563154" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01425559&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01425559
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[WHY PEOPLE KEEP COMING BACK\n\nMany people come back again and again to our Salon. Why?\n\nThey come back because they love the salon experience. They are getting a quality cut and colour, a relaxing shampoo and spa treatment, and an enjoyable, friendly service. They come back because they know they will leave feeling good. Our customers tell us how they love our staff, our warm atmosphere, our extra touches, and our positive, welcoming approach.]" time="0.013"><properties><property name="score" value="0.030964777897929396" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03096478&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03096478
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Selected Test Results\n\nWhat works\n\nInstaller; not required to finish installation of wine for use in the installer. Installs just fine on linux-gnu, after adding 32-bit windows support in winecfg. For some reason the driver doesn't get installed in Windows.\n\nWhat does not\n\nInstallation of the driver in windows\n\nWorkarounds\n\nWhat was not tested\n\nFull installation of windows; only tested enough to install drivers\n\nHardware tested\n\nGraphics:\n\nGPU:\n\nDriver:\n\nAdditional Comments\n\nThis is not a great test since the windows installation is a mess, but it seems to install the driver properly on wine.]" time="0.011"><properties><property name="score" value="0.028771158999113533" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02877116&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02877116
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[What can I say about Sweet Lou's Bakeshop and Cafe in L.A.'s Chinatown other than, well, it is a little gem. If I had to choose one thing to eat every day for the rest of my life it would be a quiche from Sweet Lou's. Seriously. When I went to the Farmers' Market on Monday I bought one. I bought a quiche at the bakery next door too but I'm not too sure how good that one is. But I am very sure that the quiche from Sweet Lou's is absolutely delicious. And it's made from scratch.\n\n\n\nLou is such a sweetie. I know he'd be a fun guy to have coffee with. He owns a gorgeous old school bakery in Chinatown and the smells of fresh baked bread and sweet treats are incredible. So if you are ever in L.A., and you happen to be in the Chinatown area, go on over and get a treat from Lou. You'll be glad you did. And he even has a menu for the cafe part of his establishment and everything on it looks pretty good too. I need to get over there soon and try it out.\n\n\n\nHis website is a little hard to find, I'm not sure if that's just me or if it's hard for everyone. But I did find it.\n\n\n\nI'm a big fan of Lou's pies. They look good enough to eat!]" time="0.013"><properties><property name="score" value="0.08525940299978219" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0852594&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0852594
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The popular Excel shortcut keys are Ctrl + C to copy and Ctrl + V to paste, Ctrl + B to bold, Ctrl + I to italicize and Ctrl + U to underline. Ctrl + M activates the Enter key. Pressing the F1 key opens the help file for the active program. The Windows key launches the Start menu and the arrow keys highlight and select items on the screen. For many other common key combinations, check out www.usingtech.com/excel.\n\nWhen you're using Excel's screen controls, it helps to know the shortcut key combinations. In addition to copying, pasting and formatting, the Ctrl key also serves as a modifier key that lets you move the cursor to another area of the screen or to navigate through lists.]" time="0.012"><properties><property name="score" value="0.042159352985721155" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04215935&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04215935
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The G20, the \u201cShangri-La\u201d and the Global Trade War\n\nThe G20 is currently gathering in the Chinese city of Hangzhou. China has been actively preparing for the summit and has invested in the improvement of its infrastructure and public services. However, the relationship between China and the US has deteriorated in the past few weeks. The US has asked the WTO to look into what it describes as unfair treatment of US companies in China. Furthermore, the EU, the US and Japan have called for retaliatory measures against China, after the announcement of a new Chinese law to allow China to bypass international patents. On the other hand, there are indications that the Chinese government will relax foreign ownership restrictions in its auto industry, and also encourage the outflow of capital from the country. The theme of the summit is &quot;Better Growth, Better World&quot;, and the three themes will be &quot;growth&quot;, &quot;inclusive growth&quot;, and &quot;sustainable growth&quot;.\n\nAccording to a recent report by the World Bank, China has a great economic potential, yet faces severe challenges. These challenges include maintaining the economic growth, the management of urbanisation, the fight against corruption, the increasing inequality and the aging population. On the other hand, the report also concludes that China has the opportunity to become a major contributor to global growth, thanks to the fact that China is an important producer of capital, the country is growing in the manufacturing sector, and is also important in terms of consumption, which is expected to grow by more than 6% this year. The report also highlights the fact that China has been the biggest contributor to the global GDP growth in the last 15 years, with a contribution of 30%.\n\nOn the other hand, China is also a top trading partner of the US, with trade between the two countries amounting to more than $600 billion. Therefore, the potential of a trade war between the US and China would be dramatic. As the leaders of the two countries are gathering in the same city, the G20 summit will be an opportunity to address many of the issues between the two countries. As an example, the summit will discuss the sustainability of trade, both for exports and imports.\n\nIt seems that the US and China may be heading towards a trade war. If such a trade war will occur, it will have devastating consequences on both the US and Chinese economies, as well as on the global economy. Furthermore, the WTO may face a severe problem, as it would probably have to be the court of the two parties involved in the trade war. Thus, the G20 summit will be an important meeting and will represent a key opportunity for the leaders of the two countries to reach an agreement.]" time="0.026"><properties><property name="score" value="0.015499767026799298" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Culture and Tradition\n\nThe Sumerians, who were the first people to build cities, believed]" time="0.011"><properties><property name="score" value="0.000514234643600453" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Juan De La Torre, a 10-year-old artist, designed a colorful new look for the school. He won the contract over five other artists because his design most closely resembled the school's own renderings, reported Milenio.]" time="0.012"><properties><property name="score" value="0.021479853667299986" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02147985&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02147985
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Takes place three months after the events of Season 4, 'Brother's Keeper'.\n\nMax and Liz's lives were put on hold after George's tragic death, but now that the madness of the wedding season has passed, they're finally ready to start their life together. They've settled into their apartment, they're adjusting to working as a team, and even Chompy's finally feeling more like family than just another new roommate. They're finally back to their normal routine.\n\nExcept now, it's not.\n\nNow it's crazy.\n\nMax and Liz are supposed to be getting ready to take their first vacation in a very long time, but it turns out Liz isn't feeling up to it. She has a persistent cough, she's not sleeping well, and she's even started getting headaches. Max tries to be supportive as she talks about calling their doctor, but the more she says it, the more certain he becomes that something is seriously wrong.\n\nAnd then Liz stops talking. She won't talk to the doctor on the phone, she won't talk to Max, and she won't talk to the doctor when he comes to their apartment. She just smiles politely and lies to the doctor's face about how she's been feeling, and then makes Max drive her home as soon as the appointment is over.\n\nMax finally gets her to talk to him once they get home, but she can't tell him what's wrong. She just says that she's been feeling off, and doesn't know why. She refuses to go to the hospital, and eventually falls asleep on the couch, leaving Max alone to think about the phone call he has to make the next day.\n\nHe tells Hank, and his mind immediately goes to the worst case scenario. It doesn't help that when he went to visit Liz earlier, he saw a few of the guys from work lurking around outside their apartment. He had a feeling they weren't just paying him a social call.\n\nHe ends up calling Hank at two in the morning, rambling and stuttering and eventually saying it all in one long string of words. When he's done, Hank doesn't even need to tell him he was right.\n\nThe next day, Max wakes up to find Liz looking better than she has in weeks. There's a huge smile on her face, and she looks so happy. He's thrilled to see her finally feeling better, but he can't help but wonder if she's just hiding the fact that she's sicker than ever.\n\nHe can't bring himself to ask.\n\nShe starts working on her first assignment with a new partner a few days later. This new partner's name is Adam, and he seems like a pretty nice guy. Max watches Liz interact with him and can't help but wonder if he's doing the right thing by keeping her away from the wedding business. He knows how much she loves the job, and this is the first time he's seen her truly smile in a while.\n\nOf course, he doesn't get the chance to mention that to her. The moment he opens his mouth, she cuts him off and says she has a date with Adam.\n\nLiz goes on a date with Adam, and comes home late again. Max's heart sinks when she opens the door and he sees the smile on her face. It's not her usual smile, the one she saves for Max. It's the smile she reserves for Chompy and her art, and it's a smile he hasn't seen in a long time.\n\nMax knows what he's supposed to do, and he knows what's best for Liz. He knows that keeping her away from the wedding business is the best thing he can do for her. But he can't bring himself to do it.\n\nHe still isn't sure how she's going to take it when he tells her.\n\n&quot;Oh my god,&quot; she says, and Max's breath catches in his throat.\n\nHe doesn't get a chance to tell her that he's sorry.\n\nShe rushes off the bed and takes the two steps necessary to cross the room. She throws her arms around him, and Max's whole body tenses at the unexpected contact. He can feel her shaking in his arms, and the sound of her sobs echo through his apartment.\n\nShe tells him she loves him, and he tells her that he loves her too. She pulls away, and he can see the tears streaming down her face. She says that she has to go, and he knows he has to tell her what he did.\n\nHe knows she'll be angry with him, and he's not sure how she's going to react.\n\nShe kisses him on the cheek, and then disappears out the door.\n\nHe doesn't go after her.\n\nMax calls Liz the next day, and tells her to come over to his apartment. She's silent the entire drive, and he's not sure what she's going to do when she walks in.\n\n&quot;I'm sorry,&quot; he says, and she freezes in the middle of the living room.\n\n&quot;I have something to tell you,&quot; he says, &quot;and I'm going to need you to promise me that you'll stay calm.&quot;\n\nShe gives him a slow nod, and he knows she's not going to like what he's about to say.\n\n&quot;I called Hank, and I told him that you were sick. That's why you haven't been going to work. I told him that you needed to take some time off, because you had a really serious case of strep throat.&quot;\n\nShe opens her mouth to say something, but he interrupts her and keeps going.\n\n&quot;He didn't want to believe me at first, but I convinced him. I told him that it wasn't a big deal, but he needs to know if you start showing signs of weakness. He's going to call in a few days to check up on you. I told him you wouldn't answer your phone, and that's why he's calling here instead.&quot;\n\nHe sees the expression on her face, and he knows it's time to tell her the truth.\n\n&quot;You're really sick, aren't you?&quot; she asks.\n\nHe nods.\n\n&quot;How sick?&quot; she asks, and he knows this is the moment he's been dreading.\n\n&quot;Very sick,&quot; he says, &quot;It's Hodgkin's lymphoma. And it's probably already in your lungs.&quot;\n\n&quot;When did you find out?&quot; she asks, and Max can't stand to look at her. He turns away, and he's silent for a few moments.\n\n&quot;I knew as soon as we got home,&quot; he says, and she goes completely still.\n\n&quot;Max,&quot; she says, and he turns around.\n\n&quot;I was afraid,&quot; he says.\n\n&quot;You didn't have any right,&quot; she says, and he can see the tears in her eyes.\n\n&quot;It was a bad idea,&quot; he says, &quot;I know it was a bad idea.&quot;\n\n&quot;What about me?&quot; she asks, and he knows what she's asking.\n\n&quot;You have to promise me that you'll stay out of the wedding business for a while,&quot; he says, &quot;I can't risk you getting hurt.&quot;\n\nShe shakes her head.\n\n&quot;You can't stop me from doing this,&quot; she says.\n\n&quot;Liz,&quot; he says, and his voice cracks.\n\n&quot;I'm not giving up on my dream, Max,&quot; she says.\n\n&quot;It's not a dream,&quot; he says, &quot;It's your life. It's your job, and I can't risk you getting hurt.&quot;\n\nShe shakes her head again.\n\n&quot;You can't tell me what to do,&quot; she says.\n\n&quot;I can and I will,&quot; he says.\n\n&quot;We're not a team,&quot; she says, &quot;We're not the Millers.&quot;\n\n&quot;You're not the Miller,&quot; he says.\n\n&quot;No,&quot; she says, &quot;I'm the Keller.&quot;\n\nHe's never seen her like this before, and he realizes she's not going to let him stop her.\n\n&quot;I'm doing this,&quot; she says.\n\n&quot;No, you're not,&quot; he says, and he can see the fire in her eyes.\n\n&quot;Don't tell me no,&quot; she says, &quot;Don't tell me I can't do this.&quot;\n\n&quot;I'm not telling you that,&quot; he says, &quot;You're going to do this, no matter what I say.&quot;\n\nShe shakes her head.\n\n&quot;I can't stop you,&quot; he says, &quot;I know I can't. I'm not going to force you to do anything, but I am going to ask you to think about what you're doing.&quot;\n\nShe just stares at him.\n\n&quot;You're not going to stop, are you?&quot; he asks.\n\nShe shakes her head, and then walks out the door.\n\nLiz calls Max from work a few days later.\n\n&quot;Why did you do it?&quot; she asks.\n\n&quot;I thought I was doing the right thing,&quot; he says.\n\n&quot;I don't want to be a burden to you,&quot; she says.\n\n&quot;I thought you were sick,&quot; he says, &quot;I was just trying to help.&quot;\n\n&quot;I know,&quot; she says, &quot;And I appreciate it. It's just\u2026I'm not really sick. At least, not anymore.&quot;\n]" time="0.071"><properties><property name="score" value="0.0550928835514631" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05509288&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05509288
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[In this section we describe and analyze the effects of different problems on the model performance. In particular we consider the following problems: the serial position effect, the order effect, the semantic (in)sensitivity, and the distance effect.\n\nFigure 6.16 shows an example of text results of three sequences of three sentences. The whole process of filling and querying a WBS model is shown in Figure 6.17.\n\nWe can observe that in all three sequences the first sentence (\u201cA citizen of A is driving a car\u201d) is more frequently retrieved than the second sentence (\u201cA citizen of B is driving a car\u201d). The number of retentions for the first sentence is also higher than the number of retrievals for the second sentence, which can be attributed to the lexical and syntactic differences between these two sentences.\n\nAnother common problem in short sentences is the inter-sentence association: if a WBS model is designed to store sentences and not words, then the grammatical order of the sentences is kept in the model. Thus, the sentences that have the same word but a different order in the two sentences will be retrieved as if they had the same word. Figure 6.18 shows an example of sentences with the same words but a different order. As expected, the same sentences will have a higher number of retentions than the same words in the same order.\n\nThe serial position effect can be observed when words that occur in a specific position in the sentences are more retrieved than the same words that occur in different positions. This effect can be attributed to the influence of the semantic context on the recognition of the words that are at the same position in different sentences, and the words that are not in the same position have a different semantic context that makes them not to be recognized as if they were in the same position. Figure 6.19 shows an example of sentences with the same words in the same order but with a different position in each sentence. In this example, the word \u201cof\u201d in the position 5 of the first sentence is more retrieved than the same word in the position 7 of the second sentence, and the word \u201cof\u201d in the position 3 of the second sentence is more retrieved than the same word in the position 6 of the third sentence. This effect is more noticeable when the word \u201cof\u201d is repeated.\n\nIn this section, we present the results of a search on the Corpus of Contemporary American English (COCA) [44] with the retrieved words and their respective sentences in a sentence. For each sentence, the average of the values of the retrieval frequencies for all the words in the sentence is computed. We show in Figure 6.20 the relative frequencies of the average values for each word-type for all the sentences of the corpus.]" time="0.022"><properties><property name="score" value="0.09981485697689434" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[In two years, after graduating with a master's in fine arts, he would do what he was paid to do: act in a studio's movies.\n\nBut then in the late 1960's, &quot;you couldn't see any kind of independent films,&quot; he said. &quot;We had two major studios, MGM and RKO. The independents didn't get a chance.&quot;\n\nSo he decided to make one of his own. After asking students at the graduate film program of the New York University Film School for help, he gathered friends and shot his first film, &quot;Mr. Mom,&quot; about a laid-off advertising executive who becomes a housewife, in 1970.\n\nThe low-budget, independent film, directed by a woman and starring a cast of talented newcomers, was ahead of its time. Many viewers laughed, he said, at the notion of a housewife as a career woman. Others did not understand the film's subtle humor.\n\nPhoto\n\n&quot;The primary thing people didn't understand was it wasn't supposed to be funny,&quot; he said. &quot;It was dramatic.&quot;\n\nThe film cost $25,000, and he sold it to Warner Brothers for $50,000. &quot;We got drunk for two weeks,&quot; he said. &quot;It was our first film. We were elated.&quot;\n\nHe went on to make six more films as a director and actor. But the money he made was eaten up by the next movie, as well as his divorce, children and a country house in New Hampshire, he said.\n\nNewsletter Sign Up Continue reading the main story Please verify you're not a robot by clicking the box. Invalid email address. Please re-enter. You must select a newsletter to subscribe to. Sign Up You will receive emails containing news content , updates and promotions from The New York Times. You may opt-out at any time. You agree to receive occasional updates and special offers for The New York Times's products and services. Thank you for subscribing. An error has occurred. Please try again later. View all New York Times newsletters.\n\nThen he met Nancy Chase, a writer and former actress who would become his second wife. She had grown up in a suburb of New York City and attended St. Agnes School in the city's West Village and a private high school.\n\nAdvertisement Continue reading the main story\n\nMs. Chase knew nothing about film, he said. But she was interested in Mr. Moore's films, especially &quot;Slackness,&quot; which came out in 1981. She asked if he would like to work on a screenplay she was writing. He agreed.\n\n&quot;I thought this was a possibility to get out of the mess I was in,&quot; he said. &quot;It was a partnership. I had a chance to work on a screenplay with her. And she had a chance to write something. And then we got married.&quot;\n\nThey met in 1983. A year later they married and went on to make several more films.\n\nMr. Moore said that he and Ms. Chase had a very &quot;civilized&quot; divorce after 23 years. &quot;We've been married long enough that we have two children who are married, who are professionals,&quot; he said. &quot;They're both people who are not into the entertainment business. So they're good children.&quot;\n\nMr. Moore also has two other children. He did not see his first wife, Kay, and their children after they divorced in 1973.\n\n&quot;I didn't find it productive to go back into my son's life,&quot; he said. &quot;I didn't want to be a hero and make up. I didn't want to see him have a second set of delusions about me. I've always believed that children should be given credit for being capable of determining who their parents are. I didn't want to be a subject of people's emotions.&quot;\n\nHe recently remarried, and now lives with his wife, Eileen Minnelli, in a three-story, six-bedroom house that sits behind a wrought-iron gate on a leafy residential street in Mill Valley, Calif.\n\nA self-described &quot;indiscriminate reader,&quot; Mr. Moore spends much of his time writing short stories and screenplays. And he has plans for at least two more films. One is about a woman who was institutionalized in the 1960's, after becoming pregnant. The other is about the son of an interracial couple who is raised by his mother's relatives and becomes a neo-Nazi.\n\n&quot;I have a lot of things to say, and I don't have a problem saying them,&quot; he said. &quot;I haven't changed my approach to making films at all.&quot;]" time="0.064"><properties><property name="score" value="0.08341929928725367" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0834193&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0834193
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Jets quarterback Mark Sanchez practiced for the first time this week and appears on track to start Sunday against the Tennessee Titans.\n\nSanchez's right shoulder strain limited him in practice Wednesday, but he took most of the first-team repetitions. He was 15 of 19 passing with no interceptions and no sacks against a second-team defense. He said after practice that he would play against the Titans if the game were today.\n\n&quot;It's great, it's great,&quot; Sanchez said of his arm strength. &quot;The most important thing with an injury is to get the strength back. We're seeing good results of that. ... That's the main thing.&quot;\n\nThe Jets quarterback was injured in the second quarter of Sunday's 24-17 victory over the Baltimore Ravens, and rookie Greg McElroy came off the bench and played well in relief. McElroy completed 13 of 19 passes for 153 yards and a touchdown and was sacked twice.\n\nSanchez said it's not hard to prepare to play in his first game since the injury.\n\n&quot;This is the first week I'm really getting to throw a lot,&quot; he said. &quot;I've got to get back into a rhythm. That's what it's about. ... I'm excited. I'm ready to go.&quot;\n\nThe Jets have to make a decision on their starting quarterback by Tuesday, when players return from their four-day Thanksgiving break. Coach Rex Ryan said it's not a tough call at this point.\n\n&quot;We'll see,&quot; Ryan said. &quot;We've got another day of practice. That's what we've got to look at and see how the situation unfolds.\n\n&quot;If you just look at where he is at this point, he looks good. You look at what Greg did, too. You look at what we did against a good team with a lot of different players.&quot;\n\nRyan said the decision to start Sanchez or McElroy has &quot;absolutely nothing to do with any kind of controversy or anything like that. We've got a good quarterback. We know we have two good quarterbacks. You guys want to make this some big controversy, but that's not the case.&quot;\n\nMcElroy said he is excited for Sanchez and happy for himself, as well.\n\n&quot;When you have two quarterbacks, you're going to get one of them going,&quot; he said. &quot;I'm excited to see Mark back. It's a good thing.&quot;]" time="0.025"><properties><property name="score" value="0.061233517678474916" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06123352&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06123352
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Livingston County\n\nLivingston County is the name of three counties in the United States:\n\n]" time="0.012"><properties><property name="score" value="5.3106439873440436e-05" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Sir Robert Hart, 1st Baronet\n\nSir Robert Hart, 1st Baronet (22 November 1823 \u2013 1 September 1905) was a British diplomat who worked as chief inspector of Chinese Maritime Customs.\n\nHart was born in Fermoy, County Cork, Ireland. His father was a physician, and his mother was the daughter of another doctor. He was educated at Fermoy Grammar School.\n\nHart's father died when he was twelve, leaving the family in financial difficulties. Hart was forced to earn a living and so took a job as a clerk in the Civil Service of the East India Company at a salary of \xa3100 per year. He became fluent in Urdu and Hindi. After six years, he left the service of the East India Company and moved to China.\n\nHe joined the Chinese Imperial Maritime Customs in 1851. He worked at Amoy in Xiamen and at Shanghai. He worked at Fuzhou and Hong Kong. He spent some time in Macau before moving to Canton in 1859. In 1863, he was promoted to Inspector General. Hart was promoted again in 1871 to Chief Inspector, and became Inspector General in 1882. In that position, he was responsible for administering and enforcing the Chinese Maritime Customs' rules and regulations and for conducting negotiations with other governments. He negotiated many new treaties with foreign powers. He retired from that position in 1897. He died in London in 1905.\n\nThe famous novel &quot;Lost Horizon&quot; by James Hilton has a British consul who is modeled after Hart.\n\nHart was created a Baronet, of Claremont in the County of Cork, on 10 August 1899. He received several honours and decorations from foreign governments, including the Grand Cross of the Order of the Bath from the British government, the Grand Cross of the Order of Leopold and the Legion of Honour from the Belgian government, and the First Class Order of the Double Dragon from the Chinese government.\n\nOn his death, he was succeeded as baronet by his eldest son Sir Ronald Hugh Stewart Hart, 2nd Baronet.\n\n\n]" time="0.024"><properties><property name="score" value="0.07627682360626342" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07627682&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07627682
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Chris Van Gorder\n\nChris Van Gorder (born March 16, 1955) is the former head coach for the women's basketball team at the University of Cincinnati. Van Gorder coached at Ohio University from 1985 to 1996, where he posted a 204-163 record.\n\nVan Gorder was the head coach at the University of Akron for ten seasons. During his tenure, he compiled a record of 157-108. Van Gorder was the first head coach of the Akron Zips women's basketball team.\n\nIn 2005, he was named Mid-American Conference Coach of the Year after his Zips finished 22-9 and went to the NCAA Tournament.\n\nIn the 2006-2007 season, the Zips started the season ranked 12th nationally but dropped out of the rankings following a 65-67 home loss to Western Carolina. Despite the loss to Western Carolina, Van Gorder's Zips won the Mid-American Conference regular season title and the Mid-American Conference Tournament. Van Gorder's Zips then lost to Tennessee in the first round of the NCAA Tournament.\n\nIn the 2007-2008 season, Van Gorder's Zips were not ranked to start the season. But after two early victories against ranked opponents, Van Gorder's Zips were ranked 25th in the nation. Van Gorder's Zips remained in the top 25 for a majority of the season and won the Mid-American Conference regular season championship for a second consecutive year. After a first-round loss in the NCAA Tournament, Van Gorder was named Mid-American Conference Coach of the Year for the second time in three seasons.\n\nVan Gorder was let go at the end of the 2012 season and hired by the University of Cincinnati on April 5, 2013. He resigned on May 4, 2015.\n\nVan Gorder is married to the former Deborah Beisecker, a native of Knoxville, Tennessee. The couple has a daughter, Jennifer, and two sons, Cody and Collin.\n]" time="0.021"><properties><property name="score" value="0.0392213689335129" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Streptomyces durhamensis\n\nStreptomyces durhamensis is a bacterium species from the genus of &quot;Streptomyces&quot;. &quot;Streptomyces durhamensis&quot; produces shikimic acid, paclitaxel, antimicrobial peptides and tyrosinase.\n\n\n\n]" time="0.011"><properties><property name="score" value="0.01562473450205104" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01562473&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01562473
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Ammi majus\n\nAmmi majus, common names greater ammi, gousle, onion grass, greater ammi, cumin-vine, is a plant belonging to the family Apiaceae. It is endemic to Central America, South America, Mexico, West Indies, and Florida in the United States. It is cultivated as a vegetable and spice in other parts of the world.\n\n&quot;Ammi majus&quot; is an annual or biennial plant that reaches heights of and has weak stems. The leaves are thin, simple, and alternate. The flowers are white or pale yellow and borne in clusters of 2 to 4. The fruits are cylindrical capsules with five angles.\n\n&quot;Ammi majus&quot; is used as a vegetable in the Caribbean, where it is known as &quot;gousle&quot;, and in Southeast Asia. It has been found to contain goitrogenic compounds which may be problematic for people with thyroid disease. The whole plant can be consumed, but is most often eaten only the tender leaves.\n]" time="0.012"><properties><property name="score" value="0.06669353215247864" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06669353&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06669353
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[2010 Spanish Grand Prix\n\nThe 2010 Spanish Grand Prix (formally the XLV Gran Premio de Espa\xf1a Telef\xf3nica) was a Formula One motor race held on 28 May 2010 at the Circuit de Catalunya in Barcelona, Spain. It was the seventh round of the 2010 Formula One season. The 60-lap race was won by Red Bull driver Mark Webber, after he started from second position. Lewis Hamilton finished second for the McLaren team and Ferrari driver Fernando Alonso came in third.\n\nPirelli's prime tyre compound was the hard &quot;primes&quot;, whilst the option tyre was the soft &quot;option&quot; compound.\n\nMark Webber took pole position for Red Bull, his first since the 2009 Belgian Grand Prix, while Fernando Alonso set the fastest time in Q2 and qualified second. The two Ferraris filled the next row with Felipe Massa ahead of Kimi R\xe4ikk\xf6nen. The two McLarens of Lewis Hamilton and Jenson Button were fifth and sixth, with Nico Rosberg in seventh place, having struggled all weekend with the handling of his Mercedes. Adrian Sutil in the Force India was eighth, followed by Sebastian Vettel in the second Red Bull and Nico H\xfclkenberg in the Williams. Paul di Resta, S\xe9bastien Buemi and Michael Schumacher rounded out the top ten. At the start, Mark Webber led away from Fernando Alonso and Felipe Massa while the two McLarens of Hamilton and Jenson Button slipped backwards, with Hamilton eventually being forced to pit with an electrical problem, and Button dropping to eleventh. Mark Webber's race soon went wrong as he came into the pit lane to switch to his wet weather tyres on lap 11, but was told to remain in his pit box as he had not switched off his engine, having forgotten to do so. This meant he was stationary for 23 seconds and dropped to fourth, before recovering to third by the time he rejoined the circuit. He also had to serve a drive-through penalty for passing Alonso on the grass, although he eventually finished fourth.\n\nAlonso and Massa pitted on laps 18 and 22 respectively, rejoining in second and third places behind the McLaren of Lewis Hamilton, who had made an early pit stop for dry tyres and was running a different strategy to the two Ferraris. With the track still wet, Hamilton, Massa and Alonso all switched to the intermediate tyres on lap 25, but at this point, the gap between the top three was only 1.3 seconds. The McLarens of Hamilton and Button were then unable to pass the Ferraris on the drying track and they began to slip backwards, eventually losing their fourth and fifth places to the two Force Indias of Adrian Sutil and Paul di Resta. The latter had passed Sutil on the main straight on lap 30 after the Force India driver had made a pit stop and had a clear track in front of him, and held on to fifth place despite being passed by Sutil a few laps later. The two Toro Rossos of Jaime Alguersuari and S\xe9bastien Buemi were the first of the lead lap cars to pit on lap 28, with Alguersuari dropping behind Rosberg in the standings after pitting earlier than the Mercedes driver. The other Toro Rosso of Sebastien Buemi managed to stay out longer and maintained his eighth place, before making a pit stop at the end of lap 34.\n\nFurther back, Jenson Button was the first of the two McLarens to pit for new tyres on lap 33, with a nine-second stop due to an electronics problem with the wheel gun which meant the front left tyre was not fitted correctly. Team-mate Lewis Hamilton followed suit a lap later and rejoined the race in twelfth place. Button rejoined the race behind S\xe9bastien Buemi and passed him on lap 36, but was behind Adrian Sutil when he stopped for fresh tyres on lap 38, although Button was able to pass Sutil a few laps later, with Buemi also getting ahead of the Force India. At the end of the 43rd lap, Lewis Hamilton had managed to pass Paul di Resta and move up to eighth place, with Sutil falling to ninth, having been passed by both Buemi and Button. Hamilton's battle with Buemi came to an end on lap 44 as the two McLarens pitted together, with Hamilton getting ahead]" time="0.064"><properties><property name="score" value="0.09640903431397016" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.09640903&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.09640903
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Carthage House is a historic home in Deland, Florida, United States. It is located at 709 South Catherine Street. On July 26, 1990, it was added to the U.S. National Register of Historic Places.\n\n]" time="0.013"><properties><property name="score" value="0.025233245383049496" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02523325&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02523325
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[S\u0103v\u0103disla\n\nS\u0103v\u0103disla () is a commune in Bihor County, northwestern Romania with a population of 2,099 people. It is composed of four villages: Aluni\u0219u (&quot;Szalonyos&quot;), Ciutura (&quot;Sztutura&quot;), S\u0103v\u0103disla and Socet (&quot;Szoket&quot;).\n]" time="0.014"><properties><property name="score" value="0.04062879766231209" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0406288&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0406288
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Agelasta cognata\n\nAgelasta cognata is a species of beetle in the family Cerambycidae. It was described by Stephan von Breuning in 1935. It is known from the Philippines.\n]" time="0.012"><properties><property name="score" value="0.01720639763355125" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0172064&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0172064
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The Don's Analyst\n\nThe Don's Analyst (, , ) is a 1968 Italian-French comedy film directed by Sergio Corbucci.\n\nDon Salvatore's sister wants to marry Don Vincenzo's son, which makes the two crime families form a union. But Don Salvatore's son is in love with the Don's brother-in-law's daughter, which makes the two families' alliance a difficult one.\n\n\n]" time="0.013"><properties><property name="score" value="0.034340207119321065" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03434021&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03434021
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[1741 in Norway\n\nEvents in the year 1741 in Norway.\n\n\n\n]" time="0.012"><properties><property name="score" value="0.006995714288471344" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00699571&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00699571
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[1997 in Spain\n\nEvents in the year 1997 in Spain.\n\n\n\n\n]" time="0.012"><properties><property name="score" value="0.002907961775699963" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00290796&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00290796
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Blitz (name)\n\nBlitz is a surname of German origin and may refer to:\n\n\n]" time="0.014"><properties><property name="score" value="0.009633253923224672" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00963325&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00963325
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Kurnatowice\n\nKurnatowice is a village in the administrative district of Gmina Lipsko, within Lipsko County, Masovian Voivodeship, in east-central Poland. It lies approximately south-west of Lipsko and south of Warsaw.\n\n&lt;br&gt;\n]" time="0.012"><properties><property name="score" value="0.04463726153769443" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04463726&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04463726
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Jane's All the World's Aircraft\n\nJane's All the World's Aircraft is an annual publication on the subject of civil and military aircraft.\n\nIt was originally produced by Charles and John Whitley in 1909. Since 1962, it has been published by Jane's Information Group. It is published in English and French editions.\n\nAt the time of writing (June 2018), the 2018 edition is in its 75th year of publication.\n\n]" time="0.013"><properties><property name="score" value="0.028087939412890783" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02808794&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02808794
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Back and forth\n\nBack and forth may refer to:\n\n\n\n\n\n\n]" time="0.012"><properties><property name="score" value="0.00020412655491464093" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00020413&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00020413
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Eucosma plurivittella\n\nEucosma plurivittella is a species of moth of the Tortricidae family. It is found in Russia (Amur, Ussuri, Sakhalin, Primorye, Khabarovsk, the Kuriles), China (Heilongjiang, Beijing, Jilin, Liaoning, Zhejiang, Hunan, Yunnan), Korea, Japan, Mongolia and Europe, where it has been recorded from Spain, Germany, Austria, the Czech Republic, Hungary, Slovakia, Poland, Latvia, Finland, Ukraine and Romania. The habitat consists of moist meadows.\n\nThe wingspan is 12\u201317\xa0mm. Adults are on wing from May to August.\n\nThe larvae feed on &quot;Rumex japonicus&quot; and &quot;Polygonum&quot; species.\n\n]" time="0.013"><properties><property name="score" value="0.12510405761070875" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.12510406&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.12510406
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Cluster switching\n\nCluster switching is a migration strategy for cloud computing, in which the users or services in a running system are transparently moved to a new cluster without disrupting their services or performance. It was first introduced in a paper presented at the 2004 IEEE International Conference on Cluster Computing (CLUSTER 2004) by the same name, and published in Cluster Computing in 2006.\n\nThe purpose of cluster switching is to maintain system services even when a node or the entire cluster fails. Cluster switching is often used in parallel or distributed systems where the nodes are not permanently connected to each other. In this case, it is possible to have one node fail without impacting the others. The challenge with cluster switching is that it requires the system to transparently move the workloads to other nodes without affecting their performance.\n\n]" time="0.012"><properties><property name="score" value="0.02629653874743898" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Lozada\n\nLozada is a surname. Notable people with the surname include:\n\n]" time="0.012"><properties><property name="score" value="0.008710523904631451" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00871052&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00871052
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[2015\u201316 UEFA Women's Champions League\n\nThe 2015\u201316 UEFA Women's Champions League was the 12th edition of the European women's championship for football clubs. The final was played between Frankfurt of Germany and Lyon of France, which Frankfurt won on a penalty shoot-out following a 1\u20131 draw. The win was the first time a German team had won the competition.\n\nGerman teams won both the main round groups, and Frankfurt also won the play-off round group, while German and French teams dominated the knockout phase. Frankfurt defeated Arsenal 3\u20131 in the final, while Paris Saint-Germain defeated Barcelona 3\u20132 in the match for third place.\n\nStarting from this season, the following changes were introduced:\n\nThe allocation of teams into qualifying and qualification is as follows:\n\nUEFA has scheduled the competition as follows (all draws are held at the UEFA headquarters in Nyon, Switzerland).\n\nThe schedule of the competition is as follows (all draws are held at the UEFA headquarters in Nyon, Switzerland).\n\nThe draw for the qualifying round was held on 22 June 2015. The first legs were played on 27 and 28 August, and the second legs were played on 3 and 4 September 2015.\n\nThe first legs were played on 17 and 18 September, and the second legs were played on 24 and 25 September 2015.\n\nThe first legs were played on 1 and 2 October, and the second legs were played on 22 and 23 October 2015.\n\nThe first leg matches were played on 6 and 7 November, and the second legs were played on 20 November 2015.\n\nThe first leg matches were played on 1 and 2 December, and the second legs were played on 13 December 2015.\n\nThe first leg matches were played on 17 and 18 January 2016, and the second legs were played on 24 and 25 January 2016.\n\nThe draw for the round of 16 was held on 27 January 2016. The first legs were played on 7 and 8 March, and the second legs were played on 14 and 15 March 2016.\n\nThe first leg matches were played on 7 and 8 April, and the second legs were played on 14 and 15 April 2016.\n\nThe first leg matches were played on 12 May, and the second legs were played on 19 May 2016.\n\nThe first leg matches were played on 25 and 26 September, and the second legs were played on 2 and 3 October 2016.\n\nThe final was played on 15 May 2016 at the Stadion im Borussia-Park, M\xf6nchengladbach, Germany.\n]" time="0.023"><properties><property name="score" value="0.0775850905594565" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Kawanishi N1K\n\nThe Kawanishi N1K &quot;Ky\u014df\u016b&quot; (\u5f37\u98a8, &quot;Strong Wind&quot;) was a long-range single-seat carrier fighter aircraft, serving the Imperial Japanese Navy Air Service during World War II. It was a single-engined, single-seat monoplane fighter designed to replace the A6M Zero, which was designed for shorter ranges.\n\nThe N1K was developed to replace the Nakajima A6M Zero in the long-range fighter role, and was the first fighter in the world to use the Type 3 &quot;Hakury\u016b&quot; 18-cylinder air-cooled radial engine, which was also used in the Kawanishi N1K1-J &quot;Shiden&quot; fighter. The Japanese Navy also gave it the name &quot;Kyofu&quot; (meaning &quot;Strong Wind&quot;) as the manufacturer's symbol, the first time a name was given to a Japanese military aircraft.\n\nIn the end, only a handful of N1K1 fighters (armed with two 20\xa0mm cannon and two 12.7\xa0mm machine guns) were deployed operationally before the end of the war, and none ever engaged enemy aircraft in combat. After World War II, the Allied Occupation forces in Japan took possession of six N1K1 fighters and four N1K2 fighters. These aircraft were studied and flight tested by the Technical Air Intelligence Unit (TAIU) at Tachikawa. Test flights and subsequent analysis revealed that the N1K2 was the faster of the two types, but its lighter armament was deemed inadequate for combat, and the aircraft was regarded as lacking in manoeuvrability.\n\nFollowing the war, some of the N1K2 aircraft were provided to the USSR, and some were used by the &quot;Taiwanese Air Force&quot; (ROCAF), until all remaining aircraft were destroyed by a typhoon in 1950.\n\n\n\nThe only two remaining airworthy N1K2-J Shiden, on display in Japan, are preserved at the following museums:\n\nThe only remaining N1K1-J is on display at the Planes of Fame Air Museum in Chino, California. The airframe was recovered from the ocean in 1994 by Paul Allen and Don Arndt, restored to static display configuration and placed on display at the museum in 2003. The N1K1-J was removed from display in late 2013 for a full restoration, including flight testing, and returned to display on March 10, 2017.\n\nThe Soviet Union took possession of several of the N1K2-Ja variants and used them as the prototypes for the development of the Mikoyan-Gurevich MiG-15.\n\nTwo Shiden's survive in Taiwan. They are located in the Republic of China Air Force Museum in Taipei.\n\n\n]" time="0.024"><properties><property name="score" value="0.0717859605393829" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07178596&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07178596
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Massagris mohammedi\n\nMassagris mohammedi is a jumping spider species in the genus &quot;Massagris&quot; that lives in Guinea.\n]" time="0.011"><properties><property name="score" value="0.01299416747673754" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01299417&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01299417
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[BLACK SABBATH star Tony Iommi says that he has &quot;no plans&quot; to reunite with Ozzy Osbourne for another album or tour following the release of the group's new album, &quot;13&quot;, in June (13).\n\nThe metal legends are about to launch their final world tour after performing in Latin America this month, but while fans can expect to see Osbourne and Geezer Butler at the shows, Iommi reveals that the band will not attempt to bring bassist Bill Ward back into the fold.\n\nHe tells Rolling Stone magazine, &quot;There's no plans for another album or tour with (Ward). And the bottom line is that I wish him well, but I just can't go back. It's been too long. I just can't do it again.&quot;\n\nHe continues, &quot;I'm not going to say that I'll never work with (Ward) again, but I'm not going to try to work with him again for the sake of the fans. It wouldn't be right. It wouldn't be fair.&quot;\n\nIommi also admits that he is baffled by the criticism Osbourne received for his behaviour during the band's Rock And Roll Hall Of Fame induction in April (13).\n\nHe says, &quot;I know he's had problems over the years with drink and that, but I don't really know. All I know is what he did at the Hall Of Fame, which was absolutely incredible, because it was one of those rare moments when you see somebody and you know they're completely sober. It was really good.&quot;\n\nHe adds, &quot;I really couldn't understand the criticism. But then again, I'm pretty thick-skinned about stuff like that. I can take it.&quot;]" time="0.013"><properties><property name="score" value="0.1096141093985814" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.10961411&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.10961411
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[&quot;When you have a lot of money and power, sometimes you think you can buy yourself out of anything.&quot; \u2015Ben Urich [src]\n\nJohn Bushmaster was a powerful crime boss and drug dealer, having inherited control of Harlem's Paradise after his brother, Quincy, was shot and killed by Luke Cage. Bushmaster was also the head of the ruthless Bushmaster drug cartel.\n\nContents show]\n\nBiography\n\nEarly Life\n\nNot much is known about John Bushmaster's early life, except that he was once involved in organized crime. He became a drug lord and took over his brother's position as head of Harlem's Paradise. He later became involved in the criminal organization known as the Hand.[1]\n\nLuke Cage's Power\n\n&quot;Tell me where they took Misty Knight.&quot;\n\n&quot;What are you going to do with me?&quot;\n\n&quot;Well, I'm not going to let you walk away.&quot;\n\n&quot;Who's going to stop me?&quot;\n\n&quot;Luke Cage.&quot; \u2015Luke Cage and Bushmaster [src]\n\nWhen Bushmaster witnessed Luke Cage throw a man out of Harlem's Paradise for harassing some young women, he asked his brother, Quincy, to help him get rid of the unruly patron. The next day, Bushmaster called the man who was threatening to sue Harlem's Paradise and threatened to kill him. When Cage returned, Bushmaster tried to persuade Cage to leave by telling him he did not belong there. Bushmaster revealed to Cage that his name was John Bushmaster and that he was the owner of Harlem's Paradise. When Cage insisted that he would be staying, Bushmaster noted that he did not know who he was messing with before Cage told him he was messing with Luke Cage.\n\nCage remained in Harlem's Paradise and caused problems for Bushmaster, much to his annoyance. Bushmaster and his men began selling the Bushmaster drug throughout Harlem. Cage overheard Bushmaster's name when it was said by his brother Quincy during a phone call. When Cage visited Bushmaster at Harlem's Paradise and accused him of being behind the Bushmaster drug, he denied it and told Cage to leave. Bushmaster's men, however, confronted Cage and attacked him. Cage easily overpowered all of them and escaped from Harlem's Paradise.[1]\n\nChasing Misty Knight\n\n&quot;I'm getting sick of your bullshit.&quot;\n\n&quot;Is that what you tell all the girls?&quot; \u2015Bushmaster and Misty Knight [src]\n\nWhen Cage attacked one of Bushmaster's drug warehouses and threatened to kill Bushmaster if he did not tell him where he was keeping Misty Knight, Bushmaster told him that he did not know what he was talking about and Cage attacked him. Bushmaster ordered his men to shoot Cage while he escaped. The bullets did nothing to Cage as they could not penetrate his skin.\n\nBushmaster visited Pop's Barber Shop and threatened to burn down Pop's Barber Shop and kill everyone inside if he did not get a haircut, including the young girl named Tilda Johnson, who could do a mean cornrows, before he left. Tilda snuck out of the shop and stole Bushmaster's car while he was inside getting a shave. Bushmaster ran outside and found Tilda attempting to get away. He then chased after Tilda in the stolen car.\n\nTilda was able to evade Bushmaster's pursuit until he was able to push her car off the road and caused her car to roll over and crash. Bushmaster dragged Tilda from her car and found out who she was. Tilda, however, was rescued by a bystander who had witnessed the entire incident and called the police. Bushmaster was arrested for kidnapping, but he was released from police custody later.[2]\n\nBusiness Proposition\n\n&quot;Now, we are making money, yes, but not as much as we should. It's a competitive market, man. You're good. And that girl, well, she's something else. I'm offering you a piece of this action. 25% for the both of you.&quot;\n\n&quot;That's not how this works.&quot;\n\n&quot;You know what? It is how this works. It's time for you to grow up.&quot; \u2015Bushmaster and Shades [src]\n\nBack at Harlem's Paradise, Bushmaster offered Shades a percentage of his drug business, but Shades refused. Bushmaster made an offer to Diamondback to sell him some of his product in exchange for a fee. Bushmaster met with Shades, who berated him for not being a better businessman and selling to Diamondback. Shades told Bushmaster that he was selling drugs on his own, but Bushmaster told him that he had a plan to sell his drugs in Bushmaster branded bullets. Bushmaster then attempted to seduce Shades' wife, Tilda Johnson, but she refused his advances.\n\nThe next day, Bushmaster gave a speech to the patrons of Harlem's Paradise as they celebrated the end of their evening. Bushmaster boasted about his power and demanded that they bow to him, but was met with little response. Shades then insulted Bushmaster, who then drew his gun on Shades. Luke Cage then attacked Bushmaster, but Bushmaster managed to get the upper hand and beat Cage into submission. As Bushmaster prepared to shoot Cage, Tilda Johnson shot Bushmaster in the chest with a concealed gun and stole his gun, with Shades shooting the other members of Bushmaster's entourage before they could react.\n\nBushmaster and his men went into hiding at a warehouse, where Bushmaster furiously scolded Shades for his betrayal. Shades explained that he had been working with Diamondback, but now wanted to make a better life for himself and his wife. Bushmaster, realizing that Shades had a point, agreed to go into hiding for the night. The next day, Bushmaster was visited by Knight, Shades, and Cage. Knight had returned to take back Harlem's Paradise. Knight gave Bushmaster a chance to surrender, but Bushmaster refused and attempted to shoot Knight, but was overpowered by Cage and Knight, forcing him to surrender.\n\nWhile Knight and Cage prepared to transport Bushmaster to Ryker's Island, Bushmaster told them that the shipment of drugs would be arriving that night. The police tried to stop the convoy, but the SWAT team was gunned down by Bushmaster's men. Knight decided to attack the warehouse to stop the shipment from getting to New York. Knight, Cage, and Bushmaster fought his men while Tilda Johnson fought Diamondback and his men, although Diamondback managed to escape from the scene.\n\nOnce Knight and Cage subdued the rest of Bushmaster's men, Knight interrogated Bushmaster about where the shipment was. Bushmaster told them that the guns had already been sold to a man named Ricky. Knight and Cage traveled to a diner to look for Ricky, where they found Ricky talking to his partner Rico. Knight threatened Ricky into giving her the location of the shipment. Once Ricky told Knight the location, she and Cage drove to the location where the guns were being shipped. Knight was able to steal the guns from the men who were shipping them, but one of Bushmaster's men threw a grenade into the truck. Cage managed to catch the grenade and toss it back at Bushmaster's men, but was then hit by the truck, knocking him unconscious. Knight ran Cage to the hospital while Bushmaster's men reclaimed their guns and returned them to Bushmaster.[3]\n\nPersonality\n\nTo be added\n\nAbilities\n\nExpert Combatant : Bushmaster was an expert hand-to-hand combatant, as he was able to easily fight off Luke Cage with little effort.\n\n: Bushmaster was an expert hand-to-hand combatant, as he was able to easily fight off Luke Cage with little effort. Skilled Leader: Bushmaster was a skilled leader, as he was able to rally his men together in order to escape Harlem's Paradise.\n\nEquipment\n\nWeapons\n\n&quot;9-milimeter. Take the cap off, and you got a little pipe bomb right here.&quot; \u2015Bushmaster to Tilda Johnson [src]\n\n9mm Pistol: Bushmaster often used this gun when threatening people. Tilda Johnson]" time="0.074"><properties><property name="score" value="0.030391300110136462" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0303913&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0303913
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[It's been more than a decade since the last major update to Britain's copyright law, and it's high time that the UK brought its copyright law into the 21st century.\n\nYou've probably heard that the Government wants to bring in a new Digital Economy Act, which includes a measure to require all ISP customers to opt out of a default-on filter for adult content. Here's how you can tell them that you're not happy with this proposal, using just your email address.\n\nWho?\n\nIf you're concerned about online porn, you might have already signed a petition on the government's website calling for the default-on filter to be dropped. And in the wake of the controversy around the DEA, you might also have signed a petition demanding that ISPs should automatically block sites known to be involved in the distribution of online child abuse imagery. (Both of those petitions are still open for signatures if you want to join in).\n\nBut now it's time to tell the Government about some of the other things they're doing that are not as sensible, and we'll help you do it.\n\nWhat?\n\nFor a long time now, the UK Government has wanted to bring in a new Digital Economy Act. Last summer, the new law was finally published. The Act is massive, and there's lots to argue with, but it looks like the Government may be preparing to rush it through Parliament without any further debate. That means there's a limited window of opportunity to discuss changes to the draft before it becomes an Act of Parliament, and the Government starts work on enforcing it.\n\nThe draft DEA proposes a number of controversial new laws, including the opt-out default filter that's being discussed so widely. But the Act also proposes lots of other worrying measures, and gives the Government a lot more power to interfere with our online communications. The Act's supporters will tell you that it will help to tackle online infringement, for example. But its many critics argue that it risks stifling innovation in online services, and giving the Government a free hand to censor the internet in other ways.\n\nOn Monday, we told you that you could email your MP to demand a debate on the DEA. Now we've got even more reason to tell the Government that we're not happy with the draft DEA.\n\nThis week, the Pirate Party UK submitted a response to the Department of Business Innovation and Skills (BIS) on the Digital Economy Act, which responds to the Government's consultation on the proposals. We're not the only ones with concerns about the proposals; the Open Rights Group has submitted its own response (pdf) to the Government's consultation, and Index on Censorship has produced a response (pdf).\n\nHere's why we think the Government's proposals are a Bad Thing:\n\nThe proposals for an &quot;Extended Liability for Platforms&quot; could have serious unintended consequences for small internet service providers and independent content sharing websites, making them responsible for policing their services and preventing copyright infringement. We think this is likely to stifle innovation in new services, and also has serious implications for our online privacy.\n\nThe plans for an automatic filtering system to block content that is not classified by the state could lead to a drastic over-blocking of content. It could block more than just &quot;extreme pornography&quot; - it could block access to anything the Government decides is inappropriate. There's no requirement for these filters to be accurate or transparent, so we won't even be able to tell what's been blocked.\n\nUnder the proposals, the Government would also have the power to block websites that it claims are &quot;blatantly&quot; or &quot;unlawfully&quot; infringing copyright. This could be a serious threat to online innovation, and to our online privacy. And again, there's no requirement for these blocks to be accurate or transparent.\n\nThe DEA would make it much easier for the Government to put in place more blocking orders, and extend their scope. The wording of the legislation is vague enough that it could be used to order ISPs to censor content that isn't even illegal. The Government has already said that it wants to use its new powers to stop us from accessing websites that might be used to share copyrighted material.\n\nThe DEA is the biggest threat to our online privacy since the Data Retention and Investigatory Powers Act. If the Government gets its way, then they'll have the power to block websites and track what we're doing online. And they'll be able to do it without even telling us that it's happening.\n\nYou can read the Pirate Party's full response to the DEA consultation here (pdf).\n\nNow we need your help\n\nIf you're concerned about the DEA, you can join us in telling the Government what you think about the proposed Act.\n\nThe Open Rights Group's response (pdf) to the DEA consultation is still open for signatures. Once they've got enough names on the petition, they'll send their response to the Government.\n\nAnd now we've got our own response to the Government's consultation on the Digital Economy Act. Please support our campaign to stop the Act by signing our response here (pdf).\n\nWe'll be sending our response to the Government later this week, and we hope that it will give people the opportunity to stop the Act before it's too late.\n\nThis is the one chance we have to stop the DEA. But we need your help. Please support our campaign to stop the Act.]" time="0.076"><properties><property name="score" value="0.013978888303071093" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01397889&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01397889
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[BevMo Beverage Superstore in California is a chain of beverage and snack retail stores. Its headquarters are located in San Carlos, California. As of May 2015, it has 41 stores in California and Nevada. All the stores offer beer, wine and liquors. They also carry various snacks, health and beauty products, household items and automotive supplies.\n\nStore Manager and Assistant Manager jobs in BevMo are offered in different parts of the state. The company also offers Benefits for its employees and Manager Trainee Program.\n\nGet more information about Manager Trainee Program at http://www.indeed.com/cmp/BevMo/jobs\n\nBevMo Beverage Superstore provides competitive pay, benefits, and other programs to help develop employees' potential and achieve their career goals.\n\nApply today!\n\nIt's fast and easy to apply online for jobs in BevMo Beverage Superstore, at http://www.indeed.com/apply?q=beverage+superstore\n\nFrequently Asked Questions\n\nCan I find the career page for BevMo Beverage Superstore on LinkedIn?\n\nNo, the company doesn't have a LinkedIn page. However, you can find more information about the company at its official website: http://www.beverage.com/About-BevMo/Careers\n\nCan I connect with BevMo on Twitter?\n\nNo, the company doesn't have a Twitter page. You can follow them on Facebook at https://www.facebook.com/Bevmo/\n\nWhat is the number for the customer service center?\n\nBevMo doesn't have a customer service center. However, you can contact the company at 1 (877) 822-6692.\n\nIf you have other questions, comments, or concerns, please leave them in the comments section below.\n\nBest Regards,\n\nIndeed.com]" time="0.034"><properties><property name="score" value="0.021146312923944876" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02114631&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02114631
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The Indigenization policy announced by Zimbabwe\u2019s new president, Emmerson Mnangagwa, in 2017, has raised a lot of questions about the process of nationalisation and the methods that will be used to allow for equal distribution of capital, land, and business interests. It\u2019s a policy that seems to favour the government at the expense of foreign interests and brings into question whether or not the government has a commitment to fighting corruption and seeking justice.\n\nThe new policy has been met with much scrutiny both in Zimbabwe and internationally. Many have expressed concern that the implementation of the policy will be flawed and that it will ultimately lead to a worsening of the economic situation in Zimbabwe. The Zimbabwean opposition party, Movement for Democratic Change (MDC), has expressed its disapproval of the policy, stating that it will further marginalise Zimbabweans who aren\u2019t involved in the government or who support the opposition.\n\nOn the international level, the Indigenisation policy is one of the latest changes to the Zimbabwean government that have raised concern. Since he took office in November 2017, President Mnangagwa has brought the central bank and the national oil company under government control. This has led to further speculation about the political and economic direction that the government intends to take.\n\nIs the policy legal?\n\nThe Indigenisation policy was initially announced in 2010 and later replaced by another policy called the Zimbabwe Agenda for Sustainable Socio-Economic Transformation (Zim-Asset). The Zim-Asset policy, which was issued in 2013, introduced a set of guidelines for the process of indigenisation. It requires that 51% of all Zimbabwean businesses be transferred to the country\u2019s black majority. However, the policy has never been enforced.\n\nThe decision to revive the policy in 2017 has been widely criticised. Human Rights Watch (HRW) issued a statement urging Mnangagwa to repeal the policy. HRW believes that the policy will only exacerbate the poor economic situation in Zimbabwe. The statement also notes that the government has never properly implemented the policy. It states that \u201cthe fact that no businesses have been indigenized in the past six years suggests that it is unworkable, and that it will not contribute to economic development or job creation.\u201d\n\nAs a member of the African Union, Zimbabwe is bound to uphold the African Charter on Human and People\u2019s Rights, which contains specific provisions regarding the protection of private property. These provisions state that the Charter guarantees the rights to private property ownership and the right to participate in the management of natural resources]" time="0.023"><properties><property name="score" value="4.660018203524543e-06" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 4.66e-06&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 4.66e-06
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[We believe that these arrangements will provide the school with sufficient financial and resource security to deliver a quality and innovative curriculum for our children and young people.\n\nThis has been a difficult decision to make. The Governing Body has been unanimous in its support for the proposal.]" time="0.012"><properties><property name="score" value="0.005499158272409543" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00549916&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00549916
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[W naszym roczniku najwa\u017cniejsze s\u0105 sprawy wzgl\u0119dne, dlatego zdarza si\u0119, \u017ce nie doceniamy swoich przyjaci\xf3\u0142 lub ich nie szanujemy. Je\u015bli jednak czujesz si\u0119 s\u0142abo, koniecznie powiedz, czy nie chcesz, by kto\u015b Ci\u0119 pociesza\u0142, czy te\u017c jak najbardziej. Czy\u017cby to by\u0142a mo\u017cliwa pr\xf3ba?\n\nGdy jest si\u0119 zmartwionym, trzeba wyci\u0105ga\u0107 wnioski. Zdecydowanie lepiej jest przyj\u0105\u0107 kogo\u015b do serca, ni\u017c zwalnia\u0107. Mo\u017cesz te\u017c zaufa\u0107 osobie, kt\xf3ra nigdy nie wyrzuca\u0142a z siebie nic niechcianego. Sp\u0119d\u017acie wi\u0119c czas wolny wraz z rodzin\u0105, przyjaci\xf3\u0142mi lub po prostu samemu. To na pewno b\u0119dzie dobry pomys\u0142.]" time="0.011"><properties><property name="score" value="0.18703040637266244" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.18703041&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.18703041
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Why do I love Saav? It's not because they sent me these for review. These are on the top of my head that I will put on at the first sign of sun when I go out this summer. Yes, they look this good on.\n\nMade from 100% organic cotton, the Organic Star Trunk Brief is the perfect summertime style for your perfect beach body. These will feel like you are not wearing anything. The material is super soft and has a lot of stretch. They will last you a long time and are very well constructed. They come in either blue or black, and are available in Small, Medium and Large. They come in a gift box, so they make a great gift too. You can purchase them from www.saavdesigns.com for $40 a pair.]" time="0.011"><properties><property name="score" value="0.08055394477491373" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08055394&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08055394
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[CALCUTTA, March 9\u2014 The daughter of an Indian freedom fighter who died recently in a Calcutta hospital may never find out the cause of her father's death because the hospital has no autopsy facility and the family has no money to send his body to another hospital.\n\nA.B. Ray, a professor of economics at Calcutta University, died last Thursday in Calcutta Medical College Hospital after a long battle with pneumonia.\n\n''I am running from pillar to post to try and get a post mortem,'' said Mrs. Ray, whose husband is also a professor of economics at Calcutta University. ''My father was very much against autopsy and it was his express wish not to have it performed.'' Ray A.B. Ray, 82, was born on the campus of Presidency College, one of the oldest seats of higher education in India. He came from a wealthy and cultured family and had gone abroad to study economics in London.\n\nOn his return he became a prominent figure in Calcutta's intellectual and literary circles, editing newspapers and translating the ''Futurology'' series of lectures by the great French writer Andre Malraux.\n\nHe was an adviser to Indira Gandhi in the early days of her political career. He served in Parliament as a member of the Congress Party until his retirement from active politics in 1976.\n\nHe was admitted to Calcutta Medical College Hospital on Feb. 4, and received treatment in the hospital's intensive care unit. His condition gradually deteriorated and he died last Thursday.\n\nHis daughter, Dr. Suchitra Ray, said she was not at all satisfied with the treatment her father received in the hospital and thought that medical staff had not done their job properly.\n\nShe said that her father had been suffering from pneumonia for three years and that he had been given an injection a few days before he was admitted to the hospital.\n\n''My father was extremely allergic to penicillin,'' she said. ''Injecting him with penicillin at a time when he was in a very bad condition was in my opinion a highly irresponsible act.''\n\nThe doctor in charge of the intensive care unit, Dr. Abhijit Gupta, said that he could not discuss individual cases but that he had had no complaints about the treatment of Mr. Ray.\n\n''We did all we could,'' he said. ''It was very sad. His condition was very grave. It was a heart attack that killed him.'']" time="0.022"><properties><property name="score" value="0.05157738747610188" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05157739&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05157739
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[RIGS Creator Platform Playstation 4\n\nPC Genre Shooter, Sports Release Date 29th October 2016 Price $59.99 (PSVR Only) $19.99\n\nRIGS is an action sports game exclusively for PSVR. It was developed by Guerrilla Cambridge (and published by Sony) and released on October 29th, 2016. It takes place in a futuristic sci-fi setting, where you play as one of a number of different characters. Each character uses a different kind of battle suit, allowing them to take part in 3 different sports. Those sports are essentially the 3 main games. They are called RIGS, Megaton, and Gravity Wars. These games play as though they\u2019re real sports. They play more or less how you\u2019d expect them to. You score points by hitting goals, and your objective is to take the ball, and score it in the goal.\n\nThe first game is RIGS, which is essentially a shooting game. There are multiple game modes. Each mode is the same, except you\u2019re playing different characters, and there\u2019s a different goal for you to hit. RIGS is the most fun game in the package, simply because it\u2019s a FPS, which is something you don\u2019t get a lot of in VR games.\n\nMegaton is a sort of bumper-car football game. It\u2019s the second most fun game in the package, just because it\u2019s an interesting spin on something we all know. It is also a lot easier to master than RIGS.\n\nGravity Wars is a very interesting concept, and very hard to master. The game has no physics, and you have to control your direction by using the stick on the right. It\u2019s very challenging, and once you get the hang of it, it can be fun.\n\nRIGS: Mechanized Combat League is a game that is good fun for those who have VR. There\u2019s nothing else like it. It has an interesting concept, and it makes for an interesting single player experience. The multiplayer is there if you want it, but if you don\u2019t, you can use bots, and they will be enough to let you master the game. It does have a lot of potential, and if Guerrilla Cambridge do a sequel, it will most likely be worth it.\n\nPros:\n\nGreat for VR\n\nVariety\n\nWell made\n\nGood concept\n\nCons:]" time="0.022"><properties><property name="score" value="0.0815304117553379" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08153041&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08153041
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Phantom ~ Opus 4 ~ Raw\n\n~ Chapter One ~\n\nTitle: Phantom ~ Opus 4 ~ Raw\n\nAuthor: Starfyre\n\nBeta: freddie-m\n\nPairing: Eventual MeixLu\n\nSummary: After 2 years of therapy, Liu is released back into the world and he returns to the only place he calls home. Only a few things have changed. He finds himself adjusting to the changes with more difficulty than he ever expected.\n\nGenre: Angst, Romance, Tragedy, Drama, Supernatural\n\nWarnings: none\n\nWord Count: 15,611\n\nAuthor Notes:\n\nA while ago I read a story called &quot;Bloodlust&quot; by Inu'sLurker. It's a beautifully written story and I recommend you all check it out. I loved it so much I was inspired to write my own story of 2 boys on a trip to Shanghai. Since I am in no way the author that Inu'sLurker is, I apologize for my lack of originality. But I thought that perhaps if I wrote about something that happened to me I would find it easier to write a story like this.\n\nAnd what happened to me was that one year, the day before my birthday, my mother and I went to China. Now that I look back on it, it was a very exciting trip, though at the time I was rather sad because I was turning twelve. I had been dreading the day, and it was nice to be with my mom because she always made me feel better.\n\nIt was also an interesting trip because we were staying with my Grandfather for a few weeks. He lived in Shanghai, and he was the one that took us to China in the first place. So it was an adventure for me to be in a place that he lived. I only saw my Grandfather on holidays, and it was a great experience.\n\nI hope you all enjoy this story as much as I enjoyed writing it.\n\n- - - - -\n\nIt was just before eleven in the morning when we arrived at my grandfather's house. We had been traveling for quite some time and I was very excited to see him. He came out of the house to greet us]" time="0.021"><properties><property name="score" value="0.1101103917345044" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.11011039&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.11011039
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Rideshare companies like Uber and Lyft are under increased scrutiny after a self-driving Uber killed a pedestrian last week.\n\nThey\u2019re called autonomous, but are they ready? It turns out not even close.\n\nYou might be surprised to learn that self-driving technology is more or less as old as the automobile itself. In fact, there are a number of cars on the road today with self-driving features, but they are only meant to assist the driver and not take over completely.\n\nOne such system, called Autopilot, was first launched by Tesla in October 2015. The main feature allows the car to drive itself on a highway, but it\u2019s still supposed to be monitored by the driver.\n\nUnfortunately, Tesla and Elon Musk have become a lightning rod for scrutiny after an accident last week that killed a Model X driver using Autopilot. A series of emails have revealed that Tesla has attempted to make it clear that Autopilot is not a \u201cself-driving\u201d feature.\n\nIn a recent letter to shareholders, Musk wrote, \u201cWhen used correctly, it is already significantly safer than a person driving by themselves and it would therefore be morally reprehensible to delay release simply for fear of bad press or some mercantile calculation of legal liability.\u201d\n\nWhat exactly does this mean? For one, this was an admission that the technology was not yet ready for public consumption.\n\nTesla is also being called to answer to whether or not the driver should have known that he was using the Autopilot feature. The company says the driver would have known since he was sitting in the driver\u2019s seat and the car\u2019s steering wheel was not in a locked position.\n\nA photo taken by a witness has been released, showing the driver looking down at something with his hands not on the wheel.\n\nThe automaker has since issued a recall to fix the issue and said it will be updating the Autopilot system to ensure that drivers stay engaged with the vehicle\u2019s operation.\n\nUber said it will be disabling the self-driving feature in its fleet of Volvo XC90s after one of its vehicles ran a red light in San Francisco last week, killing the woman. The car was in self-driving mode and the Uber employee behind the wheel wasn\u2019t able to take control of the vehicle in time.\n\nTesla and Uber are not alone in pursuing self-driving cars. Most automakers have some form of autonomous vehicle in the works, as well as companies like Google and Apple. Tesla is still the only automaker with a Level 5 vehicle that has no steering wheel or pedals, however.\n\nAccording to the National Highway Traffic Safety Administration, Google\u2019s self-driving cars have only been involved in 11 accidents over the course of 1.5 million miles of driving, but most were caused by human error.\n\nGoogle has said that the biggest hurdle for self-driving cars will be the creation of a trust between car owners and their vehicle. To that end, the company has been testing a number of prototypes over the last several years, which they claim have driven millions of miles without causing any accidents.\n\nVolvo has been developing its own version of an autonomous vehicle for some time now, but the company is not going at it alone. They\u2019ve partnered with Uber, which bought some XC90 vehicles from Volvo, to run a fleet of self-driving Volvos for ride-hailing in Pittsburgh, where Uber\u2019s research center is located.\n\nWith the recent incidents involving Tesla and Uber, the questions of what\u2019s right and what\u2019s safe have only been heightened. Tesla is going to fix the problem with its Autopilot system and Uber is pausing its own self-driving testing, but it\u2019s clear that these vehicles are a long way from being ready for mass adoption.\n\nWill the driverless cars revolution ever happen? The answer seems to be a resounding yes, but it won\u2019t be anytime soon.]" time="0.065"><properties><property name="score" value="0.052457056695312004" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05245706&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05245706
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Truly great parties start with quality lighting. Party city and Little Tikes, combine that with other decorations and you will be sure to make your little one's birthday the best party ever. Party Supplies are not just for kids. We have a great selection of adult and kids party supplies.]" time="0.012"><properties><property name="score" value="0.0228513643739936" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02285136&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02285136
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[July 13, 2015\n\nNot so great...\n\nI have this straightener and while it is a nice colour, and heats up fairly quickly, it is not really worth the price. It straightens okay, but not much better than cheaper ones, and it is quite heavy so it feels like it is burning your hair when you use it, which makes it really unpleasant to use. Also, I found it did not stay very straight the day after straightening, so I ended up having to straighten my hair every day instead of twice a week. Overall, I would not recommend this straightener.]" time="0.012"><properties><property name="score" value="0.05080072403395586" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05080072&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05080072
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The actual intent of the changes in the UK\u2019s regulation and taxation of ISAs and pensions has become muddled in translation.\n\nThe UK\u2019s changes in pension taxation may well reduce your contributions.\n\nThe investment management industry\u2019s drive to promote personal pension products as an alternative to ISAs may have got a little over excited.\n\nFirstly, the changes that the UK government has made to the regulation of pensions and ISAs has not in fact reduced the level of tax relief on ISAs. It has however changed the way in which tax relief on contributions to personal pensions is calculated.\n\nThe current situation, whereby the actual value of the tax relief received on the sum of a pension contribution, as opposed to the rate of relief, is included in the calculation of income tax, has led to some confusion.\n\nThere is a belief that the relief that is received on the sum of a pension contribution is reduced to 10 per cent and therefore is not as good as the relief received on ISAs, where the actual value of the relief is not included in the calculation of income tax.\n\nThis misunderstanding may well have led to some people contributing to ISAs rather than pensions in order to maximise tax relief.\n\nIn addition, if the actual value of the tax relief received on the sum of a pension contribution is used as a deduction from the gross amount of the pension contribution then the level of relief has not changed at all. The actual value of the relief received on the sum of a pension contribution in the income tax calculation has increased in line with the amount of gross income that is subject to tax relief.\n\nFor example, a gross contribution of \xa310,000 would result in an actual tax relief of \xa37,500 under the current arrangements (the gross contribution is grossed up by 50 per cent to give the net pension contribution of \xa37,500).\n\nUnder the new arrangements, the gross contribution would remain at \xa310,000, but the actual tax relief received on the sum of a pension contribution would be \xa37,000. The relief on the actual value of the pension contribution is not reduced to 10 per cent.\n\nThe amount of gross income that is subject to tax relief is now \xa340,000. If you earned this amount in income before your pension contributions, then the actual tax relief on the sum of the pension contribution would be \xa330,000, rather than \xa37,500.\n\nWhat has changed is the way in which the actual tax relief received on the sum of a pension contribution is calculated. The net pension contribution (the amount actually deducted from the gross income) will no longer be grossed up to produce the actual tax relief received on the sum of the pension contribution.\n\nTherefore, the actual tax relief received on the sum of a pension contribution will always be lower than the actual tax relief received on the actual value of a pension contribution.\n\nIt is important to note that it is not the rate of relief that is affected. It is the actual value of the relief that is reduced. The rate of relief is not affected.\n\nPension contributions are still exempt from tax, although the amount of income that is excluded from tax is less than it would have been if the relief was given on the actual value of the pension contribution.\n\nSecondly, the changes to the regulation of pensions and ISAs have not increased the level of tax relief on pensions. The actual level of tax relief on pensions has not changed. The maximum level of tax relief is still 40 per cent.\n\nThe regulation of pensions has changed. Under the new rules, the maximum amount that can be paid into a pension pot, subject to tax relief, is \xa31.8 million. The maximum amount that can be paid into an ISA, subject to tax relief, is \xa37,000.\n\nBoth the maximum contribution level of pensions and ISAs has increased, but the maximum tax relief that can be obtained on both types of products is the same.\n\nIf you have already opened a pension plan, then you can only increase the contribution amount by \xa32,880 per year. Any extra contribution beyond the annual limit will not be eligible for tax relief. The full value of the contribution, regardless of how much it is, will have to be paid from your gross income.\n\nThe actual level of tax relief on ISAs has not changed. The maximum amount of ISA contributions that can be paid into an ISA in any tax year is \xa37,000. This has not changed. Any amount that is paid into an ISA in excess of this amount will not be eligible for tax relief. The full value of the contribution, regardless of how much it is, will have to be paid from your gross income.\n\nSo in summary, the actual level of tax relief on ISAs has not changed. The maximum amount that can be paid into an ISA, subject to tax relief, is \xa37,000.\n\nThe actual]" time="0.065"><properties><property name="score" value="0.15966169705165892" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[How has music from Spain been perceived in its country of origin? After World War II, a new generation of composers, nurtured by the European avant-garde, began to write in a new, progressive, experimental vein. The growth of technology, which had already made its mark on the theatre, the visual arts, architecture and the cinema, had also started to make itself felt in the music world.\n\nSpanish composers, like their European colleagues, were attracted to the possibilities offered by the new electronic and experimental instruments, and soon took their place among the most important of the international avant-garde. This programme presents a series of fascinating compositions from this period of dynamic change, providing a fascinating insight into how this new generation of Spanish composers reacted to the environment in which they lived.]" time="0.014"><properties><property name="score" value="0.0014284951347975916" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0014285&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0014285
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[During the first year of life, children change more than they ever have before. At this stage, parents play an important role in influencing their child\u2019s development.\n\nPregnant mothers are able to greatly influence the infant\u2019s life in many ways. During pregnancy, a woman is able to optimize her diet and avoid environmental pollutants. Even while pregnant, it is important for her to stay healthy and happy, as this will in turn make the baby happier. All of this is important in a child\u2019s development.\n\nBelow are five milestones in a baby\u2019s development.\n\nMovement\n\nYour child has been feeling his own kicks and punches for a while now. He may have even kicked his mom and caused her some pain. The child will also be moving his arms and legs in a jerky fashion. As the baby grows, his movements will become more fluid.\n\nChild\u2019s Hand and Finger Development\n\nBabies are able to grab things because their hands are strong. They are also able to curl their fingers into a fist. During this time, the baby may be grasping onto a lot of things that are near.\n\nThe baby may also have strong movements, such as pinching your fingers. Babies begin to explore their world through their hands. They are also able to control their hands and fingers with ease.\n\nSitting\n\nA baby can begin to sit up on his own. This is because he has grown to a point where he is able to do this. The baby may also be able to hold his head up by himself.\n\nBabbling\n\nThe baby may also begin to make sounds that can be referred to as babble. Babies are able to make a lot of sounds and combine them together. This is a part of speech development.\n\nFeeding\n\nYour child may be starting to eat solid food. This may include mashed vegetables, cereal, and baby food. If your baby is too young to eat this type of food, it may not be a good idea to force it. This can cause choking or other complications.\n\nSource: BabyCenter]" time="0.026"><properties><property name="score" value="0.03812826529575906" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03812827&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03812827
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[This lecture introduces the practice of mining associations, relations, and dependencies from massive data sets.\n\nRelated Information\n\nAdditional Resources\n\nThe exercises in the course support the textbook Data Mining and Predictive Analytics in the Biomedical Sciences, by Nicolas Le Nov\xe8re, Yan Liu, Jeffrey Leek, and Balaraman Ravindran.\n\nFor background, also check out Jeffrey Leek's introductory lecture to this course.\n\nThis course is based on the book Data Mining and Predictive Analytics in the Biomedical Sciences, by Nicolas Le Nov\xe8re, Yan Liu, Jeffrey Leek, and Balaraman Ravindran.]" time="0.012"><properties><property name="score" value="0.02421095974716181" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[New !\n\nTo access the complete course listing, please log in as a member of the Medical Imaging &amp; Radiation Oncology Learning Community.\n\nMembership is free.\n\nCME Credit: 0.5 CME Credit\n\nNumber of credits allocated to this activity. The sponsoring society for this activity,\n\n, has certified that the learner may receive up to\n\n0.5 Continuing Medical Education credit(s)\n\nfor completing this activity.\n\nAccreditation Statement:\n\nSteris Corporation is accredited by the Accreditation Council for Continuing Medical Education to provide continuing medical education for physicians.\n\n\n\n\n\nThe Medical Imaging &amp; Radiation Oncology Learning Community is accredited by the Accreditation Council for Continuing Medical Education to provide continuing medical education for physicians.]" time="0.012"><properties><property name="score" value="0.06262576957805632" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[&quot; Very nice high quality item, exactly what I expected, and quick delivery. &quot; Very nice high quality item, exactly what I expected, and quick delivery.\n\n~Brian S., Australia\n\n&quot; It arrived. It fits great and looks great. Thanks! &quot; It arrived. It fits great and looks great. Thanks!\n\n~Paul L., San Luis Obispo, CA\n\n&quot; The case arrived today as promised; it fits great! I\u2019m very impressed with the quality of materials and workmanship. I plan on ordering more from you in the future, as I have one other custom amp, and will probably get more vintage ones as time goes by. Thanks again for your assistance and awesome product! &quot; The case arrived today as promised; it fits great! I\u2019m very impressed with the quality of materials and workmanship. I plan on ordering more from you in the future, as I have one other custom amp, and will probably get more vintage ones as time goes by. Thanks again for your assistance and awesome product!\n\n~Mark S., Harford, WI\n\n&quot; I received the covers today....very nice job, they look great! I haven't tried them on the amp yet but I will be for a gig Sat. Thanks for everything! PS, I bought a Pma Board a few months back and ordered one of your covers with it.....nice fit! &quot; I received the covers today....very nice job, they look great! I haven't tried them on the amp yet but I will be for a gig Sat. Thanks for everything! PS, I bought a Pma Board a few months back and ordered one of your covers with it.....nice fit!\n\n~Joe K., Williamson NY\n\n&quot; The slip arrived today and fits perfectly, you hit the nail right on the head. Thank you for your professionalism and look forward to buying one for my 4x12 flextone cab!! &quot; The slip arrived today and fits perfectly, you hit the nail right on the head. Thank you for your professionalism and look forward to buying one for my 4x12 flextone cab!!\n\n~Derek P\n\n&quot; I got the new cases for my pedal steel and they are awesome, they fit perfectly. &quot; I got the new cases for my pedal steel and they are awesome, they fit perfectly.\n\n~Earl F., Houston, TX\n\n&quot; My padded cover arrived today for my Morgan PR12. The quality is superb, and the fit is perfect. It is refreshing to get a product delivered that exceeded my expectations. &quot; My padded cover arrived today for my Morgan PR12. The quality is superb, and the fit is perfect. It is refreshing to get a product delivered that exceeded my expectations.\n\n~Larry G., Saskatoon, Canada\n\n&quot; Hey just wanted to say i got the covers last week and they are awesome!!! i will definitely purchase more in the future.my amp and cabinets are bandmate protected &quot; Hey just wanted to say i got the covers last week and they are awesome!!! i will definitely purchase more in the future.my amp and cabinets are bandmate protected\n\n~Pete G., Nova Scotia CANADA\n\n&quot; The covers arrived yesterday and they are wonderful. The Two Rock &amp; the Princeton covers fit like a glove. Thanks for your patience in getting the measurements right and the quick turnaround on my order! Appreciate it very much! &quot; The covers arrived yesterday and they are wonderful. The Two Rock &amp; the Princeton covers fit like a glove. Thanks for your patience in getting the measurements right and the quick turnaround on my order! Appreciate it very much!\n\n~Mark W\n\n&quot; I just wanted to say a great big THANK YOU for my recent order. I thoroughly am impressed with my new studio slips. &quot; I just wanted to say a great big THANK YOU for my recent order. I thoroughly am impressed with my new studio slips.\n\n~Tim S., Nashville, TN\n\n&quot; The cover fits perfectly and looks great too! &quot; The cover fits perfectly and looks great too!\n\n~Keith R., San Diego CA\n\n&quot; I received the amp cover, you do fabulous work, thank you! &quot; I received the amp cover, you do fabulous work, thank you!\n\n~Michael S., West Chester PA\n\n&quot; I received the amp cover today. Perfect fit and fantastic materials and quality work as always! Thank you for your work! &quot; I received the amp cover today. Perfect fit and fantastic materials and quality work as always! Thank you for your work!\n\n~Mel B., Taunton, MA\n\n&quot; I received my order in perfect condition, the cover fits and looks great ! It was a pleasure doing business with your company. &quot; I received my order in perfect condition, the cover fits and looks great ! It was a pleasure doing business with your company.\n\n~Jerry W., New Berlin WI\n\n&quot; Received the studio slips cover today, perfect fit! Thanks Again for a great product! Great doing business with you again..... &quot; Received the studio slips cover today, perfect fit! Thanks Again for a great product! Great doing business with you again.....\n\n~Joe V., Lindenhurst, NY\n\n&quot; I just wanted to let you know I am thrilled with the new cover. It arrived today and fits perfect. Good job and well done!! &quot; I just wanted to let you know I am thrilled with the new cover. It arrived today and fits perfect. Good job and well done!!\n\n~Kenny M.,]" time="0.066"><properties><property name="score" value="0.1574303617279136" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[For many years, and at many different homes, the grand game of monopoly was a nightly staple of our game playing regimen. Most of the time the game was played with one of my kids as the banker. The game required a bit of accounting to keep track of all the money. More often than not the game was a heated one and everyone, particularly the banker, enjoyed counting out money while money was being piled up all around the board.\n\nMonopoly is an interesting game for a number of reasons.\n\nFirst, it requires the basic accounting function of making change. The cashier has to be able to count out the correct change for all the players. This seems like a very simple, but important function. If a cashier can\u2019t count change, you would have to return the money, and you might not get it back.\n\nAlso, with any money that comes into the game there are some decisions that have to be made. For example, with a loan, how much should be charged to each player. More importantly, how much can the players afford to pay back?\n\nIn addition, how much does the bank loan to the players and what are the terms? What does the bank require to be repaid? At what time? What happens if there is not enough money to pay back the bank?\n\nThis seems like an easy concept, but a lot of my kids had a difficult time trying to figure out what they had to do to get out of debt.\n\nFinally, there are times when players simply get too deep into debt. They are forced to sell off a building, or even the entire game.\n\nThis seems like a pretty simple concept, but I found it interesting because it taught some of my kids a very valuable lesson about life. They realized how important it is to have a budget. More importantly, they realized how much money they had to spend. They had to spend wisely because they couldn\u2019t go into debt.\n\nI think it\u2019s easy to teach kids to balance their checkbook and keep track of their spending, but it\u2019s much more difficult to teach them to spend wisely and stay out of debt.\n\nThe same goes for a family\u2019s bank account. Families, just like the player\u2019s in monopoly, need to keep track of the money coming into the bank account and what is going out. They need to make sure there is enough money to cover the things they need to pay for and that there is enough money to cover the things they want to spend money on. They need to be responsible with their money and realize they can\u2019t go into debt.\n\nThe game of Monopoly teaches a lot of these principles. It is not a game of instant gratification. There are no easy money schemes. It\u2019s a game that will test the mind and the ability to be responsible with money.\n\nIf you have kids, Monopoly is a great way to learn about personal finance. You don\u2019t need to be an expert, but you do need to be willing to show them how to count and account for money.]" time="0.027"><properties><property name="score" value="0.0025921268909908576" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00259213&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00259213
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The Bengals fell just short of a win in Kansas City Sunday, falling to the Chiefs 31-13. The loss drops the Bengals to 7-9, a number that still includes a loss to the Jacksonville Jaguars in London.\n\nOne of the biggest stories to come out of this loss is the fact that star running back Jeremy Hill was benched in favor of Rex Burkhead. Hill didn't have a good game, and the Chiefs were able to put him in his place. Burkhead was also pretty good, and scored the only Bengals touchdown on the day.\n\nBoth running backs ran for over 100 yards, but the key difference was Hill fumbled twice, including once inside the Chiefs' 20. Hill had carried the ball 39 times before the benching, while Burkhead had just nine.\n\nIt was announced Monday that Hill is still the Bengals' starting running back going forward.\n\n#Bengals Jeremy Hill named starting running back in 2018, Mixon to see more action. https://t.co/jZ9624gBwO pic.twitter.com/jOKGewN4zq \u2014 Kevin Clark (@bykevinclark) January 1, 2018\n\nThis is an interesting decision, considering the way the Bengals used the two running backs in Kansas City. Hill has been a solid running back throughout his career, but has shown lapses in concentration, as well as fumbling issues. He's not been consistent, and that has made the Bengals hesitant to hand him the starting job.\n\nThe Bengals will likely be looking to get more out of their running backs next season. We saw how talented Joe Mixon is this year, and he's the one who should have the starting job next season.\n\nHill will still be involved in the offense, but this change is a sign that he won't be a starter for the Bengals next season.]" time="0.024"><properties><property name="score" value="0.07936320654928836" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07936321&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07936321
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[This is the brand new video for \u201cSave Our Souls\u201d from the German melodic rockers CASANOVA. This is the third video release of the band, and you can watch it below.\n\nCASANOVA\u2019s new album, \u201cMonte Carlo\u201d will be out on January 22, 2013 via Escape Music.\n\nCASANOVA\n\nFacebook: http://www.facebook.com/casanovaofficial\n\nWebsite: http://www.casanova.de/\n\nEscape Music: http://www.escape-music.com\n\nDirk]" time="0.013"><properties><property name="score" value="0.05002552868345611" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05002553&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05002553
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[To: SeekAndFind\n\nThose were the good old days. Now we have NDAA, signed by \u201cwho\u2019s in charge now\u201d, saying that if you are a terrorist, you can be shot on sight, anywhere in the world, even on American soil. All you have to do is become a terrorist, and that takes less than one day in this country. It\u2019s like signing your own death warrant. The silver lining is that you might be one of the ones killed before you can become a terrorist.]" time="0.014"><properties><property name="score" value="0.03693891216353107" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03693891&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03693891
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Dr. Mackle and her staff are the best. I am an ex military retiree and have been getting crowns for over 10 years and this was my best experience yet. The entire staff were all friendly and professional. And they were all on time. I felt like I was part of their family. And I really liked their prices. I will use Dr. Mackle in the future. You are all awesome.\n\n\u2013 B.S. / Google]" time="0.012"><properties><property name="score" value="0.027304090386076618" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02730409&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02730409
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Transcript\n\n1 ADNOTAMENTO A FOGLIO NUMERO INDAGINI C.P.C.O. PROFESSIONE DISPONIBILE ORDINE CLINICO ED ORTOPEDICO 1. MEDICO ORTOPEDISTA 1 MEDICO ORTOPEDISTA ORTOTTISTA 1 MEDICO ORTOPEDISTA ORTOPEDISTA 3 MEDICO ORTOPEDISTA PSICOLOGO 1 MEDICO ORTOPEDISTA RIABILITATORE VISO 1 MEDICO ORTOPEDISTA ODONTOIATRA 1 MEDICO ORTOPEDISTA NEUROPSICOLOGO 1 MEDICO ORTOPEDISTA ASSISTENTE SOCIALE 1 MEDICO ORTOPEDISTA PSICHIATRA 1 MEDICO ORTOPEDISTA 1.1 FISIOTERAPISTA 1.1 FISIOTERAPISTA INFERMIERE 1.1 FISIOTERAPISTA OPERATORE SANITARIO 1.1 FISIOTERAPISTA DIETISTA 1.1 FISIOTERAPISTA ASSISTENTE SOCIALE 1.1 FISIOTERAPISTA COACH 1.1 FISIOTERAPISTA ELETTROMIOGRAFO 1.1 FISIOTERAPISTA DENTISTA ORTOPEDICO 1.1 FISIOTERAPISTA INFERMIERE CLINICO 1.1 FISIOTERAPISTA OSS 1.1 FISIOTERAPISTA TERAPISTA DELL AUTOMOBILE 1.1 FISIOTERAPISTA TECNOLOGO DEI MATERIALI 1.1 FISIOTERAPISTA 1.2 ORTOTTISTA 1.2 ORTOTTISTA ASSISTENTE SOCIALE 1.2 ORTOTTISTA ODONTOIATRA 1.2 ORTOTTISTA 1.3 OSTETRICA E GINECOLOGO 1.3 OSTETRICA E GINECOLOGO ASSISTENTE SOCIALE 1.3 OSTETRICA E GINECOLOGO ODONTOIATRA 1.3 OSTETRICA E GINECOLOGO 1.4 PROSTETICO IGIENE ORALE 1.4 PROSTETICO IGIENE ORALE ODONTOIATRA 1.4 PROSTETICO IGIENE ORALE 1.5 RADIOLOGO 1.5 RADIOLOGO ODONTOIATRA 1.5 RADIOLOGO 1.6 REUMATOLOGO 1.6 REUMATOLOGO MEDICO DELLA PREVENZIONE 1.6 REUMATOLOGO 1.7 OSTEOPATA 1.7 OSTEOPATA ASSISTENTE SOCIALE 1.7 OSTEOPATA ODONTOIATRA 1.7 OSTEOPATA\n\n2 1.8 PSICOLOGO 1.8 PSICOLOGO MEDICO DELLA PREVENZIONE 1.8 PSIC]" time="0.013"><properties><property name="score" value="0.1578695702405375" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Will the traditional establishment defend democracy?\n\nBy Alexandra Morton\n\nUBC-O Sea Around Us Project\n\nI am scared to my bones. For the first time in my life I am terrified that what I learned as a child is wrong. There is no more mother or father to turn to. There is no more food to count on. This is my last chance to sound the alarm that we are being systematically robbed and cheated and lied to and I don\u2019t know what we are going to do about it.\n\nA few years ago I could not have imagined that I would find myself writing this. I was born and raised in the village of Old Massett on Haida Gwaii, also known as the Queen Charlotte Islands, in Canada. We fish and hunt and gather food and make our own shelter and build our own community. This is how it has been for thousands of years. In modern language this is called subsistence living. We were the last people in Canada to sign treaties with our colonial oppressors in the 1800s. These documents were the foundation of our way of life.\n\nThen it all changed in the early 2000s when the first commercial aquaculture license was issued on our coasts. At first, like most people on Haida Gwaii, I thought that farming fish would provide income for our community. Little did I know that the promise of wealth would come with a heavy price.\n\nThe first evidence I saw that something was wrong was in 2006. The sea lice were everywhere and a fish farm had started using a pesticide called Slice. I saw the sea lice on the fish and sea lions. I saw sea lions coughing up blood, the bones on their backs visible through their skin.\n\nIn 2007, the open net pen at salmon farm salmon feedlots one kilometre off the coast of my village collapsed. An estimated two to three million Atlantic salmon, with a value of $6.5 million to $8 million, swam into the ocean. There was no monitoring, no analysis of the risk to wild salmon, not a word to local communities.\n\nAnd then the toxins appeared. In 2008 and 2009 people in my village became sick and died. In the village of Klemtu next door, children were born with flipper-like hands and feet. In Norway, which had a higher concentration of fish farms, brain tumours and other cancers were rising and women were giving birth to deformed babies.\n\nI looked at these stories and thought, \u201cThis can\u2019t happen to us.\u201d\n\nThis was followed by the closure of the Haida Nation\u2019s only grocery store. People in my village and community started suffering from serious health problems, especially neurological ones. We were told it was due to influenza or SARS. We had no idea that a toxin, which our bodies have not evolved to handle, was being released into the air and our water. The Sea Around Us reported that fish farms were creating the most toxic area in the world, right next to the First Nations in Haida Gwaii. This happened, and no one told us.\n\nI realized I had to do something. For the first time in my life I realized that what my ancestors had always taught me was wrong. It was not nature that took care of us. It was we who were responsible for taking care of nature.\n\nI wrote my first letter to the local newspaper. It was my story, and I was afraid to publish it.\n\n\u201cHow many people have to get sick and die before they take this seriously?\u201d I asked.\n\nWe waited and waited, and nothing happened. I wrote my second letter. Nothing.\n\nWe waited and waited. Still nothing.\n\nPeople in my village got sick and died, and the Norwegian scientists kept warning that their food supply was being contaminated. We were told that this was the price of progress, that it was our job to clean up the mess. And we had to do it on our own.\n\nAnd then the baby seals came. There are no young sea lions, so the seals started eating the herring. At first we thought it was cute. Then we realized they were killing the herring. There were too many seals.\n\nI wrote my third letter. I looked at the numbers and saw that Norway was killing 80,000 seals a year and there were only 80,000 seals left.\n\nI started to organize rallies and a campaign called Save the Herring, which caught the attention of major news outlets across the globe. By this time I had a voice. I was a professor of biology at UBC and one of the world\u2019s experts on fish farming, open net pen fish farms, and marine pollution. I knew the science.\n\n\u201cHow many baby seals have to get cancer before they take this seriously?\u201d I asked.\n\nIn the midst of this, Canada approved a massive new open net pen farm. This was the largest salmon farm in the world. Our local paper warned us that \u201cthe massive farm will increase marine pollution in the area.\u201d This farm started dumping blood and feces into our waters. The evidence was mounting that we were at risk of a catastrophic collapse in our food web.\n\nBut no one cared. The aquaculture industry is worth $400 million a year. We can\u2019t compete with that.\n\nA new form of ocean zoning called an impact assessment was supposed to keep salmon farms out of sensitive areas, but it was all a lie. The people who wrote it are in the industry. They gave themselves exemptions.\n\nThen I realized that I could make a living as a scientist. So I created a lab at the University of Victoria to study the toxins from fish farms. We started to find the pollutants in the mussels and the herring. I published more and more scientific papers.\n\nWe held more rallies, and I wrote more and more letters to the local paper. We were ignored and dismissed.\n\nSo we went to the Supreme Court. Our case was thrown out, but the judge was extremely critical of the Canadian government and gave them 18 months to fix their laws. But the government is now granting the aquaculture industry a blanket exemption from environmental laws and protections. The judge told them that they could not do this. But they did.\n\nI have studied toxicology for 30 years and I have never seen this. We know that these toxins are not getting into our environment accidentally. The only reason they are there is because someone put them there. It is like they are trying to poison us.\n\nAnd then we started to see a new toxin. This time it was a by-product of the feed used in the salmon pens.\n\nA Norwegian scientist tested a by-product of salmon feed. The test showed that it was toxic. The Norwegian government stopped using the feed. In 2015, that same scientist tested all the salmon feed used in the British Columbia industry. The results showed that there was another new toxin, this time a carcinogen. The Norwegian scientist said, \u201cIt is extremely unlikely that it could happen at such high levels by chance.\u201d\n\n\u201cHow many toxins have to get into our food before we take this seriously?\u201d I asked.\n\nThis time I knew that there was something wrong with my brain. We found out that Norway and Scotland were using the same feed. We had a toxic lab here in British Columbia, so what about Norway and Scotland? And the feed is approved by the Canadian Food Inspection Agency and used in these salmon farms.\n\nIt\u2019s hard to see who we are going to turn to.\n\nThen a woman in Ketch Harbour started asking questions. She wrote to the local paper and was ridiculed and insulted. She wrote to the Minister of Fisheries and Oceans. He never replied.\n\nShe was not from my village. She was a white woman from Newfoundland. Her grandfather was a fisherman.\n\nI wrote to her, and then I realized I could write to everyone. And I did. I started to get hundreds of people to write letters, and we wrote to the Minister of Fisheries and Oceans and to our Premier, the head of our government. We wrote to every MLA, the Members of the Legislative Assembly, and we were ignored and dismissed.\n\nSo we went to court again. This time we were supported by 30 affidavits from fishermen and local citizens.\n\nI went to the hearings. The minister did not even show up. His lawyer said that we should stop questioning him and just accept that he is doing the right thing. He had no explanation as to why the 30 affidavits were not enough evidence to put a stop to this industrial salmon farming.\n\n\u201cIf I had a million dollars, I could save Haida Gwaii,\u201d I said.\n\nBut I have never been able to raise that money. The man who tried to stop me is now running for Premier, and he is the frontrunner.\n\nSo I wrote my fifth letter to the editor. I warned that if we continued to use this feed, we would all get sick and die.\n\nAnd then I realized that I could put this on Facebook.\n\nA few months ago, the BC government ordered the feed company to remove this toxin from their products. But they still sell it in other provinces. There is still no ban on the feed, even though Norway has banned it, the FDA has banned it, and Health Canada has banned it.\n\nAnd then I realized I could just get on a bus and go to Toronto and speak to people directly.\n\nSo that is what I am doing. I am going to make this my life\u2019s work. I am going to stand in front of people, in universities, in public hearings, and tell them my story. I am going to explain that it is not]" time="0.076"><properties><property name="score" value="0.06511083058430915" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06511083&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06511083
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[I am a very fair skinned male and this tanning lotion has worked wonders on my pale body. It was quick and easy to apply. It also was very easy to clean off my hands after I applied it, which was a big plus.]" time="0.013"><properties><property name="score" value="0.019704145734267264" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01970415&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01970415
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Veggie burgers are great! The more veg I add the better I feel about my nutrition. Plus I am not that great at taking vitamin supplements. I do eat all the veggies I need though.\n\nMy first veggie burger was just lentils, eggs, and bread crumbs. No extra seasonings or flavours. They were ok but really needed salt. This recipe is so much better. I used this recipe as a basis but made a few modifications. I used 3 cups cooked quinoa, 2 cups kale, 1 cup spinach, 4 large carrots, 4 celery stalks, 1 onion, and 1 cup corn. I added a tsp. of thyme and 1 tsp. of cayenne pepper. I had a tiny bit of ground pork that I needed to use up so I also added that. I used black beans instead of pinto beans because I prefer them. I added extra bread crumbs and an egg to bind it. I think I put in 3 Tbs. of olive oil as well.\n\nI am happy with the result. The texture is really nice and the flavour is delicious! I served it with some sriracha mayo that I made. YUM!\n\nAdvertisements]" time="0.012"><properties><property name="score" value="0.1002592118510226" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.10025921&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.10025921
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[JUMBLE CONFIDENCE\n\nDefinitions\n\nConfidence: A term describing the likelihood that a decision-making system will reach a good decision. It refers to the quality of a decision, not to the quality of the process that leads to that decision.\n\nJumble: A pattern of naturalistic decision making. It does not include extraneous information or create false information.\n\nExample\n\nA fighter pilot is confident that she will destroy the enemy aircraft even though she has only her own memory of previous aerial encounters to rely on.]" time="0.012"><properties><property name="score" value="0.022819821284630892" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02281982&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02281982
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Anselm and his ontological proof, is one of the most important Christian philosophers of the Middle Ages. He is also well known for his attempts to fuse the sciences with theology, something of a controversial topic for medieval theologians. The ontological argument itself is based upon the idea that God\u2019s nature is that of existence. Therefore, to say that God exists is to say that God has the greatest nature, which in turn leads to the conclusion that God must exist. Some scholars argue that this is the proof\u2019s most fundamental flaw, namely that God\u2019s nature is unprovable and thus, cannot be equated to the definition of existence. Other critics point out that the greatest nature is really a metaphysic and therefore cannot be understood by people. The latter claim has been supported by Immanuel Kant and many other philosophers since. Kant\u2019s work was not known to Anselm\u2019s, but it is clear that the German philosopher would not have found much use in Anselm\u2019s philosophy.\n\nThe foundation of Anselm\u2019s ontological argument is a branch of philosophy known as natural theology. Natural theology is the search for a knowledge of God through the use of reason. The argument\u2019s goal is to prove that God exists in some way, and as such is a concept that is widely accepted in the world of Christian philosophy. In many ways, Anselm\u2019s argument for God\u2019s existence shares many similarities with the arguments of many of his contemporaries. Anselm claims that God exists and that this conclusion can be proven using the method of defining terms. The proof proceeds from the claim that God is the greatest thing conceivable to the claim that God must also exist. Anselm\u2019s argument does not satisfy many people because it does not provide any evidence to support the claim that God is in fact the greatest conceivable thing. This is one of the major weaknesses of Anselm\u2019s argument, but it is also one of the weaknesses of most arguments for the existence of God. The ontological argument can be seen as an attempt to prove God\u2019s existence without supporting any specific philosophical view of God\u2019s nature.\n\nMuch of Anselm\u2019s work focuses on the attempts to explain and prove God\u2019s existence. He was one of the earliest of many philosophers to attempt to explain and prove God\u2019s existence. The debate over the existence of God has continued over the years, and the arguments have been supported and refuted on many different occasions. Anselm\u2019s work is important because he was one of the first philosophers to put forth the idea that existence itself was proof of God. He was also one of the first Christian philosophers to try and explain the idea of God in a non-theological way. Many of Anselm\u2019s contemporaries had rejected his view on God\u2019s nature, and they also criticized his work on natural theology.\n\nOne of the major debates over Anselm\u2019s work was that of the ontological argument. This is a popular concept that states that God\u2019s existence is a self-evident truth. The name of the argument comes from the phrase \u201csomething that is greater than which cannot be conceived\u201d. This concept became widely accepted in the world of Christian philosophy, and it is still being debated today. The concept of God\u2019s existence as a self-evident truth has been supported by many Christian philosophers, including Thomas Aquinas. Aquinas\u2019 view of God was one of the most influential views of God\u2019s nature in the Middle Ages. The ideas presented by Anselm and Aquinas on God\u2019s nature were among the first attempts to view God in a logical way.\n\nMost of Anselm\u2019s work focused on the idea of God and the way in which God was perceived. He focused on the idea that God was perceived through faith and that this idea had a real and substantial foundation. Many critics viewed Anselm as a philosophical follower of Augustine, and he was certainly influenced by many of Augustine\u2019s works. The Confessions of Augustine were written in the fifth century, and they discussed the views of the time on God and the manner in which people should understand God. Anselm\u2019s view of God is therefore quite different from many of his contemporaries, and he was heavily criticized for his work on natural theology.\n\nThe ideas presented by Anselm were extremely influential to his contemporaries. Although his ideas were at times criticized, many of his contemporaries found them to be the most logical views of God. His argument for God\u2019s existence was extremely popular among his peers, but it is one of his most controversial works.\n\nAnselm of Canterbury Biography\n\nAnselm of Canterbury was born in 1033 in Aosta. He was the second of four sons and his parents were wealthy. Anselm\u2019s father was an Italian who owned a noble estate. His mother was the niece of the archbishop of Rheims. Anselm was well educated, and he studied both rhetoric and literature as a young man. He studied law in 1057 at the cathedral school in Laon. During this time, he became close friends with Roscelin of Compi\xe8gne. Roscelin was an expert on logic, and he would be one of Anselm\u2019s best friends and mentors for the rest of his life. He was also close to Lanfranc and eventually became his student.\n\nAnselm studied theology in the early 1060s at the cathedral school of Chartres. He also spent a great deal of time studying theology and grammar at the school of Benevento. Anselm spent two years at the school and was then called to the cathedral school of Canterbury in 1062. Anselm was called to the cathedral school to replace the archbishop. The archbishop of Canterbury was the highest ranking prelate in England and he was charged with the administration of the Church in England. Anselm quickly gained respect in his new position, and he wrote a new rule for the cathedral school. The rule was based on the Rule of St. Benedict, which was the most important rule for monasteries at the time. He was also responsible for the cathedral school\u2019s many other duties and its administration. He established the system of electing canons and was also in charge of its finances. Anselm had a great deal of power in his new position, but he also was a reformer. He placed emphasis on the importance of education and scholarship, and he ensured that the students received high quality instruction. The rule at the cathedral school was seen as being high quality and the students at the school were trained in Greek, Latin, rhetoric, dialectic, and arithmetic.\n\nAnselm also established a library at the school. The library contained many ancient works, which were frequently read by students. He became an expert on the writings of Augustine, and he began his own writing career in the early 1060s. His writings were influential in the development of later philosophy, and his ideas on natural theology were both influential and controversial. In 1063, Anselm was appointed archbishop of Canterbury. This was an extremely important position, and Anselm would play a major role in the Christianization of England. He was a major contributor to the revival of learning in England, and he attempted to bring education to all of his subjects.\n\nHe was a skilled theologian, and he produced several important theological works. These included the Monologion and the Proslogion. The Monologion addressed the questions of the nature of God. The Proslogion addressed the questions of the nature of God\u2019s knowledge and the proof for God\u2019s existence. Anselm was influenced by Augustine in his writing of the Monologion]" time="0.070"><properties><property name="score" value="0.03544778987603392" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Advantage Resourcing Ltd - Northampton\n\nWe are currently recruiting for a Production Operative on behalf of our client based in Northampton. This role is starting immediately on a temporary to permanent basis. If you have experience in Manufacturing then please get in touch. You will be required to carry out your duties in a high quality...]" time="0.015"><properties><property name="score" value="0.03341999287240771" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03341999&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03341999
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The Blastlite Lite BTM (Blastlite Thermal Monitor) is a small, single-output controller that supports a wide range of time-programmable and event-triggered outputs. The time programmable outputs are available as dry contacts (NO/NC) or as solid-state relays.\n\nThermal sensors are available for a wide variety of applications and locations. Select the correct sensor for the task at hand using the Sensor Selector Guide.\n\nThis thermostat was designed to easily mount to the side of any standard 1 1/2 inch (38mm) junction box with the included mount bracket.\n\nBlastlite Lite BTM Features\n\nLED status indicators\n\nEasily programmable for time-triggered and event-triggered outputs\n\nThermal sensors for flame detection and appliance temperature monitoring\n\nBlastlite Lite BTM Specifications]" time="0.013"><properties><property name="score" value="0.0199097529408665" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01990975&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01990975
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[\u201cDark\u201d will be the third in a series of single-pitch sport routes at Boulder Zephyr in Georgetown, Ontario. The sport crag is located a few kilometres west of Blue Mountain, Ontario. Here are the proposed grades and estimated lengths of the routes, which are currently being developed:\n\nDark (5.12a) \u2014 35m\n\nMorning Light (5.11c) \u2014 35m\n\nDawn Patrol (5.11b) \u2014 35m\n\nLunch Break (5.10b) \u2014 35m\n\nVibes (5.10a) \u2014 35m\n\nThough none of these routes are bolted yet, and they are still only in the development stage, Boulder Zephyr is already a great place to go bouldering!\n\nSource: Peter Croft (Bouldering Ontario)]" time="0.014"><properties><property name="score" value="0.054964755067454796" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05496476&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05496476
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[In California, a convicted pedophile is on the loose. The worst part? He's a convicted pedophile.\n\nFrancisco Felix-Jimenez was released from prison last week. He was convicted of molesting an 8-year-old boy in 2006.\n\nWell, it turns out that the Los Angeles County Sheriff's Department forgot to register Felix-Jimenez as a sex offender.\n\nYou may be asking yourself, &quot;How does that happen?&quot;\n\nThey may have simply forgot. It happens all the time, although it's usually because a sheriff's department employee who should have registered the convicted sex offender doesn't think the guy will do it again.\n\nSometimes a new chief or sheriff takes over the department, doesn't like the policy, and cancels the sex offender registry.\n\nHere's another scenario: When the convicted sex offender is released from prison, he is given a deadline by the parole board to register.\n\nIf the parole board gives the guy a year, the sheriff's department is probably going to give him a year.\n\nHowever, if the parole board gives the guy a year and the county sheriff wants him off the street, he will register the convicted sex offender that day.\n\nIn the case of Felix-Jimenez, the sex offender registry is mandatory and the deadline wasn't up yet, so how he is able to be on the loose is still unclear.\n\nThe good news is that Felix-Jimenez hasn't gone into hiding.\n\nIn fact, the most recent report shows that he is staying in a homeless shelter in Van Nuys.\n\nCalifornia doesn't want him there.]" time="0.017"><properties><property name="score" value="0.022629135460372485" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02262914&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02262914
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Granted, my current policy does not meet what I want to do, so I have done some thinking on what changes would be nice. There are two major changes I think should be made.\n\nThe first would be to include the deletion policy as the first part of the \u201cFAQ\u201d section, as opposed to the second, where it is now. The issue here is that it is currently hidden away, and unless you know to go there, you might not even know it exists. This is why I believe the policy should be in the FAQ, in addition to the other policy changes.\n\nThe second major change would be to change the time periods for deleting and undeletion requests. Currently, these time periods are a month for both and can be extended for a total of 2 months. These time periods need to be shortened. If a user creates an account, and decides it is not for them within a week, they should be able to delete it without the issue being made public, since no other users have commented on it. If a user creates an account and decides that it is not for them within the first month, they should also be able to delete it without it being made public. Also, it is not fair to users who are newer, and decide that they do not wish to use the site for the first time period to only have it deleted. These users should be able to delete them after this time period if they choose. I believe these time periods should be a month for deletion requests, and 2 weeks for undeletion requests.]" time="0.012"><properties><property name="score" value="0.06822813112859483" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Perhaps best known for their exploits during the First World War and the Gallipoli campaign in particular, the ANZACs, or Anzacs, are the legendary members of the Australian and New Zealand Army Corps.\n\nTheir contribution to the military campaign in the Middle East, including the battle at Gallipoli in 1915, will be recognised this year, with the 100th anniversary of the Gallipoli campaign being commemorated by the nations involved.\n\nThe events of the Gallipoli campaign resulted in heavy casualties, with a staggering 258,000 deaths and an additional 156,000 injuries for the Allied forces, including 44,000 British deaths and 29,000 French deaths.\n\nThe staggering casualty figures have subsequently been seized upon by those who question the strategic worth of the campaign.\n\nOne figure, British historian Niall Ferguson, wrote that \u201cthe Dardanelles fiasco was one of the greatest military blunders of all time, a story of incompetence, vanity and wishful thinking out of which could have been constructed a textbook case of how not to run a war\u201d.\n\nSo why were the Gallipoli campaign and the actions of the ANZACs so significant for Australia and New Zealand, and what did the battle mean for the two countries?\n\nArrival of the ANZACs\n\nFollowing the outbreak of war in 1914, the Australian and New Zealand forces began arriving in Egypt. On February 25 1915 the first ANZAC troops arrived at Gallipoli, a total of 11,000. The rest of the 20,000 strong force arrived over the following weeks.\n\nMany of the ANZAC troops had volunteered at the outbreak of war, lured by the romantic and patriotic view of warfare and by a yearning for adventure.\n\nJust as the troops were largely made up of volunteers, the casualties at Gallipoli were also predominantly made up of volunteers.\n\nIt has been estimated that 10 percent of all the Australian and New Zealand troops at Gallipoli were killed during the campaign.\n\nAustralian Prime Minister Sir Robert Menzies commented that \u201cthe experience of war for Australians would always be Gallipoli\u201d.\n\nPrior to the arrival of the ANZAC troops, the British had launched an attack on the Gallipoli peninsula on the 25th of April 1915.\n\nThe Allied attack consisted of five divisions, with most of the troops coming from the British Army. It was hoped that the landings would help the British Navy to get through the Dardanelles, a narrow strip of water separating the Aegean Sea and the Sea of Marmara, which leads to the Bosphorus and the Black Sea.\n\nHowever, the initial landings met with great opposition from the Turkish defenders, and the Allied forces soon became pinned down on the beaches and unable to advance.\n\nThe ANZAC troops arrived at Gallipoli, the capital of the Ottoman Empire\u2019s European territories, just over a month later on the 19th of May 1915.\n\nThe ANZACs had initially been part of the British and French force in Egypt. They were moved to a separate command under General Ian Hamilton, with Australian General William Birdwood as his deputy.\n\nA General Order issued on the 3rd of May stated that the Australian and New Zealand troops were to be given \u201cseparate and distinct\u201d duties from those of the British.\n\nThis was not intended to be interpreted as any form of segregation. Rather, it was due to the fact that the Australians and New Zealanders, despite being part of the British Empire, were soldiers who had volunteered from their own country, and who had formed a unique sense of national identity and pride.\n\nIn order to maintain a clear distinction between the Allied forces and the enemy, the ANZACs also adopted their own distinctive insignia, such as a yellow or red poppy on their badges and hats.\n\nHowever, despite the greater efforts to set the ANZACs apart from their Allied allies, the treatment of the soldiers at Gallipoli was the same as that received by the British.\n\nThe ANZACs were given responsibility for the defence of the Anzac Cove, at the northernmost tip of the Gallipoli Peninsula.\n\nThe cove was strategically important to the Turks because of its proximity to the Narrows, the strip of water which separated the peninsula from the Dardanelles.\n\nThe Anzacs were also expected to hold on to the various hills and ridges along the peninsula, despite the fact that the Allied positions were in constant danger of being shelled]" time="0.067"><properties><property name="score" value="0.04029113631882201" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04029114&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04029114
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[V6 engines had too much power to safely make use of the front wheels alone. V8 engines didn\u2019t fit between the front and rear axles, making them virtually useless for most people.\n\nHowever, a minivan with four-wheel drive would have been a dream come true.\n\nSource]" time="0.013"><properties><property name="score" value="0.02574850452566335" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0257485&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0257485
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[I love hand-cranked coffee grinders, but a fully automated device that grinds directly into the portafilter has always been one of the holy grails of coffee brewing. The Ekobrew takes a lot of the guesswork out of the whole process. Fill the hopper with whole coffee beans, twist the top to open, drop the portafilter in, turn the handle a few times and you\u2019re ready to make your espresso.\n\nThe container is made of a non-porous material, so you can put ground coffee in it. It also has a couple of holes on the side, so that you can clean it and store it without having to remove the portafilter. The concept is pretty ingenious, but I had some difficulty in actually getting it to work properly.\n\nIn theory, the machine takes a lot of the stress out of making espresso. When it works, the espresso is good. It has a big enough capacity that I could probably make it twice before having to refill it. It would be nice if there was a better way to tell how full it was.\n\nUnfortunately, it\u2019s a fairly unreliable method of extracting espresso. You can only have it set to brew for a specific amount of time. You can\u2019t use any sort of fancy steaming wand to make sure the pressure is exactly right, or adjust the grind to get a specific flavour. I think this device is designed for making a quick shot in the morning and then tossing the puck after the shot is made.\n\nThis is the kind of device that I can imagine being useful to take on a trip, or maybe for an RV or camper, but not in a regular espresso bar or cafe. At the price of $50, I can\u2019t really recommend it, though I do think that it has potential. If it were a bit more inexpensive, I think it would be worth experimenting with.]" time="0.026"><properties><property name="score" value="0.08028490344212061" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0802849&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0802849
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[I like it!\n\nI like it!\n\n$750 DSQUARED2 Short dress is on sale! $375\n\nI like it!\n\nI like it!\n\n$170 CRISTINAEFFE Short dress is on sale! $72\n\nI like it!\n\n$290 CHRISTIES \xc0 PORTER Short dress is on sale! $206\n\nI like it!\n\nI like it!\n\nI like it!\n\nI like it!\n\nI like it!\n\n$340 ERMANNO DI ERMANNO SCERVINO Short dress is on sale! $254\n\nI like it!\n\n$440 Jil Sander Short dress is on sale! $323\n\nI like it!\n\n$138 TROU AUX BICHES Short dress is on sale! $85\n\nI like it!\n\nI like it!\n\nI like it!\n\n$460 Alberta Ferretti Short dress is on sale! $322\n\nI like it!\n\nI like it!\n\nI like it!\n\n$219 CRISTINAEFFE Short dress is on sale! $117\n\nI like it!\n\nI like it!\n\nI like it!\n\n$104 JUST FOR YOU Short dress is on sale! $70\n\nI like it!]" time="0.015"><properties><property name="score" value="0.041663972047661924" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Horses, rabbits and fish swimming around in glass globes are an enticing attraction at Hanging Lake Lodge in Fruita. But, for more than a century, the people have had a passion for Hanging Lake.\n\nChildren, horses, deer and pet dogs are all welcome to stay at Hanging Lake.\n\nDogs are welcome at Hanging Lake Lodge, as long as they don't swim.\n\n&quot;There were (people) that were always trying to get to the lake,&quot; said Dale Hanks, whose great grandfather built the original Hanging Lake Hotel.\n\nHanks has been running Hanging Lake Lodge since 1972, and says his customers are almost always from out of town.\n\n&quot;They've been told about it by their neighbor, they've been told about it by their son who visited last year, they've been told about it by friends in Colorado, and when they're here they just have to come,&quot; Hanks said.\n\nMany visitors are drawn to Hanging Lake by the nearby Red Canyon, as well as the Colorado National Monument. The lake is located in a canyon along the Colorado River, nestled at the base of a rock formation called the Palisades. The National Monument, also known as The Dots, is a rock formation of mostly red and orange that spans for about a mile along the Colorado River.\n\nThe Dots.\n\nHanging Lake is a water body that was formed by lava over millions of years, according to the National Park Service. The water body is fed by a small underground river and water seeps out through the canyon walls to form the lake.\n\nHanging Lake is known as one of the most picturesque lakes in the country, and often the photos seen online or on Instagram have a filter on them. The color of the water is bright blue, the rocks are a fiery red and orange, and the waterfalls are turquoise blue.\n\nA tourist sits by the Hanging Lake waterfall.\n\n&quot;It's an excellent opportunity to do some waterfall photography,&quot; said Ben Lin, a Denver resident. &quot;That is what drew me here, the waterfall, and I don't know why, but it is very hard to take a bad picture here.&quot;\n\nLin has been to Hanging Lake several times, the first time in 2015. He came with a small group of about 10 people, and the people in his group were in awe at the sight.\n\nThe fall\n\n&quot;This is the best,&quot; Lin remembers one person saying. &quot;This is the best picture, we are so lucky.&quot;\n\nThe water temperature in the summer is about 45 degrees, and not many people are willing to dive into the water, which Lin says is a shame.\n\n&quot;I feel like you should be able to swim here, it's a beautiful place to swim,&quot; Lin said. &quot;The problem is the lake is only as deep as your chin. When you're done swimming there is all that debris and gunk on the lake, you can't see it from up above, but as you're swimming you feel the gunk and rocks in the bottom of the lake.&quot;\n\nAlthough swimming is prohibited, people do still swim in the lake, which causes the algae to increase, according to the National Park Service.\n\nHanging Lake is one of the first stops for many travelers, but after leaving the picturesque sight, many people don't realize there is so much more to Fruita than Hanging Lake.\n\nNext time you are passing through the small town of Fruita, consider stopping at Hanging Lake Lodge for lunch, a short hike, or an afternoon in a hammock. The history behind the town and the visitors will not disappoint.\n\nEditor's note: This article has been updated to correct information about Hanging Lake. The photos seen on Instagram are not enhanced in any way.]" time="0.066"><properties><property name="score" value="0.041907091697777336" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04190709&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04190709
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Fantasy baseball: Six more players who may not be healthy\n\nSportsPulse: USA TODAY Sports' fantasy baseball experts tell you which players you should buy low and sell high as the season gets into full swing.\n\nJoin the conversation or respond by signing in\n\nFantasy Home\n\nPopular Tags\n\nPlease enable JavaScript to view the comments powered by Disqus.]" time="0.012"><properties><property name="score" value="0.038412379482827275" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03841238&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03841238
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[You won't see me taking the same chances\n\nYou can't expect me to start taking your drugs again\n\nAnd now I'm out of the darkness I'll never be lost again\n\nI'll never be lost again\n\n\n\nYou never knew my secret name\n\nYou never heard of my all seeing eyes\n\nYou never knew my secret name\n\nYou never heard of my all seeing eyes\n\n\n\nYou know you're walking on the ceiling and now you're kicking my ass\n\nYou're not the only one who has had the best of me\n\nNow that I'm out of the darkness I'll never be lost again\n\nI'll never be lost again]" time="0.013"><properties><property name="score" value="0.036441084355065545" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Corporate Counsel is partnering with several organizations to host an event on October 20, 2017, on the future of diversity in corporate law, featuring Keisha Lance Bottoms, Mayor of Atlanta; Deidra Lee, Atlanta District Attorney; Aisha Dew, Founder &amp; CEO of We The Change; and Connie Maysonave, CEO of Atlantans for a Fair Future.\n\nThe event will feature a moderated conversation among Mayor Bottoms, Ms. Dew, Ms. Lee, and Ms. Maysonave, followed by a networking session and a reception. The event will take place at the Loews Atlanta Hotel from 5:30 to 7:30 p.m.\n\nTo register for the event, click here.]" time="0.012"><properties><property name="score" value="0.05535225817856482" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05535226&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05535226
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[He has won numerous awards and has been honored by both the people and the government of Jamaica. These accolades are well-deserved and confirm the love and respect the people of Jamaica have for this music genius. In December of 2005, Williams won the prestigious Golden Arrow award for his music and contribution to the music industry in Jamaica. In February of 2006, he was awarded the honorary Order of Jamaica by Prime Minister Bruce Golding at a gala event held at Jamaica House.\n\nWalter &quot;Sweetie&quot; Williams has been a major force in the growth and popularity of Jamaican music, in particular]" time="0.012"><properties><property name="score" value="0.026868987820738433" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02686899&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02686899
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[During my first month of graduate school, I received a call from my aunt on a weekend when I was in the middle of working on my thesis. She needed help moving out of a rental property she owned. I went over that weekend and helped her clear out the house.\n\nAfter it was all done, she said to me, \u201cYou know, a lot of people would be really upset at me for making them do this.\u201d\n\nI smiled.\n\n\u201cThat\u2019s why I am here.\u201d\n\nIt\u2019s funny how a simple answer can change the dynamics of a relationship.\n\nI continued to help my aunt after that weekend. She would ask for my help, and I would drop what I was doing. She asked, I gave. I took it for granted that it was my responsibility to help her out, to be available. I also took it for granted that she was comfortable asking me for help.\n\nA few years ago, I went to see my aunt in the hospital. She was having a difficult time recovering from her medical procedure, and I was her first call. I didn\u2019t think twice about dropping everything and going over to her place to take care of her and her dog, especially since she had just moved to an apartment that didn\u2019t allow dogs.\n\nBut I got to the hospital and found out that my uncle was also at the hospital. I didn\u2019t realize that he was in the hospital, and, by the time I found out, he was asleep in a nearby room. I talked to my aunt and we came to the realization that it was important for her to have my uncle there. She needed him in a way that she didn\u2019t need me. I left.\n\nI don\u2019t think my aunt and uncle even realized how much I had changed. I was not as available to help as I used to be, and I knew my time was better spent on other things. I was a graduate student, after all, and I had a whole other life to live.\n\nAs a result of all of this, my aunt and uncle started to depend on others to help out more. They hired a house cleaner and a housekeeper. They began to call people in order to receive the help they needed. I had thought that they would not want to go through the hassle of finding someone to help, but they did.\n\nIt\u2019s as if they needed to be reminded that it was okay to ask for help. They needed to realize that asking for help wasn\u2019t a sign of weakness, but a sign of strength. That asking for help meant you were confident in who you were and what you had to offer, and that you weren\u2019t afraid to ask for help if you needed it.\n\nIn doing so, my aunt and uncle showed me what it meant to ask for help, and I realized that asking for help could be a sign of strength, not weakness.\n\nAdvertisements]" time="0.024"><properties><property name="score" value="0.018587339845882286" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01858734&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01858734
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Une analyse de la CNIL\n\nBac d\u2019information, traitement automatis\xe9, robots, IA\u2026 nous avons tendance \xe0 penser que les syst\xe8mes automatis\xe9s ne sont pas faits pour r\xe9pondre \xe0 nos besoins de compr\xe9hension, de sensibilit\xe9 et d\u2019empathie. Or, \xe0 l\u2019occasion de l\u2019assembl\xe9e g\xe9n\xe9rale de l\u2019institut Data &amp; Society qui a eu lieu la semaine derni\xe8re, nous avons eu l\u2019opportunit\xe9 d\u2019entendre les explications d\u2019Omer Tene, chercheur en sciences cognitives, et professeur associ\xe9 en intelligence artificielle. Il nous a d\xe9montr\xe9 que les machines n\u2019\xe9taient pas forc\xe9ment con\xe7ues pour remplacer l\u2019homme. Bien au contraire. L\u2019apprentissage automatique est aujourd\u2019hui une v\xe9ritable valeur ajout\xe9e et permet de d\xe9passer les limites des technologies humaines.]" time="0.013"><properties><property name="score" value="0.11886562208343054" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.11886562&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.11886562
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[C\xe9sar D\xedaz \u2022 29/10/2018 - 10:08\n\nSebasti\xe1n, yo te recomiendo que te hagas el seguimiento de los indicadores y que no hagas mucho caso de los medios tradicionales. El \xfaltimo sondeo se\xf1ala que las elecciones se definir\xe1n por un muy muy muy estrecho margen de diferencia. La probabilidad de que el Frente Amplio gane el 23 es muy alta y cuanto m\xe1s alta es, mayor es la probabilidad de que gane. En estos momentos hay dos variables que van a determinar que el FA gane:\n\n- Cambios en la opini\xf3n p\xfablica (es decir, la propaganda, la capacidad de atraer m\xe1s votos, etc).\n\n- Efecto S\xe1nchez (no s\xe9 qu\xe9 les pasa, se equivocaron en las \xfaltimas elecciones de Espa\xf1a y ahora si no es el chileno el que entra a saco contra Vidal, el brasile\xf1o tambi\xe9n).]" time="0.014"><properties><property name="score" value="0.13728622044044125" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.13728622&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.13728622
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Why do we need 2 statements? Isn't the first one enough to start a transaction?\n\nThe first statement is an Open. It does not actually do anything yet. It sets up the transaction, but does not actually start it. In order to do something, you must start the transaction.\n\nIt can be done, and in fact, I think that is how it is done in your example code. However, I prefer to use a begin statement for the transaction and then perform the Open within the transaction.\n\nI do this for 2 reasons:]" time="0.013"><properties><property name="score" value="0.007611471965651706" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00761147&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00761147
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Overview\n\nGreta Esta\xf1ol's Carmina \xc1urea are sacred poems meant to inspire devotion in the faithful. They are simple in form, yet profound in content. The author has combined medieval traditions with a contemporary perspective on the faith, creating a simple, sincere, and authentic rendition of the Catholic religion. It was in this vein that she wrote these lyrical and enlightening poems. The collection consists of approximately two hundred poems, many of which have been translated from Esta\xf1ol's Spanish original by Michael Coffey. Carmina \xc1urea are an introduction to Christian spirituality. This collection contains the following poem: To Learn the Way\n\n\n\nThe Value of Purity\n\nLove Conquers the World\n\nHoly Bible\n\nThe Sacraments\n\nLiturgical Calender\n\nMy Jesus\n\nA Prayer for Peace\n\nGod Loves Me\n\nThe Feast of the Immaculate Conception\n\nThe Sacrament of Baptism\n\nFaith and Confession\n\nThe Sacrament of Holy Communion\n\nOur Father, Who Art in Heaven\n\nThe Story of Our Lord's Birth\n\nThe Easter Vigil\n\nThe Holy Mass\n\nThe Sacrament of Marriage\n\nThe Sacrifice of the Mass\n\nThe Sacrament of Confirmation\n\nHoly Mary, Our Mother\n\nSacred Heart of Jesus\n\nOur Lady of Sorrows\n\nThe Our Father and Hail Mary\n\nThe Great Love\n\nBlessed Virgin, My Mother\n\nThe Holy Rosary\n\nThe Sacrament of Penance\n\nThe Gifts of the Holy Spirit\n\nHoly Spirit, Breathe on Me\n\nThe Day of Judgment\n\nThe Second Coming of Our Lord\n\nThe Epiphany\n\nOur Lady of Guadalupe\n\nThe Holy Spirit\n\nHoly Trinity, Trinity\n\nO Lord, I Will Not Keep Silent\n\nThe Most Holy Eucharist\n\nThe Trinity\n\nO Most Holy Trinity\n\nJesus, Mary, Joseph\n\nAn Angel of God\n\nOur Father, Who Art in Heaven\n\nOur Lady of Guadalupe\n\nHoly Father, Have Mercy on Me\n\nThe Sacrament of Extreme Unction\n\nThe Incarnation of Our Lord\n\nIn the Beginning\n\nAmen\n\nThe Power of Prayer\n\nTo a Loved One\n\nThe Love of God\n\nThe Holy Spirit\n\nWho Is the Holy Spirit\n\nHoly Spirit, the Breath of Life\n\nGlory Be to the Father\n\nThe Holy Family\n\nOur Lady of Lourdes\n\nBlessed Mother, Behold Your Son\n\nHail, Holy Queen\n\nTo Mary, My Mother\n\nPraise the Lord\n\nHail, Holy Queen\n\nTo the Crucifix\n\nThe Angelic Salutation\n\nA Prayer for Parents\n\nTo the Sacred Heart of Jesus\n\nA Prayer of Thanksgiving\n\nCome, Holy Spirit\n\nJesus Christ, My Love\n\nThe Holy Spirit\n\nHoly Spirit, Dwell in Me\n\nO Holy Spirit, Thou Who Art Thrice Holy\n\nO Holy Spirit, Unction of God\n\nO Spirit of Love, Uniting Soul and Body\n\nCome, Holy Spirit, Come\n\nHoly Spirit, Light of Love\n\nHoly Spirit, Send Your Spirit of Love\n\nHoly Spirit, The Spirit of Love\n\nThe Holy Spirit\n\nThe Trinity\n\nAll Creation Loves You\n\nThe Baptism of Jesus\n\nAn Angel of God\n\nHe Is a Man Who Sees\n\nBehold the Word Made Flesh\n\nSing a Song of Incarnation\n\nBehold the Door of the Kingdom of Heaven\n\nMay the Glorious Mysteries of Jesus\n\nCome to Us, O Jesus\n]" time="0.023"><properties><property name="score" value="0.07983633832672213" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07983634&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07983634
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The BB&amp;T Building (also known as the 100 North Tryon Building) is a 17-story skyscraper in Charlotte, North Carolina, United States. It is the tallest building in Charlotte outside of Uptown. It was originally known as the Bank of North Carolina Building, when completed in 1921. In 1973 it became known as the NationsBank Building. In 1998 it became the Wachovia Building and in 2004 it became the BB&amp;T Building.\n\nArchitecture\n\nThe BB&amp;T Building is designed in the Art Deco style, which became popular during the 1920s and 1930s. The building features limestone and terra cotta facing, and metal and glass curtain walls. The top of the building features two pyramid-shaped ornamental caps.\n\nTenants\n\nThe building was built in 1921 for the Bank of North Carolina. The bank grew to be the largest bank in the South, before merging with First Union in 1998, becoming Wachovia. Wachovia's Charlotte headquarters occupied three floors in the building until the company merged with Wells Fargo in 2009.]" time="0.014"><properties><property name="score" value="0.03031400632839132" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03031401&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03031401
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Investment fund transfers are back to normal now that all new subscriptions have been processed for the month of July.\n\nOnce you've logged into your account, you'll notice that your portfolio has been adjusted to reflect the purchases and redemptions of the month.\n\nWe hope that you found this update helpful. If you have any further questions, please don't hesitate to reach out to our support team.]" time="0.012"><properties><property name="score" value="0.028087939412890783" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02808794&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02808794
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The Canonical Path for Classes\n\nLast Updated on Wed, 04 Jan 2016 | Excel 2003 VBA\n\nWhen you create a class module, you must specify the following items in the declaration:\n\n\u25a0 The class name\n\n\u25a0 The objects that are instances of the class\n\n\u25a0 The properties that the class will support\n\n\u25a0 The methods that the class will support\n\n\u25a0 The events that the class will support\n\nAll classes in Excel derive from a generic base class called Object, which is an instance of the class of the same name. The Object class is an abstract class and doesn't have a constructor. If you don't have any idea what I'm talking about here, don't worry about it for now. I'll explain it in detail in Chapter 5. You can't instantiate an object from the Object class, so the only real purpose of the Object class is to serve as a base class for other classes. You can think of the Object class as being like a template that all other classes derive from. If you want to learn more about how classes work and how you can customize Excel's classes to create your own customized classes, check out my Microsoft Press book Microsoft Excel VBA and Macros (ISBN 0-7356-1845-0).\n\nWhen you create a class module in your Excel application, you must specify the class name as a string literal. For example, if you create a class module called Class1, Excel uses the following syntax in the declarations area of the class module to specify the name of the class:\n\nPublic Class Class1 End Class\n\nAs with any other module in Excel, a class module doesn't do anything until you create an object from the class module. Creating an object is a two-step process: You first instantiate the object, and then you call the methods and properties of the object. For example, to create an object from the class named Class1, you use the following code:\n\nDim obj1 As New Class1\n\nOnce you create an instance of the object, you can call the methods and properties of the object by preceding the method or property name with the name of the object. For example, if you created an instance of the class named Class1, you would use the following code to call the method named Display:\n\nobj1.Display\n\nIf the object contains an object variable that represents another object of the same class, you can also call the methods of that object. For example, if the object named obj1 created in the preceding code contained an object variable named obj2 that represented another object of the same class, you would use the following code to call the Display method of the object named obj2:\n\nobj1.obj2.Display\n\nThe property and method names of a class are preceded by the name of the class, as shown in the preceding examples.\n\nIf you create a class module that contains a custom collection object, such as a collection of cells in a worksheet, you must use a new operator to create the collection object. For example, the following statement creates a collection object named myCol1 that is a collection of worksheet cells:\n\nDim myCol1 As New Collection\n\nWhen you create a collection object, you specify the type of object that the collection can contain when you define the type of the collection.\n\nFor example, the following statement creates a collection object that can contain only the cells in a worksheet:\n\nDim myCol1 As New Collection(Worksheet.Range)\n\nWhen you create a collection object, you can add items to the collection object by using the Add method of the collection object.\n\nFor example, the following statement adds a range named &quot;Region&quot; to the collection object named myCol1:\n\nmyCol1.Add mySheet.Range(&quot;Region&quot;)\n\nThe Add method takes two arguments:\n\n\u25a0 The object that you want to add to the collection object\n\n\u25a0 The collection that you want to add the object to\n\nWhen you create a collection object, you can add an object to a collection object that is an instance of a different class, but the type of the objects in the collection must be the same.\n\nThe following code example adds the range named &quot;Region&quot; to the collection object named myCol1. This code example assumes that you've created the objects in the declaration section of the class module.\n\nSub AddObjectToCollection() Dim myCol1 As New Collection(Worksheet.Range) Dim myRow As Long, myColumn As Long Dim strName As String ' Obtain a reference to the collection ' that contains the first worksheet in ' the active workbook. Set myCol1 = Sheets(1).Range(1, 1). _ SpecialCells(xlCellTypeVisible) ' Initialize the name of the cell ' to the value of the Name property. strName = myCol1.Cell(1, 1).Value ' Determine the address of the cell ' and store it in the myRow and myColumn ' variables. myRow = myCol1.Cell(1, 1).Address myColumn = myCol1.Cell(1, 2).Address ' Create a new Range object. Set myRange = Worksheets(&quot;Sheet1&quot;). _ Range(&quot;Region&quot;) ' Add the new range to the collection. myCol1.Add myRange End Sub\n\nThe following code example calls the AddObjectToCollection method to add the range named &quot;Region&quot; to the collection object named myCol1. This code example assumes that you've created the objects in the declaration section of the class module.\n\nSub AddObjectToCollectionWithMethodCall() Dim myCol1 As New Collection(Worksheet.Range) Dim myRow As Long, myColumn As Long Dim strName As String ' Obtain a reference to the collection ' that contains the first worksheet in ' the active workbook. Set myCol1 = Sheets(1).Range(1, 1). _ SpecialCells(xlCellTypeVisible) ' Initialize the name of the cell ' to the value of the Name property. strName = myCol1.Cell(1, 1).Value ' Determine the address of the cell ' and store it in the myRow and myColumn ' variables. myRow = myCol1.Cell(1, 1).Address myColumn = myCol1.Cell(1, 2).Address ' Create a new Range object. Set myRange = Worksheets(&quot;Sheet1&quot;). _ Range(&quot;Region&quot;) ' Add the new range to the collection. myCol1.AddObject myRange ' Add another range to the collection ' and display the number of elements in the collection. myCol1.Add myRange MsgBox myCol1.Count End Sub\n\nThe final statement in the AddObjectToCollectionWithMethodCall subroutine is a good example of how you can use a method that belongs to a class to call the methods of the collection object.\n\n0 0]" time="0.080"><properties><property name="score" value="0.09882464371104277" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Title\n\nInsurance Leads\n\nDescription\n\nWe are a leading Insurance Agency in Central Ohio offering all types of insurance. Our insurance agency offers insurance services for all types of customers from those who have been with the same insurance agency for many years to new customers that have not been with an insurance agency before. Our goal is to provide outstanding customer service and the best insurance services to our customers.\n\n\n\nPlease visit our website at http://www.affordableinsuranceservices.net/\n\nKeywords\n\nBusiness Name: Affordable Insurance Services\n\nSearch Phrase: Affordable Insurance Services\n\nLocation\n\n1 Water Street, Delaware, OH 43015, USA\n\nCategory\n\nBusiness Name\n\nAddress\n\n1 Water Street, Delaware, OH 43015, USA\n\nContact Information\n\nPhone:\n\nUrl:]" time="0.012"><properties><property name="score" value="0.02374648841574345" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[India has seen a significant spike in energy demand in recent years, driven by economic growth, rapid urbanisation and the increasing electrification of the country. Energy production has increased over the same period, but the country\u2019s domestic energy supply has not kept pace. The growth of energy demand outpacing domestic production resulted in a significant reduction in oil and natural gas stocks. The country became increasingly dependent on imported energy.\n\n\n\nDomestic energy production is also likely to experience a plateau]" time="0.012"><properties><property name="score" value="0.004444144959386831" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Many people who oppose the minimum wage assume that raising the minimum wage is bad for the economy and will cause a lot of unemployment. This is mostly false. A new study has revealed that raising the minimum wage does not kill jobs. This is not just good news for the people who will now get a raise, but also for the economy as a whole.\n\nOpponents of the minimum wage often say that it will create a lot of unemployment and that this is proof that the minimum wage should be abolished. They argue that it creates a job loss on an economy-wide scale. However, this new study has proved that the minimum wage will not cause massive job loss. This is not just great news for the people who get a raise, but it also means that the economy will benefit.\n\nThe study comes from the National Employment Law Project (NELP) and reveals that raising the minimum wage does not cause massive job loss. In fact, only two percent of the workforce is impacted by a minimum wage increase. This means that the economy will benefit, but not that much. However, it is still a positive impact.\n\nHowever, the study also shows that the country would have to raise the minimum wage by $2.13 to achieve the same number of jobs as a similar increase in the 1990s. This is not really a surprise, as the economy was growing more rapidly back then.\n\nRaising the minimum wage was also seen to boost the economy. The study found that a $1 increase in the minimum wage would increase the GDP by $2.62. This is a significant amount and would be great for the economy as a whole. In fact, it could even boost the economy so much that it could lift more people out of poverty.\n\nWhile a raise in the minimum wage does not have a massive impact on the economy as a whole, it still has a positive impact. This means that the people who will now get a raise will enjoy it. This will also help the economy, and make the country a better place to live in.\n\nThose who oppose the minimum wage should know that it does not create massive unemployment. This is also great news for the economy, and it means that raising the minimum wage is great for everyone.]" time="0.023"><properties><property name="score" value="0.053425327882620555" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[\u201c\n\nI have no hesitation in recommending you to the body of workers in Australia, either in Railway or Country Tramways, or Electrical construction. Your outfit has been constantly at work for nearly three years and has been found fully to justify my own and your own high recommendation of it, both for its efficiency and economy of working. The Cableways I have had in operation for two years are found to work in the most satisfactory manner, both in operation and maintenance, in fact I do not know of any system that I would prefer for tramways, and I am glad that I have them.\n\nJ.R. Lucas, Chief Engineer,\n\nQueensland Railways\n\nNashua, New Hampshire, August 20, 1898]" time="0.012"><properties><property name="score" value="0.06303423846689138" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06303424&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06303424
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[READ MORE Wall Street: Online poker rooms are fading, but who is 6/10/2018 \xb7 As online poker room software and banking options face an uncertain future in the U.S., many of the sites that are still online are trying to\n\nREAD MORE Exclusive: Where do U.S. poker players play online poker 7/10/2018 \xb7 With the April 15th deadline for Nevada-based online poker rooms to stop offering real-money games, where will the displaced players go next?\n\nREAD MORE Online poker room ratings, reviews, and deals - Poker In the wake of Black Friday, you can\u2019t walk three feet in a Las Vegas casino without bumping into someone who\u2019s looking for a blackjack\n\nREAD MORE Nevada Online Poker Rooms On Verge Of Exiting Market Online poker rooms in the United States are in a precarious position right now. They still have a month left to get their affairs in order,\n\nREAD MORE What Happens To U.S. Poker Players Now? - Card Player 4/3/2019 \xb7 For a site to be eligible for inclusion on the Upcoming US-facing Online Poker Sites list, it must be confirmed to launch in the United States. This means\n\nREAD MORE Who are the best US poker sites in 2019? | Best US poker 3/11/2019 \xb7 The only active and legal online poker room in the U.S. is currently the 888 Poker. The American online poker has not gone anywhere since Black Friday,\n\nREAD MORE Gambling in the United States - Wikipedia Online Poker Reviews - Top Rated Online Poker Sites for 2018. Online poker rooms in the United States are not as popular as in other countries due to\n\nREAD MORE Online Poker Sites in the United States, All Active and Are Online Poker Rooms Still Profitable? Find out more about the best online poker rooms in the United States today.\n\nREAD MORE Poker News &amp; Articles: Read Online &amp; Watch the Live Videos A ranking of the best online poker rooms in the United States, including casino sites for U.S. players and trusted online poker rooms with\n\nREAD MORE Casino Game Reviews - Top Rated Online Poker Sites for 5/22/2016 \xb7 Though not as big as the gambling industry in other parts of the world, the United States does offer a number of legal gambling options to its\n\nREAD MORE Online Poker News - Latest News from Online Poker Rooms 2018 Online Poker Sites 2018 is a collection of the best online poker rooms available today for US players. You can find all of the important info\n\nREAD MORE Where are the best US online poker rooms right now? 7/17/2017 \xb7 The legal US online poker landscape could be getting a little bit clearer, as Nevada has once again made a bill that would create an online poker licensing\n\nREAD MORE Online Poker News &amp; Articles - Poker News | Upswing Poker Blog 4/29/2015 \xb7 Online poker in the U.S. has long been an industry in limbo, with the legality of the business depending on how you define it. Now the U.S.\n\nREAD MORE Las Vegas Poker Rooms, Poker Tournaments and Game The list of best US poker sites in 2019 includes the latest online poker rooms that cater to United States-based poker players. Read about the top U.S. poker\n\nREAD MORE Online Poker in the U.S. \u2013 The Ultimate Guide | Upswing Poker Online poker room reviews featuring top sites for US players. Exclusive poker bonuses, promotions and expert reviews.\n\nREAD MORE Online Poker Guide for U.S. Poker Players - US Poker Guide 4/29/2015 \xb7 Online poker is now more popular in the U.S. than in Europe. In the U.S., state laws and gambling regulations differ by state, and some states have\n\nREAD MORE Vegas Palms Online Casino, Online Poker &amp; Online Slots Join PokerStars and enjoy top quality online poker. New players get a 100% match bonus up to $600. Play on Windows PC and Mac with our instant download client.\n\nREAD MORE Gambling in the United States - Wikipedia The U.S. online gambling market is the largest in the world. By some estimates, the U.S. accounts for one-third of all online gaming revenue.\n\nREAD MORE Online Poker in the U.S. \u2013 The Ultimate Guide | Upswing Poker Online Poker Guide for U.S. Poker Players. If you are a poker player looking to play online poker in the United States, here you will find the latest\n\nREAD MORE Online Poker and Gambling Laws for 2018 - Casinolistings.com 4/15/2017 \xb7 In the wake of Black Friday, the number of online poker rooms in Nevada dwindled from six to just one, and the number of players dropped\n\nREAD MORE Poker Players on a Dime: What US Online Poker Rooms Are 2014 Online Poker Sites - US Online Poker Guide. List of Top USA Poker Sites accepting U.S. Players including all the latest bonuses and promotions.\n\nREAD MORE Online Poker USA \u2013 Online Poker Sites for US Players A brief introduction to online poker and online gambling in the United States.\n\nREAD MORE USA Online Poker Guide - UsaPokerOnline.com 5/12/2019 \xb7 &quot;We are going to be online, we are going to be regulated,&quot; said Robert Capecchi, director of federal legislation at the Marijuana Policy Project, who led a\n\nREAD MORE Online Poker and Gambling Laws for 2018 | GamingLawOnline.com Gambling in the United States has been legal in each of the 50 U.S. states since 1911, except for a brief prohibition after World War I. At the\n\nREAD MORE 5 Must-Visit Online Poker Rooms for U.S. Players 4/15/2017 \xb7 Exclusive: Where do U.S. poker players play online poker now? We asked, and you answered. We asked, and you answered.\n\nREAD MORE Online Poker in the US in 2017 - Online Poker Guides, Tips U.S. online poker has a dark history, littered with fraudulent players and hacked accounts. This guide will help you find a safe place to play.\n\nREAD MORE Online Poker Rooms - ThePokerBank.com Nevada-based poker rooms have until April 15th to exit the US market or risk being shut down. Players are looking for a new online poker home.\n\nREAD MORE Online Poker Room Rankings - ThePokerBank.com Learn more about top online poker sites and poker rooms in the United States.\n\nREAD MORE USA Online Poker Guide | Play Poker Online in the US U.S.A. Online Poker Sites. A list of all the best USA online poker sites, casino sites, and sportsbook sites with information on bonuses, games, and more.\n\nREAD MORE USA Online Poker Guide: Reviews, News, &amp; Tips - Worldwide 7/16/2015 \xb7 The status of online poker in the United States is likely to remain a confusing mess for a]" time="0.072"><properties><property name="score" value="0.01133885507222414" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Borodino (village), Moscow Oblast\n\nBorodino () is a village (&quot;selo&quot;) in Odintsovsky District of Moscow Oblast, Russia, located on the Tvertsa River.\n\nThe village of Borodino was first mentioned in 1611 as part of the Tsaritsyn &quot;oprichnina&quot;.\n\nIn 1812, the Battle of Borodino, a major battle of the French invasion of Russia took place near Borodino. The village is now a site of the museum and the memorial. The French lost the battle, but went on to win the war in 1814.\n\nIn 1898, Borodino was the place where Nikolai Basargin and Nikolai Morozov organized a commune and founded a school of commune life that lasted from 1898 to 1902.\n\nIn the middle of the 20th century Borodino was also the place where Andrey Voznesensky had a summer dacha.\n\nThere is a railway station in Borodino.\n\n\n]" time="0.015"><properties><property name="score" value="0.0731577938382022" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07315779&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07315779
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Michal Ha\u0161ek\n\nMichal Ha\u0161ek (born April 8, 1980) is a Czech professional ice hockey goaltender who is currently playing with HC Pardubice of the Czech Extraliga. He previously played in the National Hockey League for the Detroit Red Wings, Chicago Blackhawks, Colorado Avalanche, Ottawa Senators, and Florida Panthers.\n\nHa\u0161ek spent the 1996\u201397 season with HC Pardubice's junior team, and then joined the Czech Extraliga club HC Vset\xedn, appearing in 29 games. Ha\u0161ek was a member of the Czech Republic national under-18 team at the 1997 European Junior Championships, playing in two games and posting a 1.51 GAA.\n\nThe next season, Ha\u0161ek joined HC Pardubice, which was then playing in the Czech 1. Liga. Ha\u0161ek played well for the team, and was chosen by the Colorado Avalanche in the sixth round, 162nd overall, in the 1998 NHL Entry Draft. He remained with HC Pardubice, and in his final season with the club, he posted a 2.30 GAA in 37 games. Ha\u0161ek was signed by the Avalanche, and on March 9, 2000, he made his NHL debut, against the Detroit Red Wings. He played two more games for the Avalanche before returning to the Czech Republic to join HC Pardubice for the 2000\u201301 season.\n\nHa\u0161ek played in 11 games for the Avalanche the next season, but spent the majority of the 2001\u201302 season with the Hershey Bears of the American Hockey League, where he posted a 2.20 GAA in 44 games. He remained in North America the next season, playing five games for Colorado and 59 for Hershey, and was then traded to the Chicago Blackhawks for a sixth-round pick in the 2004 NHL Entry Draft on March 10, 2003. Ha\u0161ek played well for the Blackhawks in 2003\u201304, with a 2.11 GAA in 37 games.\n\nThe next season, Ha\u0161ek played well for the Blackhawks, with a 2.22 GAA in 36 games. However, he was unable to maintain his form, and was replaced as Chicago's starting goaltender by Nikolai Khabibulin, who had been acquired by the Blackhawks in a trade with the Tampa Bay Lightning. Ha\u0161ek was then traded to the Ottawa Senators on March 9, 2005, for a second-round draft pick in the 2005 NHL Entry Draft.\n\nIn his first season with the Senators, Ha\u0161ek played in 40 games, and posted a 2.26 GAA. He then missed the majority of the 2005\u201306 season after suffering a shoulder injury. He appeared in just five games for the Senators, and was traded to the Colorado Avalanche for Patrick Eaves, a second-round draft pick in the 2006 NHL Entry Draft, and a third-round pick in the 2007 NHL Entry Draft on September 25, 2005.\n\nHa\u0161ek quickly returned to form with the Avalanche, playing in 54 games and posting a 2.22 GAA. In 2006\u201307, he played in 57 games, and posted a 2.30 GAA. Ha\u0161ek was an important factor in the Avalanche's playoff run, as he posted a 1.65 GAA in eight games.\n\nOn June 19, 2007, Ha\u0161ek signed a two-year contract extension with the Avalanche. The following season, he played in just 35 games due to a groin injury, and his GAA rose to 2.88. He was then sidelined for the rest of the season after having shoulder surgery.\n\nOn June 30, 2009, Ha\u0161ek signed a two-year contract with the Detroit Red Wings. He recorded the first win of his Red Wings career on January 2, 2010, in a 3\u20131 victory over the Chicago Blackhawks.\n\nOn January 1, 2011, Ha\u0161ek was named the First Star of the Week for the week ending December 26. Ha\u0161ek compiled a 3\u20130\u20130 record, a 0.67 goals-against average and a .984 save percentage. He led the Red Wings to their fifth four-game winning streak of the season. During the streak, Ha\u0161ek stopped 135 of 138 shots and yielded only one goal in a span of 140 minutes and four seconds.\n\nOn January 20, 2011, Ha\u0161ek was selected as the NHL's Third Star of the Week. He went 2\u20130\u20131 in three appearances, with both victories coming against Central Division rivals (January 14, 3\u20130 win at Chicago and January 16, 4\u20131 win at St. Louis), and recorded a 1.00 goals-against average and a .962 save percentage.\n\nHa\u0161ek finished the 2010\u201311 season with a career-high 41 victories, just one win behind league leader Martin Brodeur.\n\nOn March 29, 2011, Ha\u0161ek scored a goal against the Los Angeles Kings. The goal, a result of a breakaway due to an errant pass by Kings defenseman Slava Voynov, was the second of Ha\u0161ek's career. His first goal was scored against the Tampa Bay Lightning in 2008. The goal also made him the oldest goaltender in NHL history to score a goal in a game. The record was previously held by Tommy Salo, who scored a goal in a game at the age of 35.\n\nOn May 31, 2011, Ha\u0161ek underwent surgery to repair a torn labrum in his right shoulder.\n\nHa\u0161ek was injured during the 2011\u201312 preseason, and on October 2, 2011, was placed on long-term injured reserve by the Red Wings. He returned from his injury to back up Ty Conklin for a game against the Phoenix Coyotes on November 29, 2011.\n\nOn July 5, 2012, Ha\u0161ek signed a one-year, $2.05 million contract with the Florida Panthers. He was named the starting goaltender in the Panthers' inaugural game in the 2012\u201313 season against the New Jersey Devils. He was subsequently pulled in the first period after allowing 4 goals on 15 shots. He was replaced by Scott Clemmensen, and on January 24, 2013, Clemmensen was named as the Panthers' starting goaltender for the remainder of the season.\n\nHa\u0161ek was placed on waivers by the Panthers on June 13, 2013. Ha\u0161ek was then bought out from the final year of his contract on June 15, 2013. He was then signed by the Detroit Red Wings for the 2013\u201314 season to backup starter Jimmy Howard.\n\nHa\u0161ek became an unrestricted free agent in the off-season, and on September 11, 2014, signed his first contract abroad, agreeing to a one-year deal with HC Pardubice of the Czech Extraliga. After one season in Pardubice, Ha\u0161ek opted to continue in the Extraliga, signing a two-year deal with HC V\xedtkovice. In his second season with V\xedtkovice, Ha\u0161ek helped his club claim its first league title since 1999, appearing in 15 playoff games. He left as a free agent to sign a one-year contract with HC Pardubice on June 19, 2017.\n\nHa\u0161ek played for the Czech Republic at the 2002, 2003, 2004, and 2005 World Championships, winning bronze in 2004 and 2005. Ha\u0161ek was the starting goaltender for the Czech Republic team that won the gold medal at the 2004 World Cup of Hockey.\n\nIn 2003, Ha\u0161ek became a Canadian citizen, but had already represented the Czech Republic in international competition, and he was therefore eligible to play for either national team. On November 7, 2005, he announced that he would continue to represent the Czech Republic, even though he had been approached by the Canadian team and was eligible to play for Canada. Ha\u0161ek's younger brother Ond\u0159ej also plays professional hockey.\n]" time="0.071"><properties><property name="score" value="0.05083850203470381" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0508385&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0508385
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Helium leak\n\nA helium leak is a leak of helium from its container.\n\nHelium is a diatomic gas and therefore more difficult to confine than many other gases. The inert gas may diffuse through even very small leaks into other parts of a vacuum system, or it may leak into the atmosphere. Diffusion is facilitated by higher pressure on one side of a barrier. One way of countering this effect is to use helium's very low solubility in common materials. Helium also may leak into vacuum systems from poorly-made, improperly-applied, or damaged seals. \n\nHelium leak detectors use a small electric current through a coil inside the container to detect leaks. They are often found in high-vacuum and ultra-high-vacuum systems.\n\nAt room temperature, helium leaks through most materials at a rate of about one molecule every 10 seconds. If a]" time="0.014"><properties><property name="score" value="0.022320817776994017" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02232082&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02232082
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Cons\n\nCompany uses a very low standard of ethics in most cases. They do not care about employees, or helping them improve their lives. They have some of the poorest HR departments in the industry. They have a very slow turnover, and the company is very willing to allow people to be mistreated for long periods of time before taking action.]" time="0.013"><properties><property name="score" value="0.018902747490190386" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01890275&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01890275
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[$46,924\n\n$23/hr\n\n$2,533\n\nThe average publicity announcer salary in Alabama, United States is $46,924 or an hourly rate of $23. This is 11% lower (-$5,552) than the average publicity announcer salary in the United States. In addition, they earn an average bonus of $2,533.]" time="0.013"><properties><property name="score" value="0.011216471932495864" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01121647&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01121647
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Bitcoins are widely considered a great way to make purchases, but it is much harder to pay someone with them. That is exactly the case with Andreas Antonopoulos. He has reportedly been bombarded with requests from individuals who want to make donations to his projects. But he has no means of receiving them.\n\nIs Andreas Antonopoulos Being Misled?\n\nMost people know Andreas Antonopoulos by now. He is one of the most popular and prolific figures in the Bitcoin world. However, the widespread adoption of Bitcoin has led to a lot of new users, many of whom do not have a great understanding of the technology. That is not necessarily a bad thing, though.\n\nIt now appears as if a lot of people think Andreas Antonopoulos accepts donations, whereas he doesn\u2019t. As such, people keep bombarding the Bitcoin community icon with messages asking how they can make a donation. This leads to quite a few jokes being made about it on social media. However, Andreas Antonopoulos is well aware of the situation, as he keeps getting the same questions.\n\nHowever, the most intriguing part of this situation is how Andreas Antonopoulos makes it clear he doesn\u2019t accept donations. For some reason, some people continue to pester him about it. Some individuals even claim they are only able to donate a small amount of money, which would make no sense whatsoever.\n\nThe Andreas Antonopoulos Donation Account\n\nInterestingly enough, Andreas Antonopoulos is quite open about Bitcoin donations, but only to the extent of signing them over to projects. That is not what most people would expect, but it makes a lot of sense to give this money to a good cause. For example, Antonopoulos has donated over US$5,000 worth of Bitcoin to help rebuild the war-torn Marawi in the Philippines. It is evident this money came from donations he received, which is a very interesting development.\n\nThat being said, donations can still be made to Andreas Antonopoulos, even though it is done in a different way. Instead of sending money to his personal wallet, a public Bitcoin wallet has been created for donations to be sent to. That wallet has been active for some time now, and users can send money to a wide range of Bitcoin and Bitcoin Cash wallets.\n\nDo you think donations should be sent to Andreas Antonopoulos? Let us know in the comments below.\n\nImages courtesy of Shutterstock\n\nread original article here]" time="0.027"><properties><property name="score" value="0.004954253164204333" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00495425&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00495425
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Rottnest Island\n\nThe second island in the chain, after its neighbours Fremantle and Rottnest, is Barrack Street. In 1831, a convict ship called the \u201cDuke of Wellington\u201d (there are other tales about the origins of the name) ran aground here and was eventually abandoned. The women were moved to Fremantle Prison. The wreck is still there, but it is being eaten by the sea. You can walk to it on a path from Waterhouse Beach. The best way to see the island is by bike, though, because it is so small and flat. The bikes are provided by a company called Island Experience Rottnest. This is a great way to see the island. It costs around $25, which gets you a bike and the unlimited use of it for three hours. They give you a map and instructions, but you can see the whole island by walking in about an hour.\n\nThere are very few buildings here, but what there are have a very interesting history. For example, there is the Weeroona Hotel, the oldest operating hotel in the state. It is run by a company that is owned by the island\u2019s inhabitants. The oldest building is called the Round House. It was built in 1833 and has been restored to the 1830s period. It used to be the men\u2019s quarters. There is also a small cathedral that was built in 1901, and is called St Mary\u2019s Cathedral. There is also a cricket ground here that was originally established in the 1840s. This island was once the penal colony. The ruins of two of the prisons, the Governor\u2019s residence and the convicts\u2019 huts can be seen. They are located at a place called Settlement.\n\nCamel Rock\n\nThis is the most important landmark of the island. It is visible for many miles, and is sometimes mistaken for a camel. The sandstone is interesting, as it was formed from quartzite, which is a very hard stone. Some of the grains have been turned to iron oxide, which is a reddish brown. This was done by chemical reactions that were the result of oxygen and iron.\n\nSome rock formations can also be seen at the Quobba station, which is a restored 19th century railway station. There are a few lookouts here, but the best is at Arch Rock, which is a short distance from the main road. There are about 200 species of birds here, and there is a sanctuary that is managed by a conservation group called the Rottnest Island Conservation Society. This is home to the Tawny Frogmouth, the Black-breasted Button-quail, the Malleefowl, and the Rottnest Island White-tailed Black-Cockatoo.\n\nThere is also an Interpretive Centre that explains the history of the island and its different birds. Another reason to visit this island is the fact that the Aboriginal people, the Wadjuk, were here first. There is an Aboriginal reserve that is in the northern area of the island, and there is also a sacred site that can be seen there. The site is called Yanjul, which means \u201cman with spear\u201d. The local Aboriginals still refer to the island as Wadjemup, which means \u201cIsland of Peace\u201d. The First Fleet stopped at this island, and as such it is the only island in the world where two different nations fought for control.\n\nPlaces to Stay\n\nThere are a few different places to stay on this island. Most of the accommodations are self catering. There are also a few bed and breakfasts, and there is even a four star hotel called the Sheraton Rottnest Island Resort. There are over 250 different cottages to stay at, some of which have been converted from old army barracks. The Great Western Hotel is the only hotel here that has electricity and telephone connections, so you may want to stay there if you want to be near the center of things. There are many hotels here, but most of them are more expensive than the average.]" time="0.068"><properties><property name="score" value="0.06808807036633267" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06808807&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06808807
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[This great vinyl adhesive cloth tape is waterproof, flexible and conformable to curves. It features easy-to-remove release liner and 1,000 per roll. Use this tape to adhere vinyl cloth to flat or curved surfaces.]" time="0.012"><properties><property name="score" value="0.01832213002453953" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01832213&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01832213
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[I can't believe it. Yesterday, I got a letter from Robert D'Orta, the writer who's suing Paramount for copyright infringement over the Star Trek reboot. He wants to settle. And I've got to say, my heart just jumped. A settlement means that Paramount's going to have to pay some kind of damages, and I get to share in it. Maybe I get a check, maybe I get some percentage of whatever they owe to Robert. I can't believe it. That one line. &quot;Paramount's counsel has been in contact with me about settlement.&quot;It just keeps me going. It's the first time since this all started that I can say for sure that I'm not just some crazy person running after a scam. Robert D'Orta and I were going to be in the history books.I can't believe it. I'm really going to have to look at a settlement now. Who do I want? What's it going to take? What are the chances of me getting something from this?What's next?]" time="0.013"><properties><property name="score" value="0.03823646306520545" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03823646&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03823646
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[So, I never thought this day would come, but after several years of wrestling with this series, I have finally been forced to give up on it.\n\nI read The Fellowship of the Ring many years ago, and loved it. It was one of the books that inspired me to get into fantasy fiction. The author, J.R.R. Tolkien, was a legend in the world of fantasy fiction, and it was considered one of his greatest works.\n\nAt the time, I hadn't heard of this series. If I had, I probably would have thought &quot;Ew, it's another fantasy book.&quot; It would take me several years to learn that it's not another fantasy book. This is the one that started it all.\n\nThe story was great, and I really liked how Tolkien put all his imagination into it, and wrote a story that was unique and innovative. The story itself was fascinating. I loved the idea of Hobbits, which are little people that lived in an idyllic area, with their own language and customs. This book had a dark side, and even the humor in it was dark. There was a lot of death, evil, and destruction in the book. That's something that stuck with me, even though I really didn't appreciate it at the time. I was in my tween years, and just didn't understand it.]" time="0.012"><properties><property name="score" value="0.03631545624632193" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03631546&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03631546
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Reviews (15) Write a Review\n\nCool costume Perfect fit and great quality. Looks just like the picture. Happy with purchase\n\nDoesn\u2019t fit like the other girls ones! The body piece is way too long for a normal sized girl, there\u2019s so much extra material, you can barely make it look like the girl in the picture. It doesn\u2019t look as cute as the other girls ones, and the extra length isn\u2019t easy to shorten, because of the netting on the bottom.\n\nFun! Love this costume! My daughter is small and the medium was perfect! Can\u2019t wait to see her in it at school on Halloween!\n\nBeautiful! This costume is beautiful! It fits true to size and looks great!\n\nGorgeous costume I love this costume and it fits true to size. I have a slender 7 year old, ordered the size 6-8, and it fits well with some room to grow into. The only drawback I see is the flippers are sewn into the gloves so she won\u2019t be able to wear them separately, but that\u2019s a small sacrifice for the beauty of this costume!\n\nCutest! Love it! My 5 yr old loves her new swimmer costume. It's made well, super cute and just the right amount of &quot;bling&quot;!\n\nSO STINKING CUTE! This costume is exactly what I was looking for! It\u2019s super cute and fits well. My daughter is a 4t and I got her a 5-6 and it fits great! It\u2019s also not super thin. It\u2019s very well made!\n\nWonderful costume! Beautiful! Super cute and made well.\n\nExcellent quality! I\u2019m always apprehensive when ordering on-line, but was very pleased with the quality of this costume!\n\nBeautiful This costume is so beautiful! The colors are vibrant and fun, the quality is fantastic and the fit is perfect!\n\nSo pretty I bought this for my 4 year old daughter and she loves it! She is tiny so I ordered her a 4-6 which fit perfectly. It's well made and super pretty. Can't wait to see her wear it for Halloween\n\nadorable It is very pretty. The top is pretty sturdy. The bottom is light. It fits my 4 year old perfectly. My daughter is little and will not fit for very long. I love it!\n\nsuper cute This costume is so cute! I ordered the size 4-6 for my daughter who wears a size 4 in girl clothes and it fit perfectly. The material on the body piece is thick and the arms and feet are easy to get on and off, without being too loose.\n\nVery pretty I am very pleased with this costume. The colors are so vibrant and the quality is good. The size fits perfect.]" time="0.024"><properties><property name="score" value="0.0619463957343387" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0619464&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0619464
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[2013 Crime Rate Indexes North Little Rock, AR 72223 Arkansas United States Total Crime Risk 170.0 91.0 100 Murder Risk 165.0 94.0 100 Rape Risk 164.0 90.0 100 Robbery Risk 157.0 93.0 100 Assault Risk 162.0 92.0 100 Burglary Risk 155.0 89.0 100 Larceny Risk 105.0 92.0 100 Motor Vehicle Theft Risk 132.0 89.0 100\n\nThe data for North Little Rock, AR 72223 may also contain data for the following areas: North Little Rock\n\n\n\nCrime Risk Index (100 = National Average): Index score for an area is compared to the national average of 100. A score of 200 indicates twice the national average total crime risk, while 50 indicates half the national risk. We encourage you to consult with a knowledgeable local real estate agent or contact the local police department for any additional information.\n\n\n\nCrime Indexes are based on numerous current and historical datasets as well as proprietary modeling algorithms which estimate values at more granular geographic levels when specific data is either unavailable or impractical to aggregate. While every effort is made to ensure accuracy, these are estimates and should only be used as a guide. For detailed information regarding crime and safety in a community, please contact local law enforcement agencies.]" time="0.012"><properties><property name="score" value="0.009923610587309062" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00992361&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00992361
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Jonathan Daniel/Getty Images\n\nAccording to a source, the Miami Heat has signed Mike Miller to a two-year, $5.5 million contract, and may be on the verge of completing a deal with Rashard Lewis for the veteran minimum.\n\nThe Miller signing comes as the Heat are preparing to play the Chicago Bulls in Game 4 of the Eastern Conference quarterfinals.\n\nMiller averaged a career-low 4.8 points and 1.8 rebounds in just 22 games during the regular season.\n\nHe did not play in Miami\u2019s first-round series against the Milwaukee Bucks, after undergoing surgery to repair a torn tendon in his left foot.\n\nHe played in just two games during the second round series against the Indiana Pacers.\n\nThe Heat's move is somewhat surprising because they're already stacked at the wing position, with LeBron James, Dwyane Wade, James Jones, Mike Miller, and James Jones.\n\nMiller will be limited to roughly 25 minutes a game, according to a source.\n\nMiller has won two NBA titles with the Heat.\n\nHe was the MVP of the NBA Finals in 2006, and has averaged 10.7 points and 4.5 rebounds for his career.\n\nThe Heat are in need of frontcourt depth, because they only have three true big men on the roster, and could use more depth, with Greg Oden having his contract purchased by the team, but not likely to play during the playoffs.]" time="0.013"><properties><property name="score" value="0.11639800560440117" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.11639801&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.11639801
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Riot seems to be slowing down its champion creation process. Since the release of the current champion rotation, only five champions have been released. Each of these champions (except for Xerath) was tested in the PBE at some point, so we knew what to expect when the champion was released.\n\nXerath was the next to last champion to be released. He came out in the middle of the season and, because of his kit, has been popular for quite a while. He didn't really have that many people against him, and it didn't take him long to climb into the current meta. He was picked a total of 11 times in the North American Championship Series and showed a 65 percent win rate.\n\nZed, like Xerath, was released at the beginning of the season, but there are still people who have not really mastered him yet. His popularity in the NA LCS was fairly low and his win rate was only 58 percent. Zed's popularity has since risen to the top and he is now played regularly in the NA LCS.\n\nBrand is another interesting case. His play rate has risen from 0 percent in the NA LCS to 3 percent. His win rate is not that great (44 percent), but because he is a burst caster, he fits in with the current meta. As long as you have the ability to lock down one person for a period of time, Brand is pretty strong.\n\nFinally, Vi was the last champion to be released, but her popularity is still a bit low. She has been played nine times in the NA LCS, with a 43 percent win rate.\n\nWith those three champions, Riot has shown that it will not just dump out new champions on the game, and that champions will still have to be tested on the PBE before being released.]" time="0.022"><properties><property name="score" value="0.022710563128002206" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02271056&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02271056
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Photo by Gracie Malley for Cannabis Now\n\nAs California rolls out its new cannabis rules \u2014 the most relaxed marijuana laws in the country \u2014 state and local law enforcement officials are attempting to implement the new system with an unprecedented blitz of raids on illegal pot businesses.\n\nWith the stroke of Gov. Jerry Brown\u2019s pen, California became the largest state to legalize marijuana for recreational use and the sixth in the nation to make cannabis available to the general public. Yet, the history of cannabis in California has been a long one \u2014 and not without its rocky moments.\n\nCalifornia\u2019s medical marijuana laws were first approved by voters in 1996 and ushered in a sea change in cannabis culture and use. Yet even in the heyday of Proposition 215, there were raids on medical dispensaries, law enforcement shutting down shops for violation of local zoning codes and municipal zoning laws, local municipalities levying taxes on dispensaries and even offering up community space in which to open shops.\n\nThe result of these early legislative setbacks? The displacement of a long-established network of dispensaries and cultivators. Those who could no longer get by in the state moved to Oregon, Washington, Colorado and other legal states \u2014 essentially exporting the cannabis culture they had created. Now, a new generation of cultivators and distributors are starting businesses in California, with new methods and processes that are a result of the new cannabis industry.\n\nEven as recreational marijuana use became legal in California in January 2018, there has been an uptick in raids of these newly legal businesses, which remains illegal under federal law. In the first week of 2018, there were a reported 26 raids of legal dispensaries in the Bay Area, with 25 arrests and the seizure of $16,000 in cash and seven pounds of cannabis.\n\n\u201cLaw enforcement is going to go after illegal operations in any case,\u201d said Sean Donahoe, executive director of the California Cannabis Industry Association. \u201cThese folks who are doing illegal activities, that\u2019s something they have been doing all along.\u201d\n\nIn California, the rules on commercial cannabis businesses are numerous and complex, he said. The fact that a business may be operating legally, that doesn\u2019t mean it isn\u2019t being targeted for a raid, according to Donahoe. In order for businesses to get permits to operate legally, they have to first get permits to be established.\n\n\u201cSome of the communities, their local government entities have decided]" time="0.023"><properties><property name="score" value="0.04088981310885129" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04088981&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04088981
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Hillary Clinton was a no-show at the opening of a new museum dedicated to women's history and equality in New York City on Tuesday, a day after she cancelled campaign events because of a stomach virus that left her dehydrated and dizzy.\n\nClinton's Republican opponent Donald Trump and his wife Melania did not attend the high-profile opening either.\n\nClinton's doctor said in a statement Monday that the Democratic presidential candidate had been &quot;diagnosed with pneumonia&quot; on Friday.\n\nThe museum is dedicated to women's history and feminist causes. Clinton's ties with the National Museum of Women in the Arts go back many years.\n\nShe presented the museum's founding patron, Mercedes T. Bass, with a leadership award in 2009. The museum's chair, Susan Henshaw Jones, has been a top Clinton donor.\n\nA spokesman for Clinton said on Monday that the former secretary of state &quot;continues to feel better,&quot; but he didn't indicate when she would resume campaigning.\n\nCampaign spokesman Nick Merrill said the Democratic presidential nominee has had an allergy-related cough, and that &quot;several hours of sleep&quot; and some antihistamines helped.\n\nThe illness struck just after Clinton flew back to New York City from a weekend trip to California.\n\nOn Monday, her doctor Lisa R. Bardack said in a statement that Clinton had had an examination of her pneumonia in a chest CT scan.\n\nBardack said the Democratic nominee &quot;is recovering nicely&quot; and has been &quot;advised to rest and modify her schedule&quot; as she continues to recuperate.\n\nClinton abruptly left a 9/11 anniversary ceremony Sunday in New York City after feeling overheated, according to her campaign. She retreated to her daughter's nearby apartment, but later emerged and told reporters: &quot;I'm feeling great. It's a beautiful day in New York.&quot;\n\nShe left Sunday's event at Ground Zero shortly after she arrived. Her campaign said she felt &quot;overheated&quot; and departed for her daughter's apartment in New York's Flatiron neighborhood.\n\nVideo captured by a bystander showed Clinton apparently struggle to get into a vehicle as staff members held her up by the arms.\n\nClinton, 68, did not immediately disclose the pneumonia diagnosis. Her aides initially said she had overheated.\n\nBardack said Clinton is &quot;re-hydrated and recovering nicely&quot; after attending a 9/11 anniversary event Sunday.\n\nThe health issue was not mentioned at a fundraising event Monday in Manhattan for Democratic Senate candidate Ted Strickland of Ohio, according to participants.\n\nAsked if the candidate's health was a campaign issue, spokesman Jesse Ferguson said, &quot;No.&quot;\n\n___\n\nAssociated Press writers Jonathan Lemire in New York and Ken Thomas in Washington contributed to this report.]" time="0.026"><properties><property name="score" value="0.05173930474615561" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0517393&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0517393
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The former CEO of St. Louis-based Boeing said he will plead guilty to charges that he approved the improper use of millions of dollars in company funds on home furnishings, a casino and a separate investor-relations matter.\n\nThomas W. Jones, 67, was charged with approving improper expenditures at the request of his predecessor as CEO, Philip M. Condit, 63.\n\nBoeing Co. fired both men after the company's internal investigation uncovered what prosecutors called &quot;improper and illegal conduct.&quot;\n\nProsecutors said that, in addition to Condit, four other Boeing executives and a Chicago interior designer would plead guilty to various charges related to the kickback scheme.\n\nJones and Condit each face up to 10 years in prison and fines of $250,000 on the conspiracy charge. The other defendants face up to five years in prison and $250,000 fines.\n\n'Systemic fraud'\n\nU.S. Attorney James A. Graves Jr. described the alleged conspiracy as &quot;systemic fraud.&quot;\n\n&quot;This was a company that had an ethical culture, and we find it deeply disturbing that a company like that could have employees commit such fraudulent acts,&quot; he said.\n\nCondit, Boeing's former chief operating officer, said in a statement that he agreed to plead guilty &quot;to the company's disappointment, but to help resolve the situation and put this issue behind us.&quot;\n\nA number of the alleged kickbacks involved funds that were earmarked for aircraft engine development, according to the criminal information filed in the case.\n\nBoeing is based in Chicago, but the scheme occurred in St. Louis, which was home to its defense and space unit until 2003, when the company moved it to Washington state.\n\nBoeing spokeswoman Bonnie Rodney said Friday that the company has cooperated with authorities and overhauled its expense approval procedures. She said the company is not aware of any illegal expenses related to the company's government contracts.\n\n'Widespread misconduct'\n\nIn a filing Friday in U.S. District Court in St. Louis, prosecutors said that &quot;certain Boeing employees, including several executives, had conspired to engage in the improper and illegal approval of corporate expenditures.&quot;\n\nThe filings said that Boeing officials believed there was &quot;widespread misconduct&quot; that included theft of company funds and illegal personal use of company funds.\n\nIn a &quot;factual resume&quot; filed in court, the government alleged that the executives conspired to authorize company expenses for a variety of items, including theater and concert tickets and interior furnishings for Condit's home, without any business justification.\n\nThe money was funneled through a design company in St. Louis. Prosecutors alleged that about $2.5 million was spent on the house, including $130,000 for a floor-to-ceiling stone fireplace, and $18,000 for a home theater with stadium seating.]" time="0.027"><properties><property name="score" value="0.03349179282350596" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03349179&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03349179
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\nThis is a group for anyone interested in all things data science.\n\n\n\nWe host a variety of talks and events, including:\n\n\n\n\u2022Talks: Short tech talks, panels and panel debates, and longer presentations on a wide range of topics\n\n\n\n\u2022Workshops: Data science and coding workshops on Python, R, data exploration and analysis, visualization, and other topics.\n\n\n\n\u2022Meetups: Events around the world with data science-focused talks, panels, and panel debates\n\n\n\n\u2022Hackathons: Hackathons where attendees work on data science projects using Python or R.\n\n\n\n\u2022Data science team building: Escape the room and board games to improve teamwork, communication, and data science skills\n\n\n\nThis group is a project of the Data Science DC meetup (http://www.meetup.com/Data-Science-DC/).]" time="0.012"><properties><property name="score" value="0.002132137632343545" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00213214&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00213214
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The mooring cable holding the stricken Rena, which has been stranded on the Astrolabe Reef off the north east coast of New Zealand's North Island since October 5th 2011, is fraying at an &quot;unacceptably rapid rate&quot; and the ship could break up in as little as a week.\n\n\n\nIt could even go as early as Tuesday night, according to New Zealand authorities.\n\n\n\nThey had been expecting the ship to sink, as currents beneath the sea would pull the vessel into the reef, and it would begin to break apart as the tides took it apart. However the &quot;churning&quot; around the ship has actually been slowing the process of the ship disintegrating.\n\n\n\nMeanwhile the crews on board are clearing the decks of the ship, as the removal of the fuel is to begin today, a job which should take a week or two. The removal of the fuel would then free up the mooring cables and allow the ship to break apart and eventually sink.\n\n\n\nHowever if the cable did break, the pieces of the ship could drift away, or worse, float across the Hauraki Gulf and into the Bay of Plenty.\n\n\n\nThe most recent attempt to free the ship from the reef failed, with the bow slipping from the rocky coral. No one has been aboard the ship since a New Zealand Navy ship fired shots at it in January, when the Rena's crew threw their passports over the side.\n\n\n\nThe Chinese-flagged vessel broke free from its tugs in the Bay of Plenty and hit the Astrolabe Reef, spilling 1,700 tons of fuel into the sea. The damage to the reef has been assessed as being very severe.]" time="0.013"><properties><property name="score" value="0.08052250259791283" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0805225&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0805225
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[To help raise awareness of depression and suicide, a group of engineering students from the University of Waterloo have designed a robot that can create artwork with the help of artificial intelligence (AI).\n\nMental illness is a huge problem that often gets ignored, or underplayed. Although the majority of the time you can tell if a person is suffering, there are some cases where it can be harder to identify, especially if you're just talking to them on the phone.\n\nThe team of eight, led by Chris Cui, spent four months designing their robot, which is called DepressionBot.\n\nDepressionBot is now available to help raise awareness of depression and suicide 1:03\n\n&quot;The robot is meant to help raise awareness for depression and suicide,&quot; said Cui, a third-year industrial design student, in an interview with CBC's The Morning Edition host Craig Norris.\n\nIt's no secret that mental health issues affect a lot of people, but it's not often that we see people, especially students, take time out of their day to help those affected.\n\n&quot;It's a really personal subject to a lot of people, but there's a lot of stigma around it, so a lot of people don't want to talk about it,&quot; said Cui.\n\nSketchbook artwork from DepressionBot. (Submitted by Chris Cui)\n\nGetting the right artwork\n\nIn order to create the artwork, the team taught the robot to look for certain patterns in its black-and-white drawings, then draw similar patterns in its next sketch.\n\n&quot;We had a bunch of example pictures, some]" time="0.012"><properties><property name="score" value="0.022496619997602885" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02249662&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02249662
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[To be blunt, the video you linked to does not really prove anything one way or the other. That video is trying to make it seem like some statements made by Dan Abnett about the use of the term &quot;boy&quot; or &quot;son&quot; when applied to the Raven Guard during the Horus Heresy was a form of apologism for using &quot;boyz&quot; when referring to Orks. This is not the case. While it is true that Abnett's statements were inappropriate, the content of that video fails to really show it. For example, where the video claims that Abnett said &quot;Raven Guard don't ever call themselves boys&quot;, Abnett was in fact saying that Raven Guard did not ever call themselves &quot;boyz&quot; or any other similar slang. He specifically mentioned &quot;boyz&quot; in the context of Orks as an example, and it is the video that is now misrepresenting his words. In other words, Abnett's statements on the subject of Orks are accurate, and the video's attempt to show him in a negative light is based on a misrepresentation.\n\nAdditionally, your claim that the Imperium is &quot;far from perfect&quot; is not true, as the Imperium's culture is based around the worship of the Emperor, who has deified himself and is often referred to as a god by Imperial citizens. This means that the Imperium has absolutely no separation between religion and government, which means that your claim that the Imperium is &quot;far from perfect&quot; is also a form of apologism, as it is ignoring a key element of the Imperium that should be readily apparent to any Imperial citizen. The use of &quot;Apologetic&quot; is a misuse of the word and also misrepresents what the author is trying to say, as you have misrepresented the claim he is making.\n\nThe author's second &quot;claim&quot; is incorrect on many levels, and is also based on a misrepresentation of facts. Firstly, let's take a look at the quote he links to:\n\n&quot;If we wanted to do a straight-up Space Marines game, we would have done Space Marines, right? They\u2019re the coolest guys.&quot;\n\nThis quote is taken out of context. The quote actually goes on to explain why they chose to do the Space Marines as they did. In fact, the quote was in response to a question about why they decided to make Space Marines so &quot;stereotypically heroic&quot; in terms of a tone. The statement that follows in the quote states that it is because Space Marines are the &quot;coolest guys&quot; in the galaxy. What the author is doing here is taking a single quote and presenting it as if it is the entirety of the explanation of why Space Marines are the way they are in this game, when that is not the case.\n\nThe author's next claim is that he does not agree with the statement about the inclusion of certain races, but instead of explaining why he disagrees, he simply says that it is &quot;wrong&quot;. This does not actually prove that the statement is incorrect, and is a form of dog-whistle politics, as it is deliberately vague and meant to be interpreted as &quot;political correctness&quot; or &quot;SJW&quot; or &quot;Cultural Marxism&quot; and similar things, all of which are highly charged words. He could very easily have explained why he disagrees with the statement, but he did not. Instead, he chose to attempt to shut down any further discussion of the topic with a rather obscure quote, while also using a misrepresentation of the quote to try and do so.\n\nThe author's final claim is based on another misrepresentation of the source material. In his article, the author linked to an article that states that the developers actually took fan-feedback into account when designing the Space Marine Chapters. In fact, the developers mentioned that they were &quot;watching&quot; the fans' reactions to the game, and were taking note of how the game was received and what fans' concerns were, which is in line with what the developers have said about their reasons for making the Space Marine Chapters they did.\n\nOnce again, the author is trying to use his article as a form of dog-whistle politics, as he is taking these claims and trying to make it seem as if the developers were lying. However, even if the developers were lying about this, which they are not, that does not invalidate the author's argument. In other words, the author's claims that the developers are lying about this are irrelevant to the argument the author is making, and should not be considered when assessing the validity of his arguments.\n\nIn the author's first sentence, he claims that it is &quot;common knowledge&quot; that the Imperium of Man is based on Nazi Germany. However, I do not believe that this is actually true, as I have never heard this before, and the idea seems rather outlandish. In fact, the idea that the Imperium of Man is based on Nazi Germany is so outlandish that it makes me question the author's sanity. Furthermore, his claim is an appeal to emotion, as the author does not really explain why he thinks this, and instead simply uses an emotionally charged accusation.\n\nIn the author's second sentence, he claims that the Space Marine Chapters, as well as the Emperor, are based on &quot;fascist leadership&quot;. I have already addressed the author's claim about the Space Marine Chapters, and I will now address his claim about the Emperor. While I cannot speak for the Emperor's actual policies, he is still worshiped as a god by most of the citizens of the Imperium, which does not actually fit with the definition of &quot;fascist leadership&quot;. Furthermore, the author does not really explain how he came to the conclusion that the Emperor is based on fascist leadership, and instead simply asserts it without any real explanation.\n\nThe author's third claim is based on a misrepresentation of facts, and is yet another form of dog-whistle politics. His claim that a representation of the logo of the White Legion is on the flag of the Imperium is a misrepresentation of the source material. In the source material, it is explicitly stated that the White Legion and the Imperial flag are two separate things, with the author's claim making it seem as if the White Legion is the official flag of the Imperium of Man. This is, of course, untrue. In fact, the White Legion is not even in the same star system as the Imperium, so the White Legion's flag is actually irrelevant to the argument that the author is making.\n\nThe author's fourth claim is also based on a misrepresentation of facts, as well as yet another form of dog-whistle politics. The author claims that the developers stated that the Imperium is &quot;an alliance between Humans and aliens&quot;. While I cannot speak for what the developers actually stated, I can point out that they did not state that the Imperium was an alliance between Humans and aliens. The author's statement that this claim is &quot;common knowledge&quot; is false. In fact, if this was common knowledge, then the author would not have needed to claim that this was what the developers had said. Instead, he would have just cited the developers as having said it. In other words, his claim that this is &quot;common knowledge&quot; is also a form of dog-whistle politics, as it is a claim that is designed to be interpreted as &quot;political correctness&quot; or &quot;SJW&quot; or &quot;Cultural Marxism&quot; and similar things, all of which are highly charged words.\n\nIn the author's final claim, he says that the Orks are portrayed as a generic enemy of the Imperium, when in fact they are a threat to the Imperium. The author claims that &quot;Boyz&quot; is used as an insult by the Imperium, when in fact it is used to refer to any Ork. In other words, his claim that the Imperium is &quot;far from perfect&quot; is incorrect, and in fact is a form of dog-whistle politics.]" time="0.075"><properties><property name="score" value="0.0294890780862119" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The general consensus in psychology is that intelligence is a singular entity that can be measured and is unchangeable, and that any issues with intelligence will remain constant throughout a person\u2019s life.\n\n\n\nIn general, the accepted view of intelligence is based on a series of experiments carried out in the 1940s and 50s, commonly referred to as the IQ tests. These intelligence tests measured intelligence by identifying patterns and memory, and tested mental agility. But the modern view of intelligence is much broader, and focuses on a person\u2019s ability to learn and their capacity to understand their environment.\n\n\n\nResearch carried out by a group of psychologists in New York have challenged the accepted view, and shown that people\u2019s IQ levels do not necessarily remain the same throughout their lives.\n\n\n\nThey found that if a person has an early traumatic experience, then the level of intelligence is likely to reduce significantly.\n\nThe study\n\nThe study used a database of IQ scores for each person, which was then divided into three groups: people who were not exposed to early trauma, people who had experienced trauma before they were 18 and people who had experienced trauma before they were 16.\n\n\n\nThe study then looked at how the group with early trauma had scored after 16 years. What they found was that people who experienced trauma before the age of 16 scored lower on intelligence tests at the age of 30. Those who experienced trauma before they were 18 scored even lower than those who were traumatized before they were 16.\n\n\n\nThey also found that the difference between the group who experienced trauma and those who did not was three to five points, which is a significant amount.\n\n\n\nThese findings highlight the importance of understanding trauma and how it can affect people throughout their lives. It also shows that traumatic events can influence how a person learns and that this will have a lasting impact on their lives.\n\n\n\nThe other important thing to note from this study is that the effect of trauma was found to last for 20 years. Even though trauma had occurred so early in a person\u2019s]" time="0.027"><properties><property name="score" value="0.055302102976256644" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Unfortunately, in both cases the occupants were intoxicated. The first house we went into was this guy's.\n\n\n\n\n\n\n\n\n\nHere's a look at the bar. The beer fridge was in the shed in the back.\n\n\n\n\n\nWhat's a bachelor pad without a recliner?\n\n\n\n\n\nThe back yard had a trampoline. That's all the beer this guy was going to need for a while.\n\n\n\n\n\nNow we move to the house that had the party. There were lots of people at this house, and not a lot of time to take pictures. However, I did get a few that I think will give you an idea of the festivities.\n\n\n\n\n\nThe Docks!\n\nGood spot for the porta potty.\n\n\n\n\n\nWho's ready to rumble?!\n\n\n\n\n\nThe front lawn!\n\nGolf tournament!\n\n\n\n\n\nWhen all was said and done, we had collected 8 trucks, a garbage truck, and a car from both houses.\n\n\n\n\n\nOh, and no injuries.\n\nThe general area of the party was right next to a creek. There were numerous signs saying how you would be arrested if you went in the creek. Well, as it turns out, there was a reason for those signs.Last year, the creek flooded, and one of the houses nearby was torn from its foundation and carried downstream. This was the home's fate.]" time="0.013"><properties><property name="score" value="0.09012610424710688" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0901261&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0901261
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Giant scale model\n\n\n\nSide note\n\nBought in New York. My first gigapan, and I don't know if it worked. I used an auto timer on the camera, and moved the camera to a tripod as well, but I think I moved too quickly. I did a slight &quot;edit&quot; on it by changing the colors a bit and used a slightly different size to get a more accurate view of the whole object. I think it worked well. It has lots of things I don't recognize. Probably because I'm from a small town in southern Indiana, USA. I'm actually kind of suprised that it even fits in the frame.\n\nFull Resolution, if you want to look at a close up of the item\n\ngigapan.org/gigapans/72917/\n\nDone]" time="0.012"><properties><property name="score" value="0.06600458360362937" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06600458&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06600458
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[A human being is a part of the whole, called by us &quot;Universe,&quot; a part limited in time and space. He experiences himself, his thoughts and feelings as something separated from the rest - a kind of optical delusion of his consciousness. The striving to free oneself from this delusion is the one issue of true religion. Not to nourish the delusion but to try to overcome it is the way to reach the attainable measure of peace of mind.\n\nAlbert Einstein (1938)\n\nWho would have thought that the the famous scientist, whose ideas took a leap of imagination that seemed to defy common sense, had so much wisdom to share.\n\nEinstein's wisdom, delivered in a simple, direct manner, is of a philosophical nature. And yet, as we have seen, his wisdom is based on a scientific perspective, not a philosophical one. It is a perspective that, as we shall see, stands in contrast to the beliefs of the average person.\n\nThe ultimate source of Einstein's wisdom is his sense of awe. He speaks of being in awe of what he observed in nature, as if these observations of nature had moved him to the core of his being. But this awe was not the end of the matter for Einstein. He does not speak of being in awe and then simply stop and say: &quot;What a wondrous world it is.&quot; Rather, he looks to this sense of awe as a source of inspiration to do more.\n\nEinstein's wisdom is the result of a personal approach to understanding and the sense of awe that this approach has produced. It is, therefore, both a personal and a shared source of wisdom. The person sharing this wisdom is not speaking on behalf of some outside authority, like a philosopher or a scientist. His wisdom is simply the result of a personal quest, not a more generalized, impersonal, public investigation.\n\nEinstein did not discover his wisdom by himself. He spoke about his awe at the &quot;dignity&quot; of nature, by which he meant that the world of nature is majestic, not cheap. Einstein was aware of the sense of awe that his discoveries had evoked in him and he wished to share this feeling with others. In this way, Einstein's wisdom was, at its roots, a shared resource, because the more people who can feel and express awe at the majesty of nature, the better it is for all of us.\n\nIt is as though Einstein had an innate wisdom that he wished to share with others, and as though he felt compelled to do so. Einstein's quest for wisdom was not an abstract exercise in which he wished to understand nature or the meaning of life. Rather, Einstein was driven by a need to share his insights. As the physicist Stephen Hawking has pointed out, the very nature of the scientific enterprise is a social activity, and this is as true of Einstein as it is of Hawking himself.\n\nLike Hawking, Einstein had the good fortune of being born into a social context that could support his quest for wisdom. He lived in an environment in which he was exposed to an intellectual culture that had been nourished by many philosophers who were also his contemporaries. He could, therefore, both draw from the wisdom of his predecessors and challenge them by articulating the insights that he had achieved.\n\nIn addition, the new physics that he helped to develop has spawned many technologies that have radically transformed our understanding of the world and has introduced new scientific disciplines, such as quantum physics, that have enriched our understanding of nature. Our scientific knowledge has become a shared resource that can be used as a basis for developing a new kind of wisdom, one that does not exist in the history of human thought before the twentieth century.\n\nAs we can see, the search for wisdom and the sharing of wisdom are based on a wide array of factors. To create a context that will allow for the growth of a new kind of wisdom, one that can benefit the common good of humanity, it is necessary to invest in science and in a new philosophical perspective.\n\nIn other words, a new kind of wisdom can only arise as the result of a new synthesis, one that links scientific knowledge with a philosophical perspective. It is this new synthesis that will be explored in the final chapter of this book.]" time="0.069"><properties><property name="score" value="0.0042884647883296816" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Gone are the days of fighting with finicky and slow 3D printers; with the new Prusa i3 MK3S, you get the benefits of simplicity, accessibility, and a whole lot more. This third generation printer is the successor to the very popular i3 MK2. This time around, Prusa has made some notable improvements.\n\n\n\nThe Prusa i3 MK3S prints in high quality with a dual extruder. It\u2019s also compatible with a wide range of filaments including PLA, ABS, PETG, Nylon, and more. This model is an upgraded version of the original, and it has all the perks that will get you printing with no hassle.\n\n\n\nThe printer has a noise-reducing design that won\u2019t bother anyone, and it features a beautiful enclosure that makes it a joy to look at. This sturdy machine is compatible with both Linux and Windows.\n\n\n\nIf you\u2019re looking for a durable, efficient, and reliable 3D printer, this model is just what you need.]" time="0.016"><properties><property name="score" value="0.057159578538285016" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05715958&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05715958
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Onze melkveehouderij\n\nIn 1990, wij waren de oprichters van Lijst Harkema en verhuisden in 2004 naar Aagtekerke, waar de melkveehouderij in het verleden ook zat. Met de komst van de nieuwe melkveehouderij kwamen de melkveehouders van het dorp ook in de gemeenteraad. In 2011 ging het dorp verder met \xe9\xe9n kandidaat, die meteen werd gekozen.\n\nDe voorbije vier jaar, sinds het vertrek van Gert Mulder, heeft Lijst Harkema vier gemeenteraadsleden gehad: Eenmaal Dorine Schrier, daarna Ad van Haren, Peter Nieuwenhuis en uiteindelijk verkozen wij twee van de twee\xebnhalve tonne melkveehouders: Gerrit van der Poel en Marco Beutels. In onze partij Lijst Harkema mogen alle beslissingen eerst goed worden doorgerekend en beoordeeld door deskundigen voordat er ook maar \xe9\xe9n stap wordt gezet. Bij onze melkveehouderij in Aagtekerke is altijd aandacht voor natuurbeheer en het landelijk gebied.\n\nWe werken vanaf een bepaald aantal dieren ook samen met FrieslandCampina. Dit omdat we geen einde hebben aan het gebied waar we melkvee houden, ook al zijn wij de grootste houderij van heel West-Friesland. Het houden van meer dan 600 melkkoeien vraagt bijna dagelijks toezicht. Onze stalen zijn zorgvuldig afgestemd op de huidige voer- en veevo]" time="0.013"><properties><property name="score" value="0.2125272094376407" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.21252721&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.21252721
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Jonathan Chait at New York Magazine says we should start ignoring Donald Trump.\n\nTrump gets endless media coverage, which he then uses to insist he should be treated fairly by the media.\n\nThe fact that Trump is now president makes his demand for fair coverage a valid concern. So Chait\u2019s right that ignoring Trump isn\u2019t the solution.\n\nBut Chait is also wrong that Trump doesn\u2019t deserve any media coverage. He\u2019s just like any other president.\n\nJust not as good at it.\n\nChait notes that Trump uses the press to get out his message:\n\nThe president of the United States is an extremely powerful figure. Even if the entire country agrees that the president of the United States is, in fact, a buffoon who speaks only in unintelligible gibberish, the president of the United States is still very powerful.\n\nChait is correct. Trump may have been a complete loser in the private sector. But in politics, Trump\u2019s clearly been a winner.\n\nThere are many reasons Trump has been a success in politics. Some of these reasons have to do with his political positions. But most of the reasons have to do with his personality.\n\nTrump is the ultimate salesman. He is a chameleon, able to change his tone and style to appeal to his audience. He\u2019s also the consummate bullshit artist, able to make false promises sound plausible.\n\nAnd Trump\u2019s made one of his signature moves since he began his political career. He completely ignores his Democratic opponents and talks directly to his voters. He\u2019s smart enough to understand that his opponents aren\u2019t really interested in actually talking to him.\n\nTrump has always understood how to get his message out, even if the message isn\u2019t all that important. It\u2019s why he never had a detailed health care plan. He wanted to talk to people who agreed with him about how terrible Obamacare was.\n\nHe\u2019s never really been interested in a detailed plan. He\u2019s interested in getting media coverage, because he knows the media will talk about what he says.\n\nTrump is constantly in search of free media coverage. The media gave him endless coverage during the campaign, and he won. And they\u2019re giving him endless coverage now, and he\u2019s still president.\n\nThe problem is that Trump is terrible at getting his message out in an organized and coherent way. He\u2019s also terrible at actually getting anything done.\n\nTrump\u2019s behavior is extremely damaging. But the damage doesn\u2019t come from the fact that he\u2019s in the news. It comes from the fact that he\u2019s so terrible at what he does.\n\nPeople who work for him never know what\u2019s going to come out of his mouth. And they\u2019re never really sure what he\u2019s going to do.\n\nBut his supporters always know where Trump stands. He\u2019s going to get rid of illegal immigrants. He\u2019s going to bring back jobs from overseas. He\u2019s going to put coal miners back to work.\n\nThat\u2019s the message he has for his supporters. And they know that even if his administration doesn\u2019t do what he says, Trump is still fighting for them.\n\nAll presidents have to use the media to get their message out. The difference with Trump is that his message is essentially nonexistent.\n\nTrump isn\u2019t president because he has an interesting policy agenda. He\u2019s president because he knows how to use the media to his advantage.\n\nThat means that we have to talk about him, whether we want to or not.]" time="0.065"><properties><property name="score" value="0.008831501760555681" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0088315&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0088315
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[On the evening of October 23, 2017, the celebrated actor Will Smith hosted an invitation-only event in Las Vegas, hosted by UNICEF to support vulnerable children around the world. During his remarks, Smith reflected on his recent trip to Madagascar, where he met a family who lost their three children to diarrhea. \u201cWhen I say my faith was tested, it was tested. You understand? It was like I was in the Bible.\u201d\n\nSmith is a committed Christian. In 2007, when he was named the most bankable star in Hollywood, Smith made the bold claim that he\u2019s the only top-grossing actor who is a professed Christian. At the time, a study by the Pew Forum on Religion and Public Life found that just 18 percent of Americans consider Hollywood stars to be good role models, and just 10 percent said the same for politicians. \u201cIf the group that you are the leader of is viewed that way, you have to examine what you\u2019re doing.\u201d Smith said.\n\nOn this day, Smith was concerned about children around the world who are not receiving the care they need to survive. As he read about the dire need for vaccines, Smith thought about his own two kids, and how helpless he would feel if they were sick and dying. \u201cI can\u2019t imagine what that must be like. So for me, the easiest thing to do is to support UNICEF. Because when my children get sick, it\u2019s my job to get them help, and if they die, they die. I\u2019m not gonna stop that. I\u2019m not God. But when children around the world get sick and die, it\u2019s God\u2019s responsibility. And I have to believe that God will ask him, \u2018Are you there?\u2019 And when He says, \u2018Yes, I\u2019m there,\u2019 he\u2019ll say, \u2018Are you there?\u2019 And he\u2019ll say, \u2018I\u2019m there,\u2019 and God will say, \u2018Well, why didn\u2019t you help these children?'\u201d\n\nIn one year, UNICEF vaccinates 46 million children against preventable diseases. Every dollar invested in immunization results in $44 saved in treatment. UNICEF reaches children in 190 countries and territories. Each year, with the help of its many partners, UNICEF supplies more than 50 million children under the age of five with immunizations against the six life-threatening diseases: pneumonia, tuberculosis, diarrhea, polio, measles, and tetanus. UNICEF also supports the vaccination of pregnant women in countries where the practice has been shown to prevent transmission of vaccine-preventable diseases to newborns. In addition, UNICEF is the largest distributor of antiretroviral medicines for children and pregnant women with HIV in the developing world.\n\nUNICEF also provides thousands of training opportunities for community health workers to deliver health services to children in their own communities. UNICEF, through its extensive network of child protection partners, also provides psychological support and counselling for children and adolescents affected by violence, abuse and exploitation. For millions of children who would otherwise die, it is UNICEF that saves lives.\n\nIn response to the global refugee crisis, UNICEF is at the forefront of emergency response efforts to help children and families who are fleeing conflict and violence. The agency has distributed more than 1 million emergency packages in Syria, Iraq, Lebanon and Jordan, and has provided over 1.5 million people with emergency supplies including tents, kitchen sets, heaters and sleeping mats. UNICEF has established 41 child-friendly spaces for traumatized children in the region and continues to provide medical and psychosocial support, protection and assistance for refugees.\n\nIn 2016, UNICEF\u2019s commitment to the protection of children in conflict was formally recognized with the Nobel Peace Prize. The Nobel Prize was awarded to UNICEF \u201cfor its work for children in developing countries, and for long-standing advocacy in favor of children\u2019s rights.\u201d\n\nThe U.S. and UNICEF first partnered in 1953, to help children devastated by the Korean War. From that time, their relationship has strengthened. In fiscal year 2015, the U.S. government provided UNICEF with more than $77 million to support its humanitarian efforts. UNICEF also works with the U.S. government on a number of global health initiatives. These include ensuring that every child is fully immunized and preventing HIV transmission from mother to child. UNICEF also works with the U.S. government on the President\u2019s Emergency Plan for AIDS Relief (PEPFAR).\n\nThrough PEPFAR, the U.S. government has invested more than $80 billion to combat global HIV/AIDS since 2003. PEPFAR is a cornerstone of U.S. global health efforts, and it is the largest commitment by any nation to combat a single disease. The program provides 11 million people in over 50 countries with lifesaving antiretroviral treatment, and has contributed to a 40 percent decline in AIDS-related deaths since 2004. In fiscal year 2017, the U.S. contributed $4.2 billion to PEPFAR.\n\nChildren born with HIV and AIDS in Africa, including those living in remote rural areas, can now receive treatment. PEPFAR is working to ensure that every child is tested, identified, and started on treatment. New initiatives are also helping to treat HIV-positive mothers and prevent them from passing the virus on to their unborn children.\n\nIn the five years since the Global Health Security Agenda (GHSA) was launched, more than 40 countries have invested in the development of a National Emergency and Medical Preparedness and Response Plan. To date, more than 15 countries have published plans. In 2016, the GHSA received a boost when the U.S. committed to invest $1.3 billion through 2020. This funding will support the purchase of 50 million doses of vaccines to help save lives from infectious diseases like]" time="0.077"><properties><property name="score" value="0.04527513290590578" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04527513&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04527513
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[1. I cannot stop smiling, I am so happy.\n\n2. I saw Ellie today and it made me feel really good.\n\n3. My boyfriend\u2019s been away for a couple of days and it made me feel really bad.\n\n4. I have so much to do and so little time to do it all.\n\n5. But I am really happy with my life at the moment.\n\n6. I have been getting drunk and naked with the birds and it makes me feel like a teenager again.\n\n7. I bought a slice of delicious cherry pie today and it made me feel really guilty.\n\n8. But the guy at the deli was really cute so I might go back to buy more.\n\n9. I went out with some friends last night and it made me feel really old.\n\n10. We\u2019re all growing up too fast.\n\n11. We\u2019re all so busy.\n\n12. And our lives are changing.\n\n13. And I am really enjoying my life at the moment.\n\n14. And that makes me happy.\n\n15. And I\u2019m not gonna lie, it makes me feel really good to write all of this stuff down.]" time="0.013"><properties><property name="score" value="0.032362911210376266" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03236291&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03236291
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[2017.01.15. Szerz\u0151: Kir\xe1ly Zolt\xe1n\n\nAz elm\xfalt id\u0151szakban t\xf6bbsz\xf6r is m\xe1r t\xe9m\xe1t tett\xfcnk a honlapra, amelyekben a kiv\xe1ndorl\xe1s \xe9s a bev\xe1ndorl\xe1s probl\xe9m\xe1j\xe1r\xf3l, valamint a bev\xe1ndorl\xe1s befektet\xe9si lehet\u0151s\xe9geir\u0151l besz\xe9lt\xfcnk. Az itt szerepl\u0151 \xedr\xe1sokat mindenkit arra k\xe9rj\xfck, hogy ismertesse aj\xe1nl\xe1sainkat, s\u0151t: k\xf6zrem\u0171k\xf6dj\xf6n vel\xfck, ha \xe9s amennyiben kedve tartja. Hogy v\xe9g\xfcl ez az egy kicsit kev\xe9sb\xe9 szeml\xe9letes le\xedr\xe1s is az id\u0151szakon \xe1ttekinthet\u0151 legyen, t\xe9m\xe1nk megoszt\xe1s\xe1val, szeretn\xe9nk sz\xednesebb\xe9 tenni azt.\n\nBev\xe1ndorl\xe1s k\xfclf\xf6ld\xf6n: jelentkez\xe9s az online k\xe9pz\xe9sekre!\n\nEz a k\xf6zel-keleti sz\xe1rmaz\xe1s\xfa konzul bejelent\xe9se, hogy magyar diplom\xe1val is lehet amerikai \xe1llampolg\xe1r. Persze ez csak amolyan k\xf6lts\xe9ges tan\xfas\xedtv\xe1ny, amivel t\xf6bbek k\xf6z\xf6tt az amerikaiakn\xe1l \xe9s a sk\xf3t f\xfcggetlens\xe9gi referendumn\xe9l is seg\xedtette ezt az orsz\xe1got. Nem akarunk err\u0151l semmit el\xe1rulni, mert igen sokan csak ezzel a k\xfcl\xf6nleges diplom\xe1val mentek t\xf6megesen az USA-ba. Szerint\xfcnk ezt a t\xe9m\xe1t egy kicsit megk\xf6nnyebb\xfclj\xfck, ha leiratjuk a \u201eRe: \xc9rdekl\u0151d\xe9s angol nyelv\u0171 m\u0171szaki k\xe9pz\xe9s\xe9re\u201d \xe9s \u201eRe: F\xf3kuszt\xe1bor aul\xe1s k\xe9pz\xe9s, IAS vagy EC\u201d feliratokat is. Ugyanis a bev\xe1ndorl\xe1s nyitott k\xe9rd\xe9s, nyilv\xe1n nem tudni, hogy megmaradhat-e a hazai di\xe1kvizsga, vagy azt is el kell-e engedni, amihez a hatalom t\xf6bb k\xf6zt k\xe9pes \xe9s akar is. \xdagy t\u0171nik teh\xe1t, hogy az elm\xfalt id\u0151szakban nem ez a t\xe9ma kapta el a f\u0151sodor f\xf6l\xe9ny\xe9t, hanem ink\xe1bb az, hogy a diplom\xe1kat haszn\xe1lva bev\xe1ndorl\xe1sra k\xe9sz\xfclnek, mert szeretn\xe9k k\xfclf\xf6ld\xf6n folytatni tanulm\xe1nyaikat.\n\nDe mi\xe9rt is t\xf6rt\xe9nne ennyire sz\xe9lesk\xf6r\u0171 bev\xe1ndorl\xe1s, ami a diplom\xe1t kap\xf3k sz\xe1m\xe1t is mag\xe1ba foglalja? Mert a diplom\xe1val rendelkez\u0151k ugyan egyre kevesebben vannak, m\xe9gis akad b\u0151ven olyan, aki tud \xe9s ak\xe1r \xf6r\xf6m is a tanul\xe1s, a munka, de legal\xe1bbis a k\xfclf\xf6ldi \xfat. Ugyanis nemcsak a magyar fels\u0151oktat\xe1s gondot szenvedett, hogy az elm\xfalt h\xfasz \xe9vben az egyre k\xe9pzettebb munkaer\u0151 l\xe9tsz\xe1m egyre cs\xf6kkent, hanem m\xe9g a diplom\xe1val is rendelkez\u0151k sz\xe1ma is egyre kisebb. Az emberek sz\xe1m\xe1ra a k\xfclf\xf6ldi p\xe1ly\xe1zatokat is n\xe9zve, igazolni tudom, hogy nincs egy\xe1ltal\xe1n olyan k\xe9pz\xe9s, amelynek ne lenne r\xe9szben- vagy teljesen online k\xe9pz\xe9se is. Ugyanis a mai informatika, oktat\xe1s \xe9s kult\xfara nagyon nagy szerepet kap a dolgokban, ez\xe9rt \xe9rdemes elgondolkodni azon, hogy nem kell k\xe9nytelen vagy b\xe1rmi okb\xf3l kiszolg\xe1ltatott \xe1llapotban v\xe9gezni az oktat\xe1st.\n\nAz online k\xe9pz\xe9sek jelentik az \xfajat, amit ma a diplom\xe1val rendelkez\u0151k a legnagyobb ar\xe1nyban vesznek ig\xe9nybe, \xe9s szerintem nem is egy rossz d\xf6nt\xe9s, ha valaki le is teszi az online k\xe9pz\xe9st. Ami m\xe9g \xe9rdekesebb az eg\xe9szben, az az, hogy nem csak a bev\xe1ndorl\xe1sra k\xe9sz\xfclnek az online k\xe9pz\xe9sek seg\xedts\xe9g\xe9vel. Val\xf3j\xe1ban m\xe9g a saj\xe1t orsz\xe1gunkban is egyre t\xf6bb jelentkez\u0151 van olyan online k\xe9pz\xe9sekre, amelyekkel m\xe1s orsz\xe1gokban t\xf6megesen jelentkeznek. Ez teh\xe1t el\u0151ny \xe9s h\xe1tr\xe1ny is egyben, de szerint\xfcnk a legt\xf6bben ink\xe1bb a jelentkez\xe9s k\xf6vetkezt\xe9ben a szerencs\xe9j\xfckre vannak. De hogy mi a jelentkez\xe9s eredm\xe9nye? Kiv\xe1ndorl\xe1s vagy bev\xe1ndorl\xe1s, de a legt\xf6bb helyen egyel\u0151re m\xe9g azt sem lehet tudni, hogy van-e lehet\u0151s\xe9g a bev\xe1ndorl\xe1sra, vagy sem. De a fenti c\xedmen \xe9rdemes bejelentkezni, \xe9s ut\xe1naj\xe1rni, hogy van-e itt online k\xe9pz\xe9s. K\xe9rj\xfck, hogy t\xf6bbsz\xf6r is ut\xe1naj\xe1rjon, mert a keres\xe9s egy m\xe1sik c\xedmre is el\xe9rhet\u0151, \xe9s egy j\xf3 okos program lehet\u0151v\xe9 teszi azt, hogy az eredm\xe9nyek szeml\xe9ltet\xe9se legyen a lehet\u0151 legjobb.\n\nA magyar diplom\xe1val is lehet amerikai \xe1llampolg\xe1r?\n\nA t\xe9m\xe1t komolyan vegy]" time="0.076"><properties><property name="score" value="0.2983433911359419" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.29834339&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.29834339
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Twitter is the perfect platform to get the word out there about anything you are trying to promote. I am a huge fan of using Twitter as a marketing tool, even more than Facebook. It\u2019s a great way to network and reach out to people who are interested in what you have to offer. Here is my list of 10 tips to help you promote anything you are selling on Twitter:\n\n1. Use Hashtags\n\nI love using hashtags on Twitter. They help to promote the things you are interested in and create a sense of community within a group. I will admit that it can be hard to remember hashtags to use for all the things you promote. One way I suggest to overcome this is to put your hashtags in a Twitter search to make sure they are not already being used. This will save you some time and help you avoid any issues down the road.\n\n2. Retweet\n\nIf you like a tweet, retweet it! I\u2019m sure most people are familiar with the concept of retweeting, but I just wanted to put it out there to make sure it was stated. Retweeting is a great way to share the things you like, and can be a great way to show your audience that you are involved in the social media community and help them feel more comfortable doing business with you.\n\n3. Mention\n\nMentioning someone on Twitter is a great way to show your audience that you are engaged and want to interact with them. It is important to mention other people within your field of interest in order to show your audience that you are well connected. I always make sure to mention people who have liked or retweeted my stuff as well. It makes me feel good when I get a mention and it shows that the people I engage with are interested in what I have to say.\n\n4. Post About Your Daily Life\n\nThere is nothing wrong with posting about your daily life and about other stuff you like. I\u2019m sure you all know about the Twitter feeds that are all business. I don\u2019t like following people like that because I feel like it takes away from the idea of the social aspect of Twitter. By being a person and talking about your life, you can show your audience that you are a real person and are interested in other people\u2019s lives as well. I\u2019m sure you have seen some tweets about people\u2019s lives that you thought were really cool, and the people that posted those tweets have engaged you because of it.\n\n5. Be Social\n\nDon\u2019t forget to be social. I\u2019m sure that when you first started out you didn\u2019t know what to do on Twitter. My suggestion to you is to find some people who are interested in the things you are and start a conversation. You will see that the conversations that start on Twitter are real and can be interesting. One of my favorite things to do on Twitter is to start conversations with people. It gives me the chance to learn from other people and also helps them learn about me.\n\n6. Use Pictures\n\nI love using pictures on Twitter. I know a lot of people like to include pictures when they post about things, and I would like to suggest that you do so as well. Including a picture with your tweet can get you more retweets and likes, and can really help to boost the effectiveness of your post. I like to use pictures from Instagram with my tweets, but you can really use any picture. If you want to make a picture of yourself that has an impact on your audience, I would suggest using a tool like Socialbakers.\n\n7. Use Hashtags That Make Sense\n\nI\u2019ve touched on this before, but I want to emphasize it again. It is important to make sure that you are using hashtags that make sense. You don\u2019t want to hashtag your tweets with irrelevant hashtags just to get a retweet or a like. You also don\u2019t want to make hashtags that are too long. I would suggest keeping your hashtags short and relevant.\n\n8. Direct Message\n\nThere are times when a DM is more appropriate than a tweet. You will be surprised at how many people DM you. I think a lot of people just use it to ask questions, but you can also use it to network. I think a lot of people don\u2019t use DMs because they don\u2019t think their audience will see it. If you are going to send a DM, make sure it is personal. Send it to someone who is following you. It\u2019s a lot less likely that they will check the people they follow to see who is sending them a DM, so if you DM a person that is following you, they are much more likely to respond.\n\n9. Use Images\n\nI want to encourage you to use images with your tweets. Images are the first thing people look at, so you want to make sure that you are using good images. One thing I like to do is to include images with my tweets. I\u2019m sure you have seen a lot of Twitter profiles that do this. It helps to keep your tweets interesting and make your tweets more unique.\n\n10. Be You\n\nThis last one is really important. You want to be yourself on Twitter. The more genuine you are, the better. People can see through a fake personality. You want to engage with your audience and show them who you are. When they see you as a real person, it will help to build a more trusting relationship between you and your audience.\n\nThere you have it! I hope these tips help you to improve your Twitter account and that your business gets better because of it. I always love getting comments from you all, so if you have any more suggestions or questions, feel free to drop me a comment below.]" time="0.079"><properties><property name="score" value="0.01776150119186483" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Auszeichnungen\n\nBuchnerstatterin des Jahres, \xd6sterreichischer Buchpreis 2017, \xd6sterreichischer Phantastikpreis 2017, Burgenl\xe4ndische Literaturpreise 2017, Internationale Literaturpreise der Wiener Buchmesse 2018, \xf6sterreichische Kurzgeschichtensammlung des Jahres 2017, Deutscher Phantastik Preis 2018\n\nAwards\n\nAustrian Book Prize 2017, Austrian Fantasy Award 2017, Burgenland Literary Prize 2017, International Literary Awards of the Vienna Book Fair 2018, Austrian Short Story Collection of the Year 2017, German Fantasy Award 2018\n\nThema des Monats]" time="0.015"><properties><property name="score" value="0.019932935362776405" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01993294&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01993294
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Watermelon Fried Rice\n\nPrint This\n\nIngredients:\n\n1 can chicken broth\n\n1 lb boneless chicken, cut in strips\n\n3 cups cooked rice\n\n2 tablespoons butter\n\n2 tablespoons olive oil\n\n3/4 cup fresh cilantro, chopped\n\n2 cups watermelon, seeded and chopped\n\n2 green onions, sliced\n\n3/4 cup red bell pepper, seeded and chopped\n\n1/2 cup red onion, diced\n\n1/4 cup cashews\n\n1/4 cup raisins\n\nDirections:\n\nCombine chicken broth and chicken in a bowl. Let soak in refrigerator for at least 30 minutes.\n\nWhen ready to cook, remove chicken from broth, but reserve broth for the sauce.\n\nIn a skillet, melt butter with olive oil over medium-high heat. Add chicken and saut\xe9 until chicken is done and lightly browned.\n\nReduce heat to low. Add rice, cilantro, watermelon, onions, bell pepper, cashews and raisins. Stir until well mixed.\n\nAdd reserved chicken broth, a little at a time, until it reaches desired consistency.\n\nServes 4.\n\nRecipe compliments of American Farm Bureau]" time="0.015"><properties><property name="score" value="0.08928288448546429" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08928288&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08928288
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Check out the info below for future tour dates, and track them to receive alerts when they tour near you\n\n&quot;The one and only&quot; as his legions of fans call him, John Lydon is the only original member of the world's most celebrated punk band, PiL.\n\n\n\n'All the very best, Johnny Rotten!' The Independent\n\n\n\n'It was awesome!' The Guardian\n\n\n\n'He's got some music to make, let him do his job' Mojo\n\n\n\n'The People's Punk. Still Rotten after all these years' Sunday Times\n\nJohn Lydon John Joseph Lydon (born 31 January 1956), also known by his former stage name Johnny Rotten, is an English singer, songwriter, and musician. He is best known as the lead singer of the punk rock band the Sex Pistols, which lasted from 1975 to 1978, and again for various revivals during the 1990s and 2000s. He is also the lead singer of post-punk band Public Image Ltd, which he founded and fronted from 1978 until 1993, and again since 2009.\n\n\n\nLydon's musical career has spanned over four decades. As a musician, he has also been a member of several bands: PiL (1978\u20131992, 2009\u2013present), the Flowers of Romance (1978), the Joneses (1979), and the Nightingales (1979\u20131980). Since 2013, he has put his PiL bandmates on hiatus to focus on his new band, the Public Image Ltd band. He has also hosted a weekly radio show on BBC Radio 6 Music, Rotten Radio.\n\nJohn Lydon tour dates]" time="0.013"><properties><property name="score" value="0.12921503547904667" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.12921504&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.12921504
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Residence No. 8,\n\nA House with a strong connection to our Imperial Russian history, is situated in a quiet suburb of St Petersburg and has every luxury one can wish for. It\u2019s a real delight to visit the beautiful private beach of your own. The property is furnished in the best Russian style of the late 19th century.\n]" time="0.014"><properties><property name="score" value="0.044465596826964715" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0444656&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0444656
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[\n\nARTIST: Hot Tuna\n\nALBUM: First Pull Up, Then Pull Down\n\nYEAR: 1970\n\nCD REISSUE: no\n\nSHUFFLE: no\n\n\n\nTRACKLIST\n\n01. Somewhere In Between\n\n02. Highway Blues\n\n03. Alabama Blues\n\n04. All Right, Ok, You Win\n\n05. Low Down Blues\n\n06. Man On The Street\n\n07. Up To My Neck In Blues\n\n08. Teardrops Will Fall\n\n09. Mandy's Song\n\n10. Something To Look Forward To\n\n11. Just Between You And Me\n\n12. Muddy's Blues\n\n\n\n]" time="0.017"><properties><property name="score" value="0.11851068020675704" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.11851068&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.11851068
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[We provide comprehensive, customized solutions and a wide array of products. Some of the products we offer include:\n\nActual LED\n\nApplications\n\nAviation\n\nBulk Acrylic\n\nBusiness Cards\n\nCable\n\nClip-on\n\nCord\n\nCustom\n\nDisplay\n\nExit\n\nFlush\n\nFlush Mount\n\nGlass Mount\n\nJoint\n\nLight\n\nLit\n\nLouver\n\nMail Box\n\nMedical\n\nMedical Grade\n\nMezzanine\n\nOccupancy\n\nPanels\n\nRail\n\nSign\n\nSlim\n\nSpiral\n\nStripe\n\nTaper\n\nTranslucent\n\nWindow\n\nVisit our online store at shop.ledwholesale.com to learn more about our products and to place an order.]" time="0.022"><properties><property name="score" value="0.08780730040941487" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0878073&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0878073
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The Ultimate One Night Stand in the Square\n\nChapter 1: Proposal\n\nAs a bridesmaid, you're required to do a number of things that I'd never consider doing voluntarily. First of all, you're expected to wear a hideous dress that you've never worn before. Second, you're required to spend your evening catering to everyone else, acting as the host of the event. Most importantly, however, you're required to smile a lot and appear happy.\n\nAt the wedding I attended, everyone was happy, especially the bride and groom. I was happy for them, but I wasn't the happiest bridesmaid in the crowd. In fact, I was the exact opposite.\n\nWhy was I so unhappy? I wasn't the one getting married; I was the maid of honor. I was so annoyed that I was even standing up for my ex. We'd broken up last month because I thought he was cheating on me with a blonde bimbo named Summer. I was at the wedding because my best friend Alicia insisted that I go, but if she knew how much I hated the groom, she would have probably asked me not to go at all.\n\nAnyway, this was the part of the wedding that was pissing me off: the proposal. I was trying my hardest to hold back my tears, but I was failing. I didn't love him anymore, and I never would again.\n\nI glanced over at my ex-boyfriend, Blake. He was really handsome, but that's not what was important right now. He was going down on one knee and was about to propose to the most beautiful girl I'd ever seen.\n\n&quot;Andi, will you marry me?&quot;\n\nI cringed. He'd asked me the same question about ten times, but I was too busy being angry at him to remember. This time was different, however, because he was now asking the blonde bimbo he was sleeping with.\n\n&quot;Yes,&quot; she said, finally accepting the proposal.\n\nAs soon as he slid the ring on her finger, everyone in the room erupted into cheers. I just stood there, my mouth wide open, in complete shock.\n\n&quot;Yes, that's my girl,&quot; said Blake, &quot;And now she's my fianc\xe9e!&quot;\n\nAs they walked out of the church and into the limo, Blake leaned over and kissed her cheek. A couple seconds later, I lost it.\n\n&quot;Bastard,&quot; I muttered. &quot;Bastard.&quot;\n\n&quot;Andrea?&quot; said Alicia.\n\n&quot;He just asked that whore to marry him,&quot; I said. &quot;It was my turn. I was the one who should have been there. I'm the one who was supposed to marry him!&quot;\n\n&quot;Yeah, well I know that,&quot; she said, rolling her eyes. &quot;But it's over now. You just have to be a big girl and accept it.&quot;\n\n&quot;He doesn't love her,&quot; I said, &quot;I know it. He was with her while he was with me, remember?&quot;\n\n&quot;Andi, I hate to say this, but it's too late now,&quot; she said. &quot;He's getting married. You can't do anything about it.&quot;\n\n&quot;I could show up at the wedding,&quot; I said, &quot;and say that I'm his girlfriend and I'm pregnant.&quot;\n\n&quot;Andi, that's insane,&quot; she said. &quot;You know it's not your baby. You're going to destroy his wedding day if you pull something like that.&quot;\n\n&quot;That's the point,&quot; I said. &quot;I want him to feel the way I feel. I want him to feel miserable for a change.&quot;\n\n&quot;Look,&quot; she said, &quot;I'll go with you, but you can't tell Blake anything. If you do, you're going to be the biggest loser in the room, and you're going to lose your best friend.&quot;\n\nI hesitated. I knew she was right. If I ruined the wedding, I'd be in a lot of trouble.\n\n&quot;Fine,&quot; I said. &quot;I won't tell him anything.&quot;\n\nA few days later, I arrived at the church. The doors were open, so I wandered in. The church was completely empty, which made it look a lot bigger. I could hear some talking coming from the altar, so I walked in that direction. The first person I saw was Blake, wearing a brand new suit and looking like the happiest man on the planet.\n\n&quot;You look so beautiful,&quot; he said to the blonde bimbo who stood next to him.\n\nI thought I would be furious when I saw her, but I wasn't. I was heartbroken and sad, but I didn't hate her. In fact, I didn't feel anything towards her. I was just jealous, and the sight of her made me remember how much I'd wanted to marry Blake myself.\n\nI walked up the aisle, ignoring the fact that I was completely alone. I sat down in the front row, and began to think about what had happened.\n\n&quot;How did it happen?&quot; I said to myself. &quot;Why did we break up?&quot;\n\nI knew that I loved him, but I was certain he didn't love me. My best friend Alicia and I had discussed it, and she agreed with me. She was convinced that Blake was cheating on me with a blonde bimbo named Summer, but I wasn't convinced.\n\n&quot;It can't be her,&quot; I said. &quot;How did he meet her? Where did they meet?&quot;\n\nJust then, a woman walked up to the front of the church. She was dressed in a black robe, and she had short, black hair.\n\n&quot;Are you Andrea?&quot; she asked.\n\n&quot;Yes,&quot; I said.\n\n&quot;I'm sorry,&quot; she said, &quot;but the wedding has been cancelled.&quot;\n\n&quot;What do you mean?&quot; I asked.\n\n&quot;Blake is getting married in Las Vegas,&quot; she said.\n\n&quot;But I'm his fianc\xe9e,&quot; I said, &quot;And I'm supposed to be here.&quot;\n\n&quot;You were never his fianc\xe9e,&quot; she said. &quot;The bimbo, Summer, is the one he's getting married to.&quot;\n\n&quot;I'm sorry,&quot; I said. &quot;I don't understand.&quot;\n\n&quot;Neither do I,&quot; she said, &quot;but I was supposed to tell you the wedding was cancelled. So, that's what I'm doing. Sorry for ruining your night.&quot;\n\n&quot;It's okay,&quot; I said. &quot;It's not your fault.&quot;\n\nI knew I had to get out of there. I could barely stand being there as it was, so I hurried out the door. I walked down the sidewalk, not knowing where I was going. I had no idea what to do next.\n\n&quot;Hey,&quot; said a voice behind me.\n\nI turned around and saw Blake running towards me.\n\n&quot;You were supposed to be at the wedding,&quot; he said. &quot;Where have you been?&quot;\n\n&quot;I'm not going to lie to you,&quot; I said. &quot;I'm really hurt, and I don't want to be in your wedding. I'm just here because your father paid for the wedding.&quot;\n\n&quot;You don't understand,&quot; he said. &quot;We're getting married in Las Vegas tonight.&quot;\n\n&quot;Really?&quot; I said.\n\n&quot;Andi, we've got to go,&quot; he said, &quot;it's going to start any second.&quot;\n\nI took one look at his face, and I knew that he was telling the truth. I felt the tears well up in my eyes.\n\n&quot;So, this is it?&quot; I said. &quot;You're really getting married?&quot;\n\n&quot;Andi, I'm sorry,&quot; he said. &quot;I didn't know how to tell you. My father forced me into this, and I couldn't say no. If I had it my way, you'd be my wife right now. But this is the way it has to be.&quot;\n\nI felt a tear escape my eye, and I quickly wiped it away.\n\n&quot;You look beautiful,&quot; he said.\n\n&quot;I'm not going,&quot; I]" time="0.098"><properties><property name="score" value="0.09133079647121296" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0913308&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0913308
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[IMG_2720\n\nJune 26, 2018. This evening we worked on wrapping a glider for travel to a new home. Pictured here is Matt Berger of Berger Aviation taking care to wrap the craft. (I love the old mobile home!)]" time="0.018"><properties><property name="score" value="0.03293150274196188" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0329315&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0329315
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[4.25 stars - Review by Trisha\n\n\n\nIn my experience, most male-on-male romance books are really hot and steamy but the storyline is lacking, but this one definitely surpassed my expectations. This book wasn't hot and steamy. Not even a little bit. There was one sexy scene between Jay and Davey but that was about it. But it was such a great story. The storyline kept me interested the whole time and I was really rooting for them to have their happily ever after.\n\n\n\nDavey just recently moved from Ireland to get away from his family and the drama that comes with them. He loves working at a video game store and is really happy there. He has a steady boyfriend but that doesn't mean that Davey is 100% out of the closet. Jay is his friend from work but Davey doesn't know that Jay is in love with him.\n\n\n\nJay is totally in love with his best friend Davey. He thinks that Davey is perfect but he can tell that Davey is hiding something. Davey has a hard time trusting people and Jay doesn't know why. But Jay can tell that Davey is hiding something so he just decides to wait it out until Davey is ready to tell him. He's hoping that one day, Davey will realize that Jay is in love with him and he will finally be able to be with the man he loves.\n\n\n\nI really liked the characters. I felt really bad for Davey and how his family was treating him. I also really liked Jay. He was such a great guy. He was just really honest about his feelings for Davey. And I loved that about him.\n\n\n\nIt was a really great story. I loved the characters, I loved the storyline, and I was really rooting for these two to be together.]" time="0.026"><properties><property name="score" value="0.023056565060857195" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02305657&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02305657
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[La ministre f\xe9d\xe9rale de la Sant\xe9, Maggie De Block (Open VLD), n'entend pas se mettre \xe0 genoux devant les r\xe9clamations de ceux qui r\xe9clament un d\xe9remboursement du soin m\xe9dical de base. &quot;J'entends les gens, mais je pense qu'il faut se tourner vers d'autres solutions&quot;, explique-t-elle au soir de l'anniversaire de la prise du pouvoir par le nouveau gouvernement de coalition f\xe9d\xe9ral. &quot;Je vais en d\xe9battre avec les partenaires sociaux, dans la soci\xe9t\xe9 civile, avec les \xe9lus r\xe9gionaux et les \xe9lus locaux&quot;, ajoute-t-elle.\n\nLes chiffres sont l\xe0. Selon les calculs de la Drees, le co\xfbt du maintien du syst\xe8me de soin de base en Belgique est de 2,8 milliards d'euros par an. Une somme qui n'est pas forc\xe9ment optimale dans un contexte budg\xe9taire tendu. En 2014, par exemple, des remboursements aussi massifs ont provoqu\xe9 un manque \xe0 gagner de 1,1 milliard d'euros pour le gouvernement f\xe9d\xe9ral. Un d\xe9ficit qui s'ajoute au budget annuel d'environ 21,5 milliards d'euros de la sant\xe9 publique.\n\nLa ministre se montre prudente, parce qu'il ne faut pas se tromper de combat. Il ne s'agit pas de r\xe9former le syst\xe8me de soins de base pour une raison financi\xe8re, mais pour l'am\xe9liorer, pr\xe9cise-t-elle. &quot;Je ne suis pas pour le d\xe9remboursement de la m\xe9decine de base, mais je veux prendre un nouveau d\xe9part, et que tout le monde y trouve son compte&quot;, ajoute-t-elle.\n\nMaggie De Block a d'ailleurs entam\xe9 une s\xe9rie de consultations, avant de rendre son rapport final. A ses yeux, le &quot;temps de l'improvisation est pass\xe9&quot;, car les gens attendent des r\xe9sultats concrets. Le gouvernement lib\xe9ral flamand et le PS, ainsi que des institutions nationales, mais aussi internationales, auront leur mot \xe0 dire dans ce d\xe9bat.]" time="0.030"><properties><property name="score" value="0.18098289524906622" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.1809829&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.1809829
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Hello, I hope I'm posting this in the right area, I'm not sure. But I am in a bit of a predicament. I got my first laptop (Dell Inspiron 15R) for Christmas, and when I put it on my first university assignment it worked fine, for like 3 days. Then all of a sudden it starts turning off by itself. I've tried it on multiple chargers and outlets, still the same. I then found out that the reason why it would do this was because I had the processor set on High Performance, which I've now changed to Balanced, but that's still not fixed the problem. I have had this laptop for 2 weeks, and it's really beginning to make me feel insecure in a laptop.\n\n\n\nDoes anyone know what this could be, and how to fix it? Thanks.]" time="0.018"><properties><property name="score" value="0.07092321544214797" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07092322&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07092322
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The beach\n\nNew Year\u2019s Eve party\n\nNew Year\u2019s Day party\n\nThe beach is the number one attraction of Rio Grande do Sul. Each day we celebrate the weather, going to the beach in the morning and afternoon. During the evenings we celebrate the seasons of the year in winter we sit around the fireplace, while in summer we dance around the bonfire. In summer we dance samba in the streets of Santa Catarina, in Rio Grande do Sul, in the middle of the city, in the suburbs, in the mountains, at the beach and on the hills. At night we can enjoy live music in one of the many bars of Santa Catarina. You can find bars and clubs that play regional music, national music, world music and, in many cases, international music. If you are an extroverted person, there is no better place than Santa Catarina.]" time="0.014"><properties><property name="score" value="0.01259174775447569" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01259175&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01259175
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Details\n\nPictures\n\nClubs Top Characters Jan Jan (\u30b8\u30e3\u30f3) (Saibara Jansai) Is a survivor of the Third World. His hair is red, but he later bleaches it.\n\nHe uses a slingshot, but eventually gains a bandit's gun after trading with Gorobei.\n\n\n\nVoice Actors Hayashi, Hiroko Japanese\n\n\n\n]" time="0.014"><properties><property name="score" value="0.06296243299945825" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06296243&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06296243
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Replies: 17 Royal Viewer 2.1.0.d7902 (with 1.7.0.d7804 Patch) Released Views: 1,625 Posted By Rick Brown Thanks, Satch, for the fast response. ...\n\n\n\nThanks, Satch, for the fast response. I have a question about that icon: Is there an option that can be set to not show it if it's not needed, and only show it when needed? Thanks, Satch, for the fast response. I have a question about that icon: Is there an option that can be set to not show it if it's not needed, and only show it when needed?]" time="0.012"><properties><property name="score" value="0.04268177765698322" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Panda's a great restaurant that does a fantastic menu and even better service. You can't go wrong with anything on the menu. It's all very tasty. The veal is very tasty.\n\nWas this review helpful to you?]" time="0.013"><properties><property name="score" value="0.009307675021230644" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00930768&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00930768
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Top Rank Inc. has signed welterweight contender Rod Salka to an exclusive promotional agreement, Top Rank president Todd duBoef told ESPN.com on Tuesday.\n\nSalka, 29, will fight on the undercard of the Timothy Bradley-Juan Manuel Marquez pay-per-view main event on Oct. 12 at the Thomas &amp; Mack Center in Las Vegas.\n\nSalka, a former featherweight and super featherweight, will face Jorge Pimentel (23-4, 18 KOs), 32, of the Dominican Republic, in a scheduled 10-round welterweight fight. The card will be televised by HBO.\n\n&quot;I've signed with Top Rank and will be fighting in Las Vegas on Oct. 12,&quot; Salka said. &quot;I'm very excited about this opportunity. This is a great chance for me to go against one of the biggest names in boxing. I'm so thankful for the support that I've received from my family, friends and fans.&quot;\n\nSalka, who lives in Pittsburgh, is 23-3 with 13 knockouts and has not fought since July 31, 2011, when he lost a 12-round unanimous decision to Danny Garcia on the undercard of the Amir Khan-Zab Judah card at Mandalay Bay in Las Vegas.]" time="0.013"><properties><property name="score" value="0.13398705674391165" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.13398706&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.13398706
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[From Wikipedia, the free encyclopedia\n\nHarry P. Robinson\n\nHarry Payne Robinson (May 11, 1872 \u2013 May 10, 1955) was an American Democratic politician from Philadelphia.\n\nHe was born on May 11, 1872 in Pennsylvania.\n\nRobinson was a member of the Philadelphia City Council for three terms. He was elected to the 71st United States Congress and the four succeeding Congresses. He did not seek renomination in 1920. He was elected again to the House of Representatives in 1930, this time]" time="0.014"><properties><property name="score" value="0.04617139563191652" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0461714&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0461714
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The Adrian School District is dedicated to providing the best possible learning environment for all students. This school district is composed of approximately 5,000 students in grades K-12 and 125 full-time and part-time teachers.\n\nAdrian is a town of approximately 16,000 people, located in southwest Michigan. Adrian is best known for its variety of community events, from festivals and parades to musical concerts. A new football stadium opened in 2015, and other athletic facilities are planned for the future.\n\nSince 1873, the Adrian School District has been dedicated to providing students with the best education possible.\n\nAdrian Schools is one of the three districts within the Jackson-Adrian Area Schools, which includes Jackson and Adrian. This collaboration allows our students to take advantage of an excellent education and opportunities for athletic competition in a small, personal community.\n\nWe are very proud of our students and staff and are very excited about the future of our district.\n\n- Chris Hettinger\n\nSuperintendent]" time="0.014"><properties><property name="score" value="0.004383286409315179" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00438329&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00438329
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[fot. J\u0119drzej Skiba\n\nWizualna uczta dla wielbicieli Harry\u2019ego Pottera. Ju\u017c niebawem w Olsztynie!\n\nZdj\u0119cia z filmu zrealizowane przez Alfonso Cuar\xf3na, do kt\xf3rego graj\u0105 gwiazdy jak Daniel Radcliffe czy Emma Watson, zapowiadaj\u0105 si\u0119 niezwykle okaza\u0142e. Zaanga\u017cowanie mieszka\u0144c\xf3w, \u015bwiat\u0142o, charakterystyczne elementy miasta b\u0119d\u0105 w tym filmie cz\u0119\u015bci\u0105 wizualnej uczty dla ka\u017cdego fanu Harry\u2019ego Pottera.\n\nPlanowana dat\u0105 startu projektu jest pa\u017adziernik 2019 roku. Projekt b\u0119dzie realizowany przez Filmowy Ob\xf3z Gier, Wsp\xf3\u0142pracuj\u0105cych ze Stacj\u0105 Filmoznawcz\u0105 im. Andrzeja Wajdy i Stowarzyszenie Pomocy Inicjatywom Kulturalnym Miasta Olsztyna.\n\n- Miejscem realizacji filmu b\u0119dzie miasto Olsztyn, zar\xf3wno star\xf3wka, jak i inne cz\u0119\u015bci miasta, a prawdopodobnie tak\u017ce same Podlasie. Trudno jednoznacznie wskaza\u0107 przeznaczone obszary, gdy\u017c pozyskamy najlepsze miejsca do filmowania. W pobli\u017cu rzek, na szlakach turystycznych, w zabytkowych okoliczno\u015bciach, w obiektach, kt\xf3re by\u0142yby dost\u0119pne dla produkcji, przy budynkach kulturalnych, w wojskowych obiektach - m\xf3wi Mariusz \u0141azicki, prezes Filmowego Obozu Gier.\n\nTymczasem pocz\u0105tek pracy tw\xf3rc\xf3w zaplanowany jest na 21 wrze\u015bnia w Centrum Kultury w Olsztynie. Wej\u015bcie na teren projektu jest bezp\u0142atne, a pierwsze pr\xf3by rozpocznie si\u0119 w pa\u017adzierniku, r\xf3wnolegle z realizacj\u0105 obiektu.\n\nPodczas wyprawy filmowc\xf3w po Europie, prac\u0119 tw\xf3rc\xf3w poprowadzi ameryka\u0144ski re\u017cyser.\n\nW\u015br\xf3d uczestnik\xf3w b\u0119dzie mia\u0142a r\xf3wnie\u017c swoje miejsce Olszty\u0144ska Fabryka Form Filmowych - eksperymentalna scenografia - najwi\u0119kszy tego typu obiekt w Europie.\n\n- Ten nowy obszar pracy filmowej przejdzie gruntown\u0105 renowacj\u0119, w tym nowa przepustowo\u015b\u0107 dla t\u0142umu widz\xf3w. Olsztyn i teren przy Fabryce Form Filmowych wyr\xf3\u017cniaj\u0105 si\u0119 na tle innych miast w Polsce, ale i w Europie. Dzisiaj ju\u017c niebawem mo\u017cemy wystawi\u0107 szyld w kraju i na \u015bwiecie - m\xf3wi \u0141azicki.\n\nOd 2008 roku, czyli od czasu pojawienia si\u0119 pierwszych prac re\u017cyserskich w tym miejscu, powstaje kilkana\u015bcie prac, dzi\u0119ki czemu ma miejsce w Europie, wykorzystywane na potrzeby filmowe i komercyjne.\n\nObecnie w przysz\u0142ym roku zaplanowana jest realizacja kolejnego filmu o losach czarodziejki, gdzie Olsztyn b\u0119dzie g\u0142\xf3wnym miejscem akcji.\n\n- Jedna z produkcji ma za zadanie pokaza\u0107 mieszka\u0144com miasta, a za po\u015brednictwem filmu z ca\u0142ego \u015bwiata, jak Olsztyn zosta\u0142 wybrany do filmowania. W przysz\u0142o\u015bci obiekt wykorzystany jest do realizacji dalszych produkcji z unikalnym potencja\u0142em. Takie miejsce ma szans\u0119 powsta\u0107 tylko raz na kontynencie - podsumowuje prezes.\n\nW przysz\u0142o\u015bci, w obiekcie b\u0119d\u0105 realizowane kolejne filmy, nie tylko te komercyjne.\n\n- Konsekwencj\u0105 pojawienia si\u0119 Fabryki Form Filmowych, jest potrzeba dalszej rozbudowy obiektu. Tw\xf3rcy nie zapomn\u0105 o Olsztynie. Chc\u0105, aby fabryka sta\u0142a si\u0119 pr\u0119\u017cnie dzia\u0142aj\u0105c\u0105 wydarzeniem kulturalnym i edukacyjnym, daj\u0105cym prac\u0119 tysi\u0105com ludzi w Polsce. Jak i te\u017c nie zapominaj\u0105, \u017ce fabryka powsta\u0142a dzi\u0119ki Polakom, kt\xf3rzy dzia\u0142ali w tym miejscu pod nazw\u0105 Fabryka Filmowa &quot;Kolor&quot; - m\xf3wi prezes.]" time="0.028"><properties><property name="score" value="0.31186982087162707" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.31186982&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.31186982
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Share Tweet Email Copy Link Copied\n\nTimeless creator Eric Kripke wants a season 4 to resolve the show's cliffhanger ending.\n\nThe time-travel series, starring Abigail Spencer, Matt Lanter, Malcolm Barrett and Goran Visnjic, was canceled after two seasons on NBC. However, after a last-minute bid from Sony Pictures Television, it was ordered to a special two-part finale. The shows were meant to resolve the cliffhanger, and answer why Rufus was shot and where Lucy and Wyatt ended up. However, when the episode aired, nothing was resolved.\n\nRelated: Timeless Fits Perfectly Into The Revival TV Movement\n\nIn an interview with Syfy Wire, Kripke revealed he's hoping for a season 4 so that the show can get a proper ending.\n\n&quot;We always conceived of [the finale] as a potential series finale, because we couldn\u2019t get canceled! So that\u2019s just what it ended up being. It\u2019s like, \u201cOK, we can\u2019t end the show, so let\u2019s end the arc.\u201d It felt really appropriate and full circle, and everybody at NBC, Sony and Warner Bros. was like, \u201cThat\u2019s a perfect place to end this.\u201d We\u2019re not ruling out a future for the show. Who knows? We still feel passionately about it. Maybe there will be a season 4. I think everyone at the studio and the network felt really good about where it ended.\u201d\n\nIf]" time="0.013"><properties><property name="score" value="0.13509322066243" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.13509322&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.13509322
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Arnold R\xfc\xfctel\n\nArnold R\xfc\xfctel (born 3 April 1937) is an Estonian politician who was the third President of Estonia from 2001 to 2006. He was born in Rakvere, L\xe4\xe4ne-Viru County. He served as the Mayor of Tallinn from 1992 to 2001 and was the Speaker of the National Assembly from 1999 to 2001.\n\nR\xfc\xfctel served as a Colonel in the Soviet Air Defence Forces from 1967 to 1992. He joined the pro-independence party &quot;Vaba Eesti&quot; (Free Estonia) in 1988.\n\nIn the last Soviet-era election of the Supreme Soviet of the Estonian SSR, held on 17 March 1990, R\xfc\xfctel was elected to the parliament, representing the Tallinn constituency. On 20 August 1991, he was one of the speakers at the Congress of Estonia, which proclaimed independence. He was elected to the new Riigikogu (Parliament) on 18 September 1991, and became Minister of Foreign Affairs in the government of Mart Laar.\n\nR\xfc\xfctel joined the conservative Pro Patria Union party in 1993 and was elected chairman in 1996. He ran for president in 1999 and came in second to Arnold R\xfc\xfctel. He became a member of the Riigikogu again in 2003, this time for the Reform Party. R\xfc\xfctel is considered to be a pro-Western politician and a leading proponent of Estonia's membership of NATO. In February 2004, he was the first person to publicly endorse IRL politician Ene Ergma as the Reform Party's candidate for president in the scheduled elections. On 12 September 2004, he announced that he was joining the Union of Pro Patria and Res Publica party, whose candidate for president he had been in 1999.\n\nOn 23 September 2004, the Riigikogu elected R\xfc\xfctel to the office of President of Estonia. R\xfc\xfctel was sworn into office on 9 October 2004, and was re-elected in 2006.\n\nR\xfc\xfctel has held a number of positions in the Estonian government since 1991, including as Minister of Foreign Affairs, Minister of Education and Research, Speaker of the Riigikogu, and a two-year term as the Mayor of Tallinn (from 1992 to 1994).\n\nHe has also served as the Chairman of the National Defense Committee and was the leader of the European Movement in Estonia from 1996 to 2004.\n\nOn 20 February 2007, R\xfc\xfctel announced his intention not to run for re-election in the scheduled presidential election of 2007.\n\nHis first wife Tiina Urm (1942\u20132011) died on 9 May 2011, aged 69, from undisclosed causes. R\xfc\xfctel has one son, Peeter, born in 1968, from his first marriage.\n\nHis second wife is Kadri M\xe4lk, with whom he has a daughter, Triin, born in 1989.\n\n]" time="0.023"><properties><property name="score" value="0.09783608147406417" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.09783608&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.09783608
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Crescent-chested whitestart\n\nThe crescent-chested whitestart (&quot;Myioborus luteoventris&quot;) is a species of bird in the Parulidae family. It is found in the humid lowland forests of tropical South America. Its range extends from southeastern Colombia, eastern Ecuador, and northeastern Peru to northern Bolivia, southeastern Venezuela, and northern Brazil. It is a fairly common species within its small range, and the IUCN has rated it as a species of &quot;least concern&quot;.\n\nThe crescent-chested whitestart is a medium-sized songbird. The adult has a fairly long tail and a thin bill. It is mainly yellow-olive, with a white belly and vent, a yellow breast, and a yellow eyebrow. It has a black stripe over the eye and a whitish crescent on the breast. It is similar to the yellow-bellied white-start and the turquoise-fronted whitestart.\n\nThe crescent-chested whitestart is native to eastern South America, its range extending from southeastern Colombia, eastern Ecuador, and northeastern Peru to northern Bolivia, southeastern Venezuela, and northern Brazil. It inhabits humid tropical forest between sea level and about above sea level. It can be found at altitudes of , but generally below .\n\nThe crescent-chested whitestart is a fairly common species. No particular threats have been identified, and the International Union for Conservation of Nature has rated its conservation status as being of &quot;least concern&quot;.\n]" time="0.014"><properties><property name="score" value="0.05458754227638529" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Alejandro Leguizam\xf3n\n\nAlejandro Leguizam\xf3n (born October 30, 1961 in Santa Fe) is an Argentine journalist. He is a columnist of &quot;Noticias&quot; and &quot;P\xe1gina/12&quot;. Leguizam\xf3n also is a member of &quot;The Clinic&quot;, a leading group of Argentine intellectuals, and a writer of a weekly column in &quot;P\xe1gina/12&quot;.\n\nAlejandro Leguizam\xf3n is the son of Eduardo Leguizam\xf3n, founder of the &quot;C\xedrculo Militar&quot;, and brother of the TV presenter Mariana Leguizam\xf3n.\n\n]" time="0.012"><properties><property name="score" value="0.040226083644084865" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04022608&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04022608
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Friedrich Hofmeister\n\nFriedrich Hofmeister was a German piano and music publisher in Leipzig. His company, Friedrich Hofmeister Musikverlag, is the oldest music publishing company in Germany and one of the oldest in the world. It was founded in 1798. His catalogue contained more than 25,000 works, many published as composers' copies, without printed music.\n\nHofmeister was a friend and patron of Robert Schumann, a composer whose music was frequently published by Hofmeister, and a generous supporter of Felix Mendelssohn Bartholdy. A close friendship developed between Mendelssohn and Schumann, who lived at the same time.\n\nOne of the most important works of Schumann, his Piano Quintet in E-flat major, op. 44, is dedicated to Friedrich Hofmeister, as is Schumann's Piano Concerto in A minor, op. 54.\n\nIn his will, Schumann left his house to Mendelssohn.\n\nIn 2011, a new edition of the full catalogue was launched. The company today is a subsidiary of the German publisher B\xe4renreiter, founded in 1525.\n]" time="0.014"><properties><property name="score" value="0.04510148535546987" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04510149&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04510149
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Manigramam\n\nManigramam is a village in the Orathanadu taluk of Thanjavur district, Tamil Nadu, India.\n\nAs per the 2001 census, Manigramam had a total population of 2169 with 1068 males and 1101 females. The sex ratio was 994. The literacy rate was 74.41.\n]" time="0.012"><properties><property name="score" value="0.048502757883039926" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04850276&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04850276
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Claudia Escobar\n\nClaudia Esther Escobar Ram\xedrez (born 9 April 1995) is a Mexican professional football Midfielder who currently plays for Puebla in the Liga MX Femenil.\n\nEscobar debuted in the 2013\u201314 season in a match against Pachuca.\n\nIn July 2015, Escobar was transferred from C.D. Guadalajara to Puebla F.C..\n\nShe won the first Women's Football Mexican Championship in 2017, with Puebla F.C..\n]" time="0.013"><properties><property name="score" value="0.07030083485505823" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07030083&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.07030083
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Bucculatrix pomifoliella\n\nBucculatrix pomifoliella is a moth in the family Bucculatricidae. It is found in North America, where it has been recorded from Maryland, Ohio, Oklahoma, Texas and Maine.\n\nThe wingspan is 8\u201310\xa0mm. The forewings are white with pale brownish-grey lines. Adults have been recorded on wing in August.\n\n]" time="0.012"><properties><property name="score" value="0.0435448770183422" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04354488&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04354488
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Transport and Communications Workers Union\n\nThe Transport and Communications Workers Union (T&amp;CWU) is a trade union in Trinidad and Tobago, with the bulk of its membership employed by the state-owned National Petroleum Company (Petrotrin).\n\nThe union was founded in 1952 as the Oilfield Workers Trade Union, and in 1974 it became the Petroleum Workers Trade Union. In 1977, the union renamed itself as the National Oilfield Workers Trade Union, and then in 1987, the name was changed again to the National Oilfield Workers Union.\n\nThe union changed its name again in 1997 to its current form. In 2005, it merged with the Civil Service and Allied Workers Union, forming the Petroleum and Allied Workers Trade Union. The union re-split in 2007.\n\nFor many years, the union was led by Ancel Roget, and in 1997 he was convicted for soliciting a $40,000 bribe from an executive at the Shell Oil Company. In 2000, Roget was forced to retire from his position at Petrotrin due to a management investigation.\n\nIn 2006, the union was involved in a strike over the state of Petrotrin. In 2010, the union was accused of illegally striking.\n\nIn 2008, a branch of the T&amp;CWU was reported to have hired gang members to threaten striking Petrotrin workers.\n\n\n]" time="0.012"><properties><property name="score" value="0.0007963832759749767" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Raymond T. McNally\n\nRaymond Thomas McNally (March 23, 1917 \u2013 March 15, 2005) was an American author, best known for his book, &quot;The Dracula Scrapbook&quot;.\n\nHe was born in Connecticut and was of Irish descent. He began work at the RKO film studios at the age of 19 as a film librarian. During the 1940s he was a freelance writer and served in the United States Army.\n\nMcNally was best known for his &quot;Dracula Scrapbook&quot;, a work which consisted of a compilation of newspaper reports, letters, affidavits, and legal depositions detailing the various efforts to suppress Bram Stoker's novel &quot;Dracula&quot;. He also wrote books on the films of Marilyn Monroe]" time="0.013"><properties><property name="score" value="0.06371043287601852" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06371043&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06371043
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[A/N: Before I start, I must confess. I lied. In an interview with Josh Radnor from earlier this year, he claimed he would only do a spin-off if it was centered around Ted and Barney, with the two of them going off to some far off place (New Zealand? Australia?) where they lived in a beach house and watched sports. I laughed. A lot. And now that I have come up with a sequel for The Only Person Who Didn't Run from Ted Mosby, I don't really have a plot for it...\n\n(Disclaimer: This is a work of fan fiction. Any resemblance to any other work of fiction is entirely coincidental. Not to mention a little insulting)\n\nI sat on my bed, looking down at my phone. There was only one person who could help me write this...\n\nI knew it was kind of a lame excuse. I mean, my sister works at a museum, for Pete's sake. She can't help me with a stupid paper. What was I thinking? And I had a really good excuse to tell my girlfriend when I get home. We had a date. But, in order to give me that good excuse, I have to write this stupid paper and get it done. I have to talk to my sister.\n\nI dialed the number to my sister's work, and it started ringing. My sister, Gloria, picked up.\n\n&quot;What's up, D?&quot;\n\n&quot;Glor, do you think you can help me with a paper? I really need some help.&quot;\n\n&quot;Sure. What is it about?&quot;\n\n&quot;It's a paper about JFK.&quot;\n\n&quot;Oh. Cool. Yeah, sure. What do you need?&quot;\n\n&quot;I need some help finding some good sources for this. I need to know how to start a paper.&quot;\n\n&quot;I'm glad you asked.&quot;\n\nAnd I was off to a good start. I sat at my desk, making notes. The paper was going to be really easy to write. It would probably only take a couple hours. I knew it. And I would be home before Lily.\n\nThe paper was starting to look really good. I had a bibliography in place, a couple of good paragraphs, and a nice introduction. Everything was great.\n\nI had a paper ready to turn in the next morning. It was almost done, and I was free to go home. But I had a date to get to.\n\n&quot;Bartender!&quot; I called. &quot;One more, please!&quot;\n\n&quot;You sure, man? You're not gonna make it home tonight.&quot;\n\n&quot;Nah, man. It's all good. I'll just be getting home, like, ten minutes later than I would.&quot;\n\n&quot;Okay, man. Whatever you say.&quot;\n\nI started walking out the door, but I took the bar stool with me. I had to carry it around to make it look like I was in a drunken stupor. If my girlfriend saw that I was drunk when I got home, I would never be able to keep my secret.\n\nAs I made my way through the train station, I heard a strange beeping sound. I turned my head to the side, trying to find the source of the noise. It was coming from my phone. There was a text message. I opened my phone, and read it.\n\nI was starting to feel a little tipsy. It was a good thing I was carrying a bar stool around. It made me look like I had a reason for not walking right.\n\nThe message was from Lily. I read it.\n\nThe good thing about living in New York was the night life. I was lucky enough to be able to spend an evening at a nightclub, so I would be home in no time.\n\n&quot;Yeah, man,&quot; I said. &quot;I don't know why I even did this, dude.&quot;\n\nI walked out of the nightclub and back into the train station. I knew it was late, but the train wasn't too crowded. I sat down in a seat, trying to get some sleep. I would be home soon enough.\n\nI woke up as the train pulled into the station. I was close to home now. It was getting a little crowded on the train, but I managed to squeeze myself into the seat. I was a little more aware of the fact that I was still carrying my bar stool, but I decided to go on anyway.\n\nI stumbled out of the train station, still carrying the bar stool. I made my way through the crowd, out into the parking lot, and across the street to my car. I put the bar stool in the trunk, and got in the car. I turned on the engine, and drove off.\n\n&quot;Hey, Lily. Yeah, I'm on my way home. I'm sorry, but I ran into some trouble. It shouldn't take me long to get home.&quot;\n\n&quot;You said that you were just going to be a few minutes late. You said you were on your way home, and then you don't show up for three hours. You call me and say you're sorry? That doesn't even make sense!&quot;\n\n&quot;I know, I know. It's just that it took me a little longer to get home than I thought.&quot;\n\n&quot;You lied to me! You said you were going to be home at 8:30! I waited until midnight! You were just trying to get out of our date!&quot;\n\n&quot;It wasn't like that, Lily. I had some trouble with a paper, and then I was at a nightclub. I had a lot of stuff on my mind.&quot;\n\n&quot;So, you didn't want to come home. You just didn't want to spend time with me. And now you're calling me after it's all over and trying to explain yourself. And I don't want to hear it! Goodbye!&quot;\n\nI was standing at the door to my apartment, keys in my hand, when I heard my phone ring. It was Lily. I stood there for a moment, trying to decide whether or not I should answer. I looked down at the floor, and then back at my phone. I answered.\n\n&quot;Lily, I-&quot;\n\n&quot;I don't want to hear it, Ted. I told you that I don't want to hear it.&quot;\n\n&quot;No, no, it's just that I really have to tell you something. I can't wait until tomorrow. I really need to tell you right now.&quot;\n\n&quot;Well, fine. What do you need to tell me? Spit it out.&quot;\n\n&quot;It's just that...I think that...&quot;\n\n&quot;Ted, you're babbling. What is it?&quot;\n\n&quot;I...I...well...&quot;\n\n&quot;Ted? Ted! You're breaking up. I think the connection is getting worse. What was it?&quot;\n\n&quot;I think we should break up.&quot;\n\n&quot;What?&quot;\n\n&quot;I think we should break up.&quot;\n\n&quot;What?&quot;\n\n&quot;I...I...you were right.&quot;\n\n&quot;What?&quot;\n\n&quot;I didn't want to spend the night with you. I didn't want to be around you anymore. I wanted to get out of the date. I know it's not fair. I know I was a jerk. I know I lied to you. I just...I don't think we should be together anymore.&quot;\n\n&quot;What? Why?&quot;\n\n&quot;I don't know. I just don't.&quot;\n\n&quot;Ted, I love you. You can't just say that we should break up like this.&quot;\n\n&quot;You were right. You shouldn't have to date me.&quot;\n\n&quot;Ted...&quot;\n\n&quot;I'm sorry. I really am. I didn't mean to hurt you. You're a wonderful girl. I just...I don't know. I'm sorry.&quot;\n\n&quot;Ted...&quot;\n\n&quot;Goodbye, Lily.&quot;\n\nI hung up the phone, and dropped it to the floor. I was standing there for a moment, trying to get my emotions in check.\n\nI had done it. I broke up with my girlfriend of three years. I had broken up with the girl I loved. I had no idea why I did it, but I was doing it.\n\nI opened my door, and made my way to my bedroom. I dropped my keys on the dresser, and sat down on my bed. I leaned against the wall, and closed my eyes.\n\n&quot;Hey, Ted. What are you doing home so early?&quot;\n\n&quot;Lily, I...&quot;\n\n&quot;Ted, I know I was mad at you before, but I've calmed down a lot. You know I love you. I can't live without you.&quot;\n\n&quot;Lily, I...&quot;\n\n&quot;Ted, you just scared me, is all. I thought I lost you. I thought you didn't want to be with me anymore.&quot;\n\n&quot;Lily, I...&quot;\n\n&quot;I know it was wrong for me to keep pushing you to make up your mind about me, but you couldn't see that.&quot;\n\n&quot;Lily...&quot;\n\n&quot;And I know that I can be a little difficult to get along with sometimes, but I know you'll come around.&quot;\n\n&quot;Lily, I...&quot;\n\n&quot;We're supposed to be together, Ted. We always knew that. You and me, we're meant to be. You can't just do something like this, Ted.&quot;\n\n&quot;Lily...&quot;\n\n&quot;I love you, Ted. You can't just throw that away. I can't just throw that away. I can't...&quot;\n\nI opened my eyes, and saw Lily]" time="0.073"><properties><property name="score" value="0.06518460002957543" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0651846&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0651846
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Former Arizona Sen. John McCain died after a yearlong battle with brain cancer.\n\nMy fellow Republicans, John McCain was a true American hero. https://t.co/YlNEJ2KRAS \u2014 Jake Tapper (@jaketapper) August 26, 2018\n\nArizona Sen. John McCain, the maverick Republican whose war record and outspokenness against abuses of power made him an American icon, died on Saturday. He was 81.In a statement, McCain\u2019s office said he died at 4:28 pm local time at the family\u2019s home in Arizona surrounded by family members. The cause of death was not immediately known.The former Navy pilot, who was imprisoned and tortured during the Vietnam War, rose to the top of American politics after emerging as a prominent critic of President Donald Trump.An upstart presidential bid in 2000 did not succeed but after he won the Republican nomination in 2008, he lost to Democrat Barack Obama.Despite the loss, his image soared and he emerged as one of the most influential members of the Senate, willing to work with Democrats on issues like campaign finance reform.By the time he ran for president again in 2016, he seemed like a throwback to a bygone era of civility and bipartisanship. McCain ran against Obama again in 2008.He continued to play a pivotal role in American politics, becoming chairman of the powerful Senate Armed Services Committee after the Republicans regained control of the Senate in 2015.McCain in the past had been a thorn in the side of Republican presidents \u2014 George W. Bush and Trump.He was also well known for reaching across the aisle.He took a secret trip to visit Syrian dictator Bashar al-Assad during the early years of the country\u2019s civil war, and was one of the final Republicans to speak with the president before his firing last year.He was a long-time critic of the U.S. use of torture against detainees.He held out against Republicans\u2019 efforts to dismantle Obama\u2019s signature healthcare legislation and refused to support the Graham-Cassidy bill last year.He was a harsh critic of Russia, and one of the earliest supporters of sanctions against Russian officials. He played a role in seeking to keep the United States from launching a pre-emptive attack against North Korea in the face of its nuclear provocations.McCain was one of the first Republicans to support Trump\u2019s presidential campaign in 2016. But when the \u201cAccess Hollywood\u201d tape emerged, he withdrew his support, and famously said in his 2017 memoir \u201cThe Restless Wave,\u201d \u201cI do not know how I could have spoken so disrespectfully about the commander in chief of the Armed Forces.\u201dHe also repeatedly denounced what he called Trump\u2019s \u201ccheap rhetoric and insults\u201d against U.S. allies, and after Trump\u2019s summit with Russian President Vladimir Putin in July 2018, called it a \u201ctragic mistake.\u201d He was one of the loudest Republican voices criticizing the president for his performance at a summit with Putin in Helsinki.Trump mocked him as \u201ca clown,\u201d and even at the end of his life, he was still a target for the president, who reportedly resisted hosting a state funeral.On Saturday, Trump said he sent his \u201cdeepest sympathies and respect\u201d to McCain\u2019s family.\u201cThe nation mourns the loss of a great American patriot, a statesman who put his country first and enriched this nation every day he served.\u201dFormer President Barack Obama called him a \u201cgiant of American politics,\u201d and his one-time opponent in the 2008 presidential race said he was \u201ca man of deep conviction and a patriot of the highest order.\u201dPresidential candidate Sen. Bernie Sanders, I-Vermont, called McCain \u201can American hero\u201d on Twitter.McCain\u2019s funeral service will be held at the Washington National Cathedral and the ceremony will include eulogies from former President George W. Bush, former President Barack Obama, former Secretary of State Henry Kissinger and former Sen. Joe Lieberman, the former vice-presidential candidate who was a close friend.After the service, he will be buried in Annapolis, Md., at the U.S. Naval Academy.McCain was born at a naval station in the Panama Canal Zone, where his father was stationed. After his father\u2019s career in the Navy, the family settled in the Arizona desert, where he attended high school.McCain studied at the Naval Academy and became a pilot in the Navy, serving on a ship that was shot down in Vietnam. He spent five-and-a-half years in captivity, which included torture, before he was released in 1973.McCain returned to the U.S. and was assigned to the Senate as a liaison to the Navy. He became a member of the House of Representatives in 1983 and eventually the chairman of the Senate Committee on Commerce.During his 2008 presidential bid, he secured the GOP nomination in the face of much skepticism. At a dinner in 2008, then-Sen. Barack Obama joked that after losing to Obama, McCain would be on a beach sipping a drink with \u201can umbrella in it.\u201dThe Republican-led Senate, which was much closer to McCain than to the president, overwhelmingly passed a resolution earlier this year that censured Trump for his \u201cbigoted and racist\u201d remarks about four members of Congress.In his memoir, McCain said that he regretted that his choice of Sarah Palin as a running mate did not propel him to the White House. He said he made the choice based on who he thought could become president, but that was wrong.\n\nJoin Jerusalem Post Premium Plus now for just $5 and upgrade your experience with an ads-free website and exclusive content. Click here&gt;&gt;\n\n\n\n]" time="0.076"><properties><property name="score" value="0.0946661645550535" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.09466616&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.09466616
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Avolites V-Color 2\n\nV-Color is an all-in-one visualisation system for lighting professionals which empowers operators with unprecedented control and flexibility, while also offering total ease-of-use. A complete visualisation solution for the modern production environment.\n\n\n\nNew V-Color 2 Software\n\n\n\nUsing the same hardware, V-Color 2 offers full compatibility with existing shows, but with brand new software including some great new features and exciting innovations.\n\n\n\nMore Detailed, Better Feedback\n\nThe addition of 3D colour bars gives a better representation of colour than before. The inclusion of warm/cool &amp; complementary information makes diagnosing colour temperature easier than ever. An inbuilt screen lock function and scene copy capability further streamlines the user experience.\n\n\n\nEasier Control of Everything\n\nWith a large, backlit LCD screen, V-Color 2 gives the operator total control of the system from a distance. The touchscreen interface makes everything intuitive and gives instant access to show information and control options.\n\n\n\nSeamless Integration into Live Production Workflow\n\nPowered by ProfilerX (firmware version 2.3 and above), V-Color 2 gives the operator seamless and continuous control over their show using the same industry standard protocol as their other devices. This means that they can operate their visualisation system in parallel with their lighting console, creating the ideal platform for shows involving complex automated and DMX control.\n\n\n\nBigger, Better, More Powerful\n\nThe increased size of V-Color 2 gives it the space to fit a larger screen with a better resolution. The dual menu and page system makes navigation easier than ever and the addition of a USB port and SD card slot allows the operator to save and play back their show in a simple and fast manner.\n\n\n\nFeatures:\n\n\u2022 Upgradeable for full ProfilerX support.\n\n\u2022 Highly responsive and intuitive touchscreen interface.\n\n\u2022 Fits into any lighting desk.\n\n\u2022 Dual menu system.\n\n\u2022 USB &amp; SD card slot for show playback.\n\n\u2022 Upgraded 3D colour bars.\n\n\u2022 Larger display for better visibility.\n\n\u2022 ProfilerX compatible.\n\n\u2022 Industry standard OSD protocol.\n\n\u2022 Relay channel output for all screens.\n\n\u2022 Separate input and output cables.\n\n\u2022 Quick calibration system.\n\n\u2022 Assignable RS232 port.\n\n\n\nSpecifications:\n\n\u2022 VGA &amp; HDMI Input.\n\n\u2022 TFT screen.\n\n\u2022 WVGA screen resolution.\n\n\u2022 USB &amp; SD card slot.\n\n\u2022 Can be fitted to V-Pad stand.\n\n\u2022 Size: 10.1&quot; x 5.9&quot; x 1.7&quot;.\n\n\u2022 Weight: 1.2kg.]" time="0.031"><properties><property name="score" value="0.058428087567701695" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05842809&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05842809
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Video\n\nIt was the final episode of the ITV sitcom series Night and Day.\n\nSet in a TV station, the programme starred Barbara Flynn and William Gaunt as the heads of rival departments - the news and the features.\n\nAs BBC reporter Paul Collard watched the final episode in 1989, he wondered what might have happened to some of the characters.\n\nProducer: Bill Powell]" time="0.014"><properties><property name="score" value="0.04602549184041793" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04602549&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04602549
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Jason Kreis became a bit of a legend at the University of Pennsylvania. Not only did he make two trips to the College Cup, leading the Quakers to their first ever national championship in 1998, but he also played a critical role in the Penn men's soccer team's decision to make the historic transition from the NCAA Division I to Division III. While he was at the helm of the Red and Blue, he produced such a respectable record (70-32-16) that Penn was actually invited to compete in the Northeast Conference, a Division I conference at the time. Although he is now best remembered for that phenomenal campaign in 1998, he remains a very active member of the Penn community.\n\n\n\n&quot;It's pretty interesting that my last name is 'Kreis' because in Germany, that means 'square' or 'squared,' but]" time="0.013"><properties><property name="score" value="0.06322681414705489" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06322681&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06322681
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Power Consumption\n\nIdle: Windows Vista Aero sitting at the desktop (1280x1024 32-bit) all windows closed, drivers installed. Card left to warm up in idle mode until power draw was stable.\n\nMulti-monitor: Two monitors connected to the tested card, both using different display timings. Windows Vista Aero sitting at the desktop (1280x1024 32-bit) all windows closed, drivers installed. Card left to warm up in idle mode until power draw was stable.\n\nAverage: Crysis 2 at 1920x1200, Extreme profile, representing a typical gaming power draw. Average of all readings (12 per second) while the benchmark was rendering (no title/loading screen).\n\nPeak: Crysis 2 at 1920x1200, Extreme profile, representing a typical gaming power draw. Highest single reading during the test.\n\nMaximum: Furmark Stability Test at 1280x1024, 0xAA. This results in a very high non-game power-consumption that can typically be reached only with stress-testing applications. Card left running stress-test until power draw converged to a stable value.\n\nBlu-ray Playback: Power DVD 9 Ultra was used at a resolution of 1920x1200 to play back the Batman: The Dark Knight disc with GPU acceleration turned on. Playback started around timecode 1:19, which has the highest data rates on the BD with up to 40 Mb/s. Playback was left running until power draw converged to a stable value.\n\nCooling modern video cards is becoming more and more difficult, especially when users are asking for quiet cooling solutions. That's why the engineers are now paying much more attention to power consumption of new video-card designs.For this test, we measured the power consumption of the graphics card only via the PCI-Express power connector(s) and PCI-Express bus slot. A Keithley Integra 2700 digital multimeter with 6.5-digit resolution was used for all measurements. Again, the values here reflect only the power consumption of the card measured at DC VGA card inputs, not of the whole system.We choseas a standard test representing typical 3D gaming usage because it offers the following: very high power draw; high repeatability; is a current game that is supported on all cards because of its DirectX 9 roots; drivers are actively tested and optimized for it; supports all multi-GPU configurations; test runs in a relatively short time and renders a non-static scene with variable complexity.To measure the power consumption of our testbed we used a digital power meter (Keithley Integra 2700) connected before the power supply. The power consumption measurement of the system is based on the power consumption of the drives, not of the graphics cards. All cards were tested with a total of four SSDs: one in the system, one used as a primary test drive in a slot without a powered connector, and two additional drives in the system (not used for testing). The SSDs are Western Digital WD360GD (362GB), WD1500HLHX (150GB), Samsung SSD 830 (128GB) and OCZ Agility 3 (240GB).Idle power consumption of the whole test system (at the desktop) is measured after sitting idle at the desktop for 30 minutes. The cards are not closed. The result is presented as the average value of all four cards.Please note that after extensive testing, we found that measuring idle power consumption only, as seen in many other reviews, can be misleading. Due to different testing methodology and environments, this number can be significantly lower than the real power consumption. Also, in many cases, the card does not reach the desired idle temperature, so it will draw more power. In our test setup, the cards reach their idle temperature very quickly (in a few minutes) and in a climate controlled room (22\xb0C).Therefore, we use this only as a rough indication of the cards' power consumption.Please note we are using different adjustements to test the performance of our graphics cards. Our testing methodology is detailed here . It's worth mentioning that our method allows us to perform a more thorough test that's free from GPU bottlenecks. It's also free from variations caused by dual-gpu/hybrid configurations.The following two values are the peak power consumption and the maximum power consumption of our testbed (without speakers/headphones) measured after long periods of use (at least 15 minutes) with Windows Aero enabled. The peak power consumption is the maximum power draw of the graphics card only, measured in the most stressful situation (generally, when it's overclocked). The maximum power consumption is the maximum power draw of the complete testbed, including the monitor, speakers, hard-disk, SSD, etc.The results are impressive. Idle power consumption is a tiny bit higher than that of the GTX 580. All cards operate at fairly low power levels in idle. NVIDIA has done a good job reducing idle power consumption of its GTX 680. Under full load, the GTX 680 is only slightly more power hungry than the GTX 580. Compared to AMD, the GTX 680 fares even better as it uses slightly less power than the HD 7970.All graphics cards tested consume quite a lot of power during Blu-ray playback. Interestingly, this is one area where NVIDIA can't keep up with AMD.]" time="0.094"><properties><property name="score" value="0.0171307456105443" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01713075&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01713075
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[This is an experiment in one-shot plays. I've been inspired by the weekly news and what it tells us about the human condition. I call this &quot;Instructional Theatre&quot; because these plays are intended to instruct the audience in a point of view. And they are performed in one sitting. Here is the first one, 'If I Only Had A Wife' by Henrik Ibsen.]" time="0.016"><properties><property name="score" value="0.02527519558497282" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0252752&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0252752
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[A bi-partisan bill that would give states the option to spend welfare funding on faith-based charities is making its way through the U.S. Senate.\n\nThe legislation is called the Combating and Preventing Crimes of Violence Act, and it would give states the ability to use federal funds on programs that promote &quot;the traditional values of marriage, family, religion, and morality.&quot; The bill would also authorize $5 million annually to states to promote chastity and healthy marriage.\n\nState governments would be permitted to use federal welfare funding to teach young people &quot;the social, psychological, and health gains to be realized by abstaining from sexual activity,&quot; according to the text of the bill, which was introduced by Sen. Orrin Hatch (R-Utah) and 23 other co-sponsors.\n\nSen. Hatch said the legislation would reinforce &quot;state efforts to strengthen families and communities by increasing opportunities for the social and economic advancement of young people.&quot;\n\nSen. Hatch's home state of Utah passed a similar law in 2013, and it was successfully implemented, according to a statement from the senator's office.\n\n&quot;This is not an effort to allow taxpayer funding for religious groups, but to give states the option to partner with groups to promote healthy relationships,&quot; the senator's statement said.\n\nHowever, the American Civil Liberties Union (ACLU) criticized the legislation as &quot;misguided&quot; and &quot;misinformed,&quot; saying that it would restrict the freedom of organizations that provide critical services to women and children.\n\n&quot;We have grave concerns about this bill because it's divisive, discriminatory, and fiscally irresponsible,&quot; Laura Murphy, director of ACLU's Washington Legislative Office, said in a statement. &quot;It would allow states to use federal funds to provide child welfare services in a way that would exclude those who are not religious or who don't subscribe to traditional family values. This means that women and children who need assistance would be turned away, and those who provide critical services would be fired or not hired.&quot;\n\nThe bill, which was approved by the Senate Committee on Finance on Wednesday, now heads to the Senate floor.]" time="0.025"><properties><property name="score" value="0.032560776180395115" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03256078&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03256078
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[HUNDREDS OF people are expected to attend a public meeting in Clonmel this evening which will focus on the recent developments in the Irish Water saga.\n\nClonmel Borough Council has granted permission for the meeting to be held in the Tower Hotel at 7pm.\n\nGarda\xed have also been notified of the meeting but a spokesperson said they are not expecting any trouble.\n\nThe meeting comes in the wake of a massive turnout at a public meeting on Sunday last where thousands of people flocked to a hall in Carrick-on-Suir, Co Tipperary to hear speakers on the Irish Water issue.\n\nThe Tipperary Star reported today that a Fine Gael representative has complained to the party\u2019s headquarters about a leaflet produced by the Carrick on Suir branch of the party.\n\nThe leaflet called on people to attend the meeting on Sunday last in order to oppose Irish Water and the installation of water meters.\n\nSpeaking to the Star, Fine Gael TD for Tipperary South Noel Coonan said the leaflet had damaged the image of Fine Gael in the area.\n\nHe also described it as a \u201cdistortion of facts\u201d.\n\n\u201cIf this is being done in the name of Fine Gael, it is very serious,\u201d he said.\n\n\u201cI was very disappointed to see this leaflet which was distorting the facts. I\u2019m very annoyed about this and I will be taking it further.\u201d\n\nCouncillor Coonan also questioned the language used in the leaflet and the manner in which the information was presented.\n\nFine Gael councillor in Carrick on Suir, Councillor Martin Browne defended the leaflet, saying that it was a statement from the party\u2019s national executive and was a totally legitimate use of the party\u2019s logo.\n\n\u201cIt was used to promote a public meeting which was organised by a totally independent group,\u201d he said.\n\nThe group was not aligned to the party in any way and was not part]" time="0.027"><properties><property name="score" value="0.01555237933846007" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01555238&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01555238
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Horsehead\u2013Saxonburg Bridge\n\nThe Horsehead\u2013Saxonburg Bridge was a single-track, long railroad bridge in Armstrong County in the U.S. state of Pennsylvania. It carried the Bessemer and Lake Erie Railroad (B&amp;LE) main line across the Allegheny River and was the highest rail line on the B&amp;LE, with a maximum elevation of .\n\nThe bridge connected the small hamlet of Saxonburg on the west side of the river with Horsehead Landing on the east side.\n\nIn 2016, the Pennsylvania Department of Transportation announced plans to replace the bridge with a new one. The current bridge will be removed in the summer of 2019.\n\n]" time="0.015"><properties><property name="score" value="0.0376707010281363" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0376707&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0376707
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[John Astley (diplomat)\n\nJohn Francis Melhuish Astley, CMG (born 2 January 1950) is a former British diplomat.\n\nAstley was educated at St Paul's School, London, and the University of Cambridge, where he was a member of St John's College.\n\nAstley joined the Foreign and Commonwealth Office (FCO) in 1973. He served as Consul in Santa Cruz, Bolivia, 1973\u20131975, Deputy Consul General in San Francisco, California, 1981\u20131984, Deputy Head of the Middle East Department, FCO, 1985\u20131987, Private Secretary to the Secretary of State for Foreign and Commonwealth Affairs, 1987\u20131989, and Ambassador to Bahrain, 1990\u20131993.\n\nIn 1993 Astley became the first Permanent Under-Secretary of State at the FCO since the Second World War. He was appointed Companion of the Order of St Michael and St George (CMG) in the 1995 New Year Honours. From 1996 he was Ambassador to the Netherlands.\n\nIn 1999, during his tenure as Ambassador to the Netherlands, he was informed that his name appeared on a list of paedophiles. He had no connection to the accusations, but he resigned anyway, rather than face public humiliation.\n\nFrom 2000 to 2006 Astley was Chair of the Northern Ireland Audit Committee, an all-party group of the Northern Ireland Assembly, reporting to the Assembly and the Northern Ireland Audit Office.\n\nAstley was a director of Intraco Management Services Ltd, a recruitment firm.\n\nIn 2006 he was appointed a non-executive director of the Ministry of Defence's Defence Business Services organisation, where he remained until 2010.\n\nAstley is a member of the Athenaeum Club.\n]" time="0.030"><properties><property name="score" value="0.10124731541556524" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.10124732&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.10124732
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Beata D\u0105browska\n\nBeata D\u0105browska (born 8 January 1967) is a Polish fencer. She competed in the women's individual and team \xe9p\xe9e events at the 1988 Summer Olympics.\n]" time="0.014"><properties><property name="score" value="0.03228424677662045" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03228425&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03228425
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Wilfred Harris (cricketer)\n\nWilfred James Harris (1884 \u2013 16 January 1935) was an English cricketer. He played two first-class matches for Cambridge University Cricket Club in 1905.\n\n]" time="0.015"><properties><property name="score" value="0.04307397989687378" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04307398&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04307398
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Stephen W. Kearney\n\nStephen Watts Kearney (June 17, 1794 \u2013 October 31, 1848) was an officer in the United States Army who is best known for his leadership in the Mexican\u2013American War, in particular the Battle of Monterrey and the Battle of Buena Vista.\n\nKearney was born in Newark, New Jersey. He was a descendant of John Kearney (1635\u20131701), a native of Ireland who was the first of the family to settle in what became the United States. John's son Richard (1668\u20131734) was born in Bristol, England, but became a fur trader in what was then the English colonies and later became New Jersey. Richard was commissioned as an ensign by British colonial governor Jonathan Belcher in 1709.\n\nRichard's son James (1703\u20131785) was a general in the New Jersey colonial militia during the French and Indian War. James's son Stephen Watts Kearny, the namesake of the general, was a captain in the War of Independence and was at the surrender of General John Burgoyne. His home, the Kearney Mansion, has since been preserved in Newark.\n\nIn 1794 he moved with his parents to Louisville, Kentucky. His father, Alexander Morris Kearney, a native of South Carolina, was a captain in the American Revolutionary War. In the War of 1812, Kearney participated in the Battle of Lundy's Lane, where he was wounded and captured by the British. After being released, he joined the Missouri Rangers and, in 1813, married Mary Radcliffe Matthews. In 1814 he joined the 5th United States Infantry as its sixth lieutenant and served in the war against the Creek nation. In 1819 he was promoted to captain and assigned to the 1st U.S. Infantry. Kearney resigned from the Army in 1822 and became superintendent of Indian affairs at St. Louis, Missouri.\n\nIn 1832 Kearney was appointed as governor of the New Mexico Territory, where he was active in promoting settlement and negotiating treaties with the Native Americans. Kearney participated in the Gila Expedition of 1837 and then commanded the Army's Third Dragoon Regiment in Florida in 1837 and 1838.\n\nDuring the Mexican\u2013American War, Kearney was appointed colonel of the 1st New York Dragoons on June 30, 1846. Kearney's new regiment was sent to Santa Fe, New Mexico, to reinforce the Army of New Mexico under Col. Stephen Watts Kearny (no relation). There they joined the march on the capital of Mexico City. On the way, Kearney's regiment carried out several successful raids, including the Battle of La Mesa, where he captured a large amount of ammunition and supplies. He took the Puebla de Los Angeles without opposition on December 7, 1846. There he received word that his promotion to brigadier general had been approved.\n\nKearny, reinforced by the 3rd Infantry and Mormon Battalion, took the city of Monterrey on September 25, 1846. During the battle, a cannonball tore away part of Kearny's jaw and neck.\n\nFollowing the battle, Kearney was carried to Saltillo for medical treatment. While he was recovering, Lt. Colonel John Garland was given command of the 1st Dragoons. In February 1847, Kearney returned to the army and resumed command of the dragoons. Kearney served in the Battle of Buena Vista, and later he fought in the Taos Revolt. He returned to the United States in October 1847 and was assigned to Fort Leavenworth in the Kansas Territory.\n\nKearney, a bachelor, died of cholera at Fort Leavenworth at the age of 54. He was buried in the Fort Leavenworth National Cemetery.\n\nSeveral U.S. Army installations are named in honor of Kearney, including Kearney Barracks in Wilhelmshaven, Germany, the Kearney Building at Fort Leavenworth, and Fort Kearny, Nebraska, which is now a state historic site.\n\n\n]" time="0.083"><properties><property name="score" value="0.10281943332789088" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.10281943&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.10281943
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Los R\xedos Province, Chile\n\nLos R\xedos (Spanish for &quot;The Rivers&quot;) is one of four provinces in the southern Chilean region of Los R\xedos (IV). The capital is Valdivia.\n\nAs a province, Los R\xedos is a second-level administrative division of Chile, governed by a provincial governor who is appointed by the president.\n\nThe province comprises eleven communes (Spanish: &quot;comunas&quot;), each governed by a municipality consisting of an alcalde and municipal council:\n\nThe province spans an area of , the second largest in the region. According to the 2002 census, Los R\xedos Province had a population of 253,331 inhabitants (133,315 men and 120,016 women), giving it a population density of . Of these, 119,262 (46.4%) lived in urban areas and 134,069 (53.6%) in rural areas. Between the 1992 and 2002 censuses, the population grew by 1.9% (4,243 persons).\n\nAs a province, Los R\xedos is a second-level administrative division of Chile, which is further divided into eleven communes.\n\nWithin this province are the most important archaeological ruins in all of southern Chile. The ancient city of &quot;Puquios&quot;, capital of the &quot;Lentas&quot; nation, is located in the commune of Valdivia. Other ruins include those at Chep\xe9n, Chichihualco and the ancient fortifications of Valdivia. Also of interest is the &quot;Cueva del Milod\xf3n&quot;, which features the remains of an extinct species of giant sloth.\n\nThe native tree &quot;Fitzroya cupressoides&quot;, locally called &quot;lauca&quot; or &quot;lauco&quot;, is a common tree in the forests of the coastal mountains.\n\nThe Rupanco and Quinchao Islands in the Chonos Archipelago are part of Los R\xedos.\n\nLike other southern regions of Chile, Los R\xedos experiences a moderate climate. Its weather is affected by the cold Humboldt Current which flows along the west coast of South America from Antarctica. During the summer months, temperatures can reach , while in the winter temperatures hover around .\n\nThe province is one of the most forested ones in Chile. Its economy is based on timber, salmon and tourism.\n\nRupanco and Quinchao Islands are part of the commune of Quinchao. The latter, together with a few minor islands, forms the &quot;Ciudad Insular&quot; (Inhabited Island) of Quinchao, with approximately 3,500 inhabitants, and is administered as part of the &quot;Comuna&quot; of Quinchao, in the province of &quot;Llanquihue&quot;.\n\n]" time="0.029"><properties><property name="score" value="0.06851588131418039" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06851588&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06851588
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[In many places there are no longer any young men at all. According to the World Health Organization, &quot;in countries with very low fertility, the proportion of young people may halve in the next 40 years&quot;. At this rate, by 2100 a majority of the world's population will be over 50. We will then have to find a way to accommodate the needs and aspirations of an unprecedented number of very old people in a society that has been moulded for decades around the needs of youth.\n\nNowhere is this more true than in Japan, the world's fastest-ageing country. The greying of Japan \u2013 &quot;silver tsunami&quot; or &quot;grey goo&quot; as it is known \u2013 is by now well known. There are a number of reasons for this. Japanese have one of the longest life expectancies in the world, with women expected to live to 87 and men to 80. Meanwhile, a long-standing national emphasis on having small families and long working hours \u2013 for both men and women \u2013 have kept the birth rate below replacement levels. And Japan's postwar economic miracle has not been kind to the young. The country's population is shrinking, partly because of a low birth rate and partly because of a high level of immigration.\n\nBut what is less well known is that this graying of Japan \u2013 and other nations such as Germany, China, South Korea and Thailand \u2013 is not just due to longer life expectancy and lower birth rates. There is another factor in play, and one that could have dramatic implications: very low fertility rates, particularly among younger women. It's estimated that around a quarter of the population will be aged over 65 by 2050.\n\nIn Japan, the &quot;total fertility rate&quot; (the number of children a woman is likely to have in her lifetime) is just 1.4 \u2013 and has been for the past 20 years. Among women aged between 15 and 24, the rate is even lower at 0.8, lower even than the famously low rate in Germany (which is 1.38). And in Tokyo, that rate is even lower at 0.7. By contrast, in 2010, the average rate in Europe was 1.6. The fertility rate in England and Wales in 2010 was 1.9.\n\nThe world's media are full of breathless stories about Japan's crisis of depopulation. Last year, an article in the International Business Times said that &quot;Japanese women are rejecting marriage and motherhood at a staggering rate&quot;. In the New York Times, another piece claimed that Japan is &quot;exporting its young women&quot; and is now &quot;facing a demographic crisis of epic proportions&quot;. More soberly, the BBC's business editor Robert Peston has written: &quot;Without a dramatic increase in its birth rate, Japan will face economic and social ruin.&quot;\n\nThe sense of panic is understandable. &quot;The number of young people will go down by two-thirds by 2035,&quot; says demographer Toshio Kanazawa. &quot;That means not only a decline in the size of the workforce but a dramatic increase in the number of retirees. Who will pay for them?&quot;\n\nThe answer is &quot;women&quot; \u2013 or more specifically, their wombs. And in this, there is something not quite right.\n\nIn Tokyo, I meet Nana, a 29-year-old who is employed in the financial sector. She is smartly dressed in a suit and speaks good English. She tells me she has a boyfriend and they are very happy. She is, she says, the typical &quot;She Economy&quot; \u2013 a young woman working in the city and her parents are very pleased. &quot;They want me to settle down,&quot; she says, &quot;and have children.&quot; But she's not so sure.\n\n&quot;It's too late,&quot; she says. &quot;I've got to think of my future. And I just don't know if I could afford it \u2013 having children. I'd need a bigger flat, and pay for childcare, and a new car.&quot;\n\nFor Nana, and for many Japanese women like her, the economic reality of having children is daunting. &quot;I have two brothers,&quot; she says, &quot;and they have wives and children. The wives don't work, and they have one or two children. One of my sisters-in-law doesn't even work outside the house. They don't have much money, but my parents love them. They are very happy. So I know what my parents want for me. But it's impossible. I can't do it.&quot;\n\nThis is a scenario that is playing out in homes across the country. The situation is, in many respects, rather unique. As Peston points out, Japanese women who stay in employment tend to have children later. As a result, while there is a generation of Japanese in their early 30s who have given up on having children, there is a generation of younger women who are still trying to get pregnant.\n\n&quot;I think it's pretty much impossible to have two or three children in Tokyo,&quot; says Seiko, a 35-year-old marketing executive. &quot;I don't think I could afford it. You have to pay so much for childcare. And then, you've got to give up work. I just couldn't do it.\n\n&quot;But, still, my mother is pressuring me to have children. If you have a boyfriend, you can't get married. It's impossible. You have to marry first, then have children. So, if you are a working woman, you have to get married, have children, and then stay at home with your children. That's it. And that's what everyone wants. That's the ideal.&quot;\n\nAt the root of this ideal is Japan's traditional gender-role society \u2013 one that has stayed remarkably the same since the end of the second world war. It is not, perhaps, surprising that such a rigid society is under strain. And not only for women. It is also under strain for men.\n\n&quot;Women in Japan are not treated well,&quot; says Toshio Kanazawa. &quot;They don't earn a lot of money. They are discriminated against in many ways. That's why there aren't enough babies.&quot;\n\nNot everyone agrees. &quot;I don't think it's true,&quot; says Nana, who, I've noticed, has referred to herself as &quot;boku&quot; (which means &quot;I&quot;) all evening, rather than the more feminine &quot;watashi&quot; (which means &quot;I&quot;). &quot;Women in Japan are strong. They do well at work and they earn a lot of money. They can do everything. So why should they have children?&quot;\n\nIn some respects, Nana has a point. Japan is a country in which young women are breaking glass ceilings. The number of women in senior management in Japan's 3,000 largest companies is at a record high. And in 2013, the World Economic Forum rated Japan the fourth most &quot;gender-equal&quot; nation in the world.\n\nThe Japanese government has encouraged this development. Ever since the Second World War, successive prime ministers have tried to get more women into the workplace and have even gone so far as to tell men to &quot;say goodbye to their bonuses if they refuse to let their wives and daughters work outside the home&quot;.\n\nBut while some Japanese women are finding their way into senior positions, many are also facing a harsh reality: most large Japanese corporations have an unofficial &quot;marriage bar&quot; that prohibits the promotion of women who are married and have children. This is particularly true of men. According to a study by the Japanese government, 61.5 per cent of male managers are married. By contrast, only 42.2 per cent of female managers are married. This means that women with children can be in a very difficult position, because \u2013 as Shihoko Goto of the Wilson Center points out \u2013 they face a choice between marriage and career. And that is not a choice that men face.\n\n&quot;If you look at people in their 30s,&quot; says Seiko, &quot;I think many of them are not getting married and having children. If you don't have children, you have more time and money to enjoy your life. So why not?\n\n&quot;I don't think that women are rejecting marriage and family,&quot; she says. &quot;It's just that if you want children and you are a career woman, it's impossible. There is no way. You can't do both. And so you have to make a choice. Either you go for your career or you go for marriage and family. That's what I think is happening.&quot;\n\n&quot;In Japan,&quot; says Goto, &quot;it's very difficult for men and women to be able to balance work and family. So, if you are a career woman, you will have to sacrifice family. That's the problem. That's why so many women don't have children. Because they want to focus on their career.&quot;\n\nThis is a view echoed by Ai Kato, a 29-year-old tax accountant. &quot;It's true that if you work in a large Japanese company, you cannot get married or have children if you want to keep your job,&quot; she says. &quot;That's the reality. It's a matter of the company culture. It's difficult to balance both. And so people are having fewer children.&quot;\n\nJapan has never really known anything else. &quot;Since the Meiji era [from 1868 to 1912] Japan's population has been very stable,&quot; says Kanazawa. &quot;So they think that is normal. But it's not. If you look at Germany, their population is falling too, but they are getting immigrants. And in the US and the UK, the population is stable, but people are having more children. So there are many different types of demographic change, and Japan is one of the most extreme.&quot;\n\nWhat is less clear is why this is happening. The popular explanation is that Japan is a victim of the phenomenon known as the &quot;demographic time bomb&quot; or &quot;]" time="0.083"><properties><property name="score" value="0.05277351950351836" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05277352&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05277352
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Australian Computer Emergency Response Team\n\nThe Australian Computer Emergency Response Team (AusCERT) is the recognised national computer security incident response team of the Australian Government. AusCERT works closely with the Australian Communications and Media Authority (ACMA), the Australian Federal Police (AFP), the Department of Defence (Defence), the Australian Taxation Office (ATO), and other government departments and agencies to deal with computer security incidents on an operational level. The organisation also provides specialist security incident response advice and services to the private sector.\n\nAusCERT was established in 1987 under the auspices of the Australian Computer Society (ACS) and is funded by the Australian Government through the Department of Communications and the Australian Government Information Management Office (AGIMO).\n\nAusCERT provides services to the public and private sectors in the following areas:\n\nThe headquarters for AusCERT are in the inner-city Melbourne suburb of Fitzroy, and are situated at the ACS Victoria. Other AusCERT team members are located around the country, including in Sydney, Brisbane, Adelaide and Perth.\n\nIn the early 1980s, the Australian federal government recognised the need for a computer emergency response capability and developed the Computer Emergency Response Team (CERT) concept. The first CERT team was established in December 1987 as the Computer Incident Advisory Capability (CIAC) under the auspices of the Australian Computer Society (ACS) at the University of Melbourne. In the mid-1990s, CIAC changed its name to AusCERT and moved its headquarters to the Department of Communications in Canberra. Since then, it has moved to the AGIMO building in Belconnen, ACT. In January 2013, AusCERT relocated to the City of Melbourne.\n\nAusCERT is a specialist organisation that provides advice and assistance in computer security incidents to its constituent members. As an operational unit, AusCERT:\n\n\nAusCERT's incident management process incorporates the Australian Signals Directorate's Incident Handling Working Model, an approved Australian Standard, AS 4801-2005.\n\nAusCERT is the designated lead agency for computer security incidents within Australia's Critical Infrastructure community, as defined by the Communications and the Arts portfolio of the Commonwealth Government.\n\nSince 2001, AusCERT has been responsible for issuing national warnings and alerts on a number of security issues, as directed by the Australian Communications and Media Authority (ACMA).\n\n\nAusCERT is active in a number of national and international computer security forums. AusCERT regularly participates in the Cyber Storm exercises conducted by the US Department of Homeland Security, as well as the Cyber Storm exercises hosted by the Singapore Infocomm Development Authority. AusCERT has also participated in the Cyber Security Exercise Series (CSES) developed by the Government of New Zealand. In 2003, AusCERT attended the UK's National Infrastructure Security Coordination Centre (NISCC) / British Computer Emergency Response Team (BCERT) course, and was the first international agency to receive the NISCC's Advanced Incident Handling course.\n\nOn 8 September 2005, AusCERT signed a Memorandum of Understanding with the US Department of Homeland Security's United States Computer Emergency Readiness Team (US-CERT) to formalise cooperation in the areas of:\n\n\n]" time="0.084"><properties><property name="score" value="0.03395354471992132" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Dick Carr\n\nDick Carr (born March 23, 1941) is a former American football player and coach. He served as the head football coach at Washington &amp; Jefferson College from 1978 to 1980 and at Allegheny College from 1990 to 1999, compiling a career college football record of 55\u201383\u20132.\n]" time="0.015"><properties><property name="score" value="0.03836829067943026" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03836829&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03836829
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Jandiala Guru railway station\n\nJandiala Guru railway station is a railway station in Sangrur district, Punjab. Its code is JG. It serves Jandiala Guru city. The station consists of 3 platforms.\n\nThe following trains stops at Jandiala Guru station :\n\n]" time="0.014"><properties><property name="score" value="0.001746023471016489" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00174602&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00174602
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Balantiocheilos melanopterus\n\nBalantiocheilos melanopterus is a species of fish in the family Cyprinidae.\n\nIn the wild they are found in southern Borneo, the lower portions of the Mekong River basin, and in the Chao Phraya and Mae Klong River basins. They are usually found in large, slow-flowing, and often turbid rivers. They live near the bottom, sometimes swimming over sandy areas, with many aquatic plants. They are often found in medium-sized shoals. They are mostly omnivorous. They are one of the smallest cyprinids. They have a maximum length of .\n\nThey are well-known for their remarkable abilities to tolerate very low water temperatures. They are known to survive long periods at temperatures as low as . However, when exposed to high water temperatures of , their metabolism starts to fail, which eventually leads to death. For this reason they are not recommended for tropical aquariums, but can make an excellent and hardy aquarium fish for more temperate regions.\n\n]" time="0.023"><properties><property name="score" value="0.06055693413573593" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06055693&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06055693
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Ahmed Harkat\n\nAhmed Harkat (, ; born April 27, 1971 in Ottawa, Ontario, Canada) is a Pakistani Canadian and Muslim who was detained as a threat to national security and imprisoned for two years in Canada.\n\nOn May 1, 2005, a Canada-wide arrest warrant was issued for Harkat by the Federal Bureau of Investigation (FBI) of the United States. On July 18, 2005, Ahmed Harkat was arrested under a Security Certificate by the Government of Canada, at his home in Ottawa, on the grounds that he was involved in terrorist activities, in particular in association with the al-Qaeda organization. On November 23, 2005, Ahmed Harkat was detained indefinitely by the Federal Court. He was released on strict house arrest conditions on February 1, 2006, after the Supreme Court of Canada ruled that his detention was unconstitutional. On December 13, 2007, the Court of Appeal for Ontario found that Harkat was detained lawfully. However, the Court ordered a new detention review hearing to be held before the end of June 2008. The Supreme Court of Canada refused to hear his appeal on February 15, 2008.\n\nHarkat applied for a judicial review of the decision of the Minister of Public Safety to order his detention. The Federal Court of Canada dismissed his application on December 14, 2006. Harkat has appealed to the Supreme Court of Canada.\n\nThe Federal Court decision came a week after the United States Senate rejected a bill to close the Guantanamo Bay detention camp.\n\nIn August 2008, Harkat was living under 24-hour house arrest, with some restrictions.\n\nOn June 17, 2009, the Supreme Court of Canada ruled that the security certificate under which Harkat was detained was valid and that he must be deported.\n\nOn July 29, 2015, the Federal Court of Appeal upheld a lower court decision and ruled that Harkat could stay in Canada. The ruling meant that Harkat could not be removed from Canada and remained under a strict form of house arrest.\n\nOn January 12, 2017, the Supreme Court of Canada refused to hear his appeal.\n\nAhmed Harkat was born in Pakistan in 1971, and he came to Canada in 1995. He speaks some English and has a wife and three daughters. He is a pizza delivery driver in Ottawa, Ontario, Canada.\n\nHarkat worked in Peshawar, Pakistan as a relief coordinator for Care International Pakistan, an international humanitarian organization, until 1999. He then went on to work as an accountant and cashier at Afghan Relief in Peshawar, a project of the National Council for Voluntary Social Services (NCVSS).\n\nHarkat came to the attention of Canadian intelligence officials after being deported from Dubai in April 2000, after he failed to obtain a visa for the United States. He has never visited Afghanistan, a key requirement of a security certificate. Harkat also states that he was not involved in any terrorist activity, and has never supported or associated with a terrorist organization.\n\nHarkat was arrested on December 10, 2002. He was charged with knowingly facilitating a terrorist activity and knowingly participating in the activities of a terrorist group.\n\nOn January 23, 2003, a judge granted the Royal Canadian Mounted Police (RCMP) an order to monitor his activities and to search his home. On March 20, 2003, Harkat was released under a strict curfew. The surveillance and house arrest was lifted on September 11, 2003.\n\nOn February 10, 2004, the judge in the case ruled that Harkat would not be allowed to work as a relief coordinator in his community or to interact with a list of specified persons, including the media. On March 10, 2004, Harkat's curfew was changed from 11 p.m. to 6 a.m., and the list of people he was not allowed to contact was expanded to include more than 100 people, including several Members of Parliament.\n\nOn May 11, 2004, the judge in the case ruled that Harkat would not be allowed to attend community events, and that he would be required to meet regularly with RCMP officials.\n\nOn August 16, 2004, Harkat was accused of writing love letters to two married women, one of whom was an RCMP officer.\n\nOn September 15, 2004, the Attorney General of Canada, under Paul Martin, asked that Harkat be detained for an additional six months, which was approved by a judge on October 12.\n\nOn September 30, 2004, Harkat's wife, Sophie, appeared on a news show, the &quot;National Magazine&quot; on CBC.\n\nOn January 21, 2005, the Federal Court of Canada ordered Harkat to remain under a 24-hour surveillance and under house arrest. On January 23, 2005, he was required to surrender his passport.\n\nOn May 1, 2005, a Canada-wide arrest warrant was issued by the FBI for Hark]" time="0.096"><properties><property name="score" value="0.038700719568423726" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03870072&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03870072
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[By DontHassleDaHoff\n\nWe're heating things up with PROJECT: Overcharge in the upcoming Patch 7.24b!\n\nPROJECT: Overcharge\n\nAssault up, minions! Project: Overcharge replaces Stormrazor and Stormguard Banner, empowering your entire team with 10% increased attack speed and 8% increased movement speed for 15 seconds. In addition, you'll gain Overcharge stacks while attacking enemy champions and playing as Summoner's Rift objectives. The more you participate, the faster you charge!\n\nWith the PROJECT pre-season approaching, we're giving Overcharge a test run on the PBE!\n\nHow do I get Overcharge?\n\nYour team starts with a fixed number of Overcharge tokens that reset when you recall. You earn more Overcharge by killing enemy champions and playing as objectives.\n\nObjective Total tokens Rift Herald 2 Scuttle Crab 2.5 Dragon 4\n\nKilling an enemy champion generates 1 stack, but if you earn more than 5 stacks, you'll go into &quot;overdrive&quot; and get a short burst of additional movement speed and attack speed!\n\n1-5 stacks: +8% AS / 10% MS\n\n6-10 stacks: +16% AS / 20% MS\n\n11+ stacks: +24% AS / 30% MS\n\nOverdrive lasts 1.5 seconds and grants up to +120% bonus AS and +180% bonus MS!\n\nOverdrive ends immediately if you're dealt damage greater than 4% of your maximum health or if you activate Stormguard Banner.\n\nPROJECT: Hunters\n\nPROJECT: Hunters rewards map objectives and creating map pressure for your team. If you slay a jungle monster or help your allies take down an enemy Rift Herald, Dragon, or Baron, you'll earn a stack of PROJECT: Hunters.\n\n1 stack for Rift Herald\n\n2 stacks for Dragon\n\n4 stacks for Baron\n\n3 stacks for a secured Baron kill\n\nWhen you earn a stack of PROJECT: Hunters, you and your teammates are empowered for 15 seconds with +10% bonus movement speed and +10% bonus attack speed.\n\nPlay the objective and rack up those stacks to hunt for the enemy team!\n\nExhaust changes\n\nWe're making some changes to Exhaust that are designed to let it work better with aggressive crowd control effects.\n\nWe want to remove Exhaust's bonus AD ratio because it makes it really difficult to fight champions with both Exhaust and high attack speed. However, Exhaust should feel more effective against high attack speed targets because its slow should be more powerful.\n\nWe're doing this by reducing the duration of the slow by 10% but increasing the duration of the debuff that prevents the target from autoattacking by the same amount. Overall, this means that it'll be easier to reach the minimum Attack Speed slow but harder to reach the maximum Attack Speed slow.\n\nAlso, this means Exhaust won't affect attack speed-based champs as much because it'll take more time for the debuff to tick down.\n\nWe'll continue to assess Exhaust and other individual adjustments to hit our goals in the pre-season.]" time="0.031"><properties><property name="score" value="0.05585111916950085" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05585112&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05585112
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Unemployed people who refuse a job offer will have their benefits cut under Government plans.\n\nThe Government will tighten the conditions which allow unemployed people to refuse work offers.\n\nWork and Income (WINZ) has been told it must provide greater support for people to get into work.\n\nBeneficiaries who refuse a suitable job offer will have their benefits reduced under new measures announced by Social Development Minister Paula Bennett.\n\nREAD MORE:\n\n* Government plans crackdown on parents who won't work\n\n* PM takes aim at beneficiaries who refuse to work\n\n* Thousands of solo parents' benefit slashed under 'tough love' plan\n\n* More sanctions on beneficiaries as Government claims welfare reforms are working\n\n* Benefit sanctions have forced hundreds to quit job search\n\n* Beneficiaries told 'get a job' in get tough letter\n\n* Benefit sanctions 'designed to make you feel like you're doing something wrong'\n\n* No Job? No Benefit.\n\nThere will be exceptions for those who cannot work due to illness or disability, and those in serious financial hardship.\n\nThose who had been receiving a sickness or invalids benefit, such as the invalids benefit, the sickness benefit, or the severe disabilities benefit for at least six months in the past two years, will be exempt from the tougher requirements.\n\nHowever, all job seekers who had been on a sickness benefit for six months or more would be required to actively look for a job, even if they were on a sickness benefit.\n\nThe measures, contained in a Ministry of Social Development discussion paper, would apply to new and existing beneficiaries.\n\nThey would also apply to beneficiaries who are looking after a child under six months old, or a child who is in the &quot;prescribed age range&quot; of six to 12 months, or is aged under two years old.\n\nPeople with children over two years old would be expected to look for work once the youngest child turned two.\n\nBennett said jobseekers needed to be active and look for work, rather than wait to be told what to do.\n\n&quot;We don't think the system is working very well at the moment. The policy is asking for more people to go into paid employment.&quot;\n\nThe Government is under pressure to get more people into work, particularly with a labour shortage.\n\nJob vacancies were at a 10-year high, but there were not enough skilled workers to fill them, Work and Income chief executive Glenys Coughlan said.\n\nThere were also not enough job offers to meet the needs of unemployed people, Bennett said.\n\n&quot;We know there are around 28,000 job seekers who have been unemployed for a year or more, and nearly 30,000 unemployed who have been unemployed for at least two years,&quot; Bennett said.\n\n&quot;We want to get these people into the workforce.&quot;\n\nThe Government's target is for 55 per cent of people aged between 18 and 64 to be in paid work by the end of this year.\n\n&quot;This is a national problem. It's going to require a lot of hard work from all of us to solve it.&quot;\n\nThe discussion document is out for public consultation, and the measures are likely to be introduced in September.\n\nIf the measures are adopted, there will be increased reporting requirements, and consequences for not meeting work obligations.\n\nThere will be four levels of compliance measures.\n\n&quot;Where a person is placed at the highest level of compliance with their obligations, it would mean that a person may be subject to reduced or no benefit payments if they are not meeting their obligations,&quot; the discussion paper said.\n\nPeople on a benefit would be expected to have regular interviews with a case manager, and have mandatory obligations to get training and development.\n\n&quot;The expectation is that the person will be actively looking for work, and not sitting on the benefit,&quot; Bennett said.\n\n&quot;They can say no to jobs, but they will still need to look.&quot;\n\nThose who refused a job offer three times would be placed on the work preparation programme. They would be expected to actively look for work, and attend job interviews, if they were suitable for the position.\n\nIf the person did not have a suitable job offer, they would be expected to participate in the work preparation programme.\n\nThose on a benefit who refused to participate in the programme would have their benefit reduced.\n\nWork and Income would also actively seek out job opportunities, and encourage beneficiaries to apply for the jobs.\n\nBennett said the work preparation programme would not be mandatory.\n\nBut &quot;if you have two years of not being in work, you should probably have a little bit of pressure on you to do a little bit more&quot;.\n\n&quot;I wouldn't want to see a situation where there was lots of sanctions. I think there will be exceptions.&quot;\n\nThe work programme already had a requirement for those who were work-ready to actively seek out work.\n\nThe changes were &quot;very targeted&quot;, she said.\n\n&quot;If you've been on a benefit for a long time, we want to make sure you are taking steps to get into work.\n\n&quot;It's about saying to people 'get with it or we'll move you along'.&quot;\n\nThe tighter requirements for beneficiaries were part of a wider review of the welfare system.\n\nThe changes could see beneficiaries' obligations reduced if they were taking care of a seriously ill or dying family member.\n\nBut if the sickness was self-inflicted, such as drug and alcohol abuse, there could be tougher requirements.\n\nCoughlan said there was already a work-ready requirement for job seekers, but that it was flexible.\n\n&quot;I think we're not as prescriptive as Australia. I think they're much more prescriptive.&quot;\n\nShe said it was important to keep the labour market up to date.\n\n&quot;That's part of our focus, to make sure that we've got the market that matches up with the right skills.&quot;\n\nThere had already been a move towards a more results-based approach, she said.\n\n&quot;We're already in the business of 'what are the outcomes for the people we serve'?&quot;\n\nThe discussion document said the proposed measures would help reduce the number of beneficiaries, while supporting those who were genuinely unable to work.\n\n&quot;It is proposed that these policy changes will provide a tighter fit between the requirements for beneficiaries to engage in paid work, and the exceptions to those requirements,&quot; it said.\n\nThe new measures are part of the Government's welfare reforms.\n\n&quot;There is a real concern that people are falling out of the benefit system. We]" time="0.088"><properties><property name="score" value="0.010885932061366501" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Filth was first made into a play for the BBC\u2019s series of television plays in 2003, featuring Ewan McGregor, as a detective called Bruce Robertson. I don\u2019t think it\u2019s too far fetched to say that Bruce is the Scottish version of Neil Buchanan in American Psycho. Not necessarily in terms of the content of the story, but in the sheer fact that he is a very messed up guy, one who people can\u2019t stand, but at the same time can\u2019t stop looking at him.\n\nThis character is played to perfection by Jonny Lee Miller in the new film adaptation of Filth. The novel follows Bruce Robertson (Miller) a detective who, after starting the story off as a drunken, bitter asshole, finds out that he has cancer and so takes leave to get himself checked out. This action in turn leads to Bruce discovering that he is actually a massive pervert. He becomes obsessed with performing the most unspeakable of actions in order to be like the perverts that he so despises. The film follows Bruce as he conducts an investigation to find the person responsible for him being a terrible person and while he is still struggling with his illness.\n\nWhat is great about this adaptation is that it keeps the nastiness and the gritty, unpleasant content that is in the book. That might be what stops this film from being a complete success because this film is a difficult watch. It\u2019s not so difficult because of the content but because you will be so incredibly frustrated by the main character. He is an unpleasant person who you don\u2019t want to spend time with.\n\nIt\u2019s a testament to Miller\u2019s performance that he makes you want to keep watching even though you can\u2019t stand him. It\u2019s this nastiness that he exudes that makes him so engaging to watch. You will want to know how his investigation turns out and what his next move is, because he is truly awful. It also helps that he looks incredibly creepy, like he would kill someone at the slightest provocation.\n\nBruce is helped by a fantastic supporting cast. Jon Skellern plays Bruce\u2019s best friend Simon. You can tell that Simon has his concerns for Bruce and that his concern for him will only continue to grow as Bruce\u2019s behaviour becomes worse. Simon acts as the voice of reason when it comes to Bruce and does his best to help him, only to be shut down by Bruce\u2019s barbs and by Bruce himself. Kelly Macdonald plays Bruce\u2019s wife Susan. While Bruce gets into scrapes with his fellow detectives, Susan is struggling to cope with him and his illness. It\u2019s obvious that Susan is struggling with her husband\u2019s condition, and you will feel for her throughout the film.\n\nThere are a lot of big names in this film that will be recognisable to most people. Jonny Lee Miller, so good as Sherlock Holmes, plays Bruce with a brilliant blend of ferocity and uncertainty. While he is fully committed to the disgusting nature of his character, he also seems like he could completely lose it at any minute. Stephen Fry is outstanding as Inspector Siobhan Clarke, who is one of the few people who is able to make Bruce think twice about the things he does. Everyone is perfectly cast and help to make this film feel incredibly authentic.\n\nThe film does do a good job of adapting the book. Although the source material is sometimes over the top, the film takes that feeling and amplifies it to the next level. What is great about this film is that you do believe that all the characters are real people. You get a real sense of the community that Bruce has around him and that is what helps to make the film so engaging. You want to know what is going to happen to Bruce and how he is going to get out of the situations that he has put himself into.\n\nFilth is a good adaptation of the novel by Irvine Welsh. It does have the same nastiness that is in the novel, which is great for those of you who enjoy watching people doing things that you shouldn\u2019t be watching. The acting is excellent, with a standout performance from Jonny Lee Miller, who takes the idea of a horrible person to a whole new level. It\u2019s not the most pleasant of films to watch, but it\u2019s certainly one of the most engaging.]" time="0.093"><properties><property name="score" value="0.06772003288910435" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06772003&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.06772003
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[This board was created for the discussions of Nintendo's two major flagship gaming systems, the Nintendo DS and the Nintendo Wii. Some discussion about previous Nintendo systems is also welcome, but as always please try to keep the conversation in the current console's area.\n\n\n\nNext Display topics from previous: All Topics 1 day 7 days 2 weeks 1 month 3 months 6 months 1 year Sort by Author Post time Replies Subject Views Ascending Descending\n\nReturn to Board index\n\nJump to: Select a forum ------------------ Forums Info &amp; Announcements Read me first! Hints &amp; Tips Gaming Marketplace Recommended Source Books Role Playing - Tales of The Mind Forums Games &amp; RPG Rules The Sinister Spire Promethean: The Created The Dreaming Stone Vampire: The Masquerade Werewolf: The Apocalypse Exalted: The Emerald Dawn Wraith: The Oblivion Promethean: The Created Wraith: The Oblivion Promethean: The Created 3rd Edition Promethean: The Created 2nd Edition Werewolf: The Apocalypse Vampire: The Masquerade Werewolf: The Apocalypse Exalted: The Emerald Dawn Wraith: The Oblivion Wraith: The Oblivion Promethean: The Created Wraith: The Oblivion Vampire: The Masquerade Wraith: The Oblivion Vampire: The Masquerade Werewolf: The Apocalypse Vampire: The Masquerade Exalted: The Emerald Dawn Werewolf: The Apocalypse Exalted: The Emerald Dawn Vampire: The Masquerade Exalted: The Emerald Dawn Wraith: The Oblivion Wraith: The Oblivion Promethean: The Created Vampire: The Masquerade Werewolf: The Apocalypse Exalted: The Emerald Dawn Vampire: The Masquerade Promethean: The Created Werewolf: The Apocalypse Promethean: The Created 2nd Edition Werewolf: The Apocalypse Exalted: The Emerald Dawn Wraith: The Oblivion Promethean: The Created 3rd Edition Werewolf: The Apocalypse Promethean: The Created Werewolf: The Apocalypse Exalted: The Emerald Dawn Vampire: The Masquerade Promethean: The Created 2nd Edition Vampire: The Masquerade Promethean: The Created 3rd Edition Vampire: The Masquerade Exalted: The Emerald Dawn Werewolf: The Apocalypse Exalted: The Emerald Dawn Werewolf: The Apocalypse Exalted: The Emerald Dawn Wraith: The Oblivion Wraith: The Oblivion Wraith: The Oblivion Vampire: The Masquerade Vampire: The Masquerade Wraith: The Oblivion Wraith: The Oblivion Promethean: The Created Vampire: The Masquerade Promethean: The Created Werewolf: The Apocalypse Werewolf: The Apocalypse Promethean: The Created Exalted: The Emerald Dawn Wraith: The Oblivion Vampire: The Masquerade Promethean: The Created 3rd Edition Vampire: The Masquerade Vampire: The Masquerade Exalted: The Emerald Dawn Promethean: The Created 2nd Edition Werewolf: The Apocalypse Promethean: The Created Werewolf: The Apocalypse Exalted: The Emerald Dawn Vampire: The Masquerade Vampire: The Masquerade Exalted: The Emerald Dawn Promethean: The Created Wraith: The Oblivion Wraith: The Oblivion Vampire: The Masquerade Promethean: The Created Exalted: The Emerald Dawn Werewolf: The Apocalypse Vampire: The Masquerade Vampire: The Masquerade Exalted: The Emerald Dawn Promethean: The Created 2nd Edition Werewolf: The Apocalypse Vampire: The Masquerade Wraith: The Oblivion Werewolf: The Apocalypse Vampire: The Masquerade Wraith: The Oblivion Promethean: The Created Exalted: The Emerald Dawn Werewolf: The Apocalypse Vampire: The Masquerade Promethean: The Created 3rd Edition Wraith: The Oblivion Werewolf: The Apocalypse Vampire: The Masquerade Promethean: The Created Wraith: The Oblivion Vampire: The Masquerade Exalted: The Emerald Dawn Vampire: The Masquerade Exalted: The Emerald Dawn Wraith: The Oblivion Wraith: The Oblivion Wraith: The Oblivion Vampire: The Masquerade Vampire: The Masquerade Exalted: The Emerald Dawn Werewolf: The Apocalypse Vampire: The Masquerade Wraith: The Oblivion Wraith: The Oblivion Vampire: The Masquerade Promethean: The Created Exalted: The Emerald Dawn Werewolf: The Apocalypse Wraith: The Oblivion Promethean: The Created Werewolf: The Apocalypse Vampire: The Masquerade Vampire: The Masquerade Promethean: The Created 3rd Edition Werewolf: The Apocalypse Vampire: The Masquerade Exalted: The Emerald Dawn Wraith: The Oblivion Wraith: The Oblivion Vampire: The Masquerade Promethean: The Created 2nd Edition Vampire: The Masquerade Promethean: The Created Wraith: The Oblivion Wraith: The Oblivion Vampire: The Masquerade Vampire: The Masquerade Exalted: The Emerald Dawn Wraith: The Oblivion Wraith: The Oblivion Vampire: The Masquerade Exalted: The Emerald Dawn Wraith: The Oblivion Wraith: The Oblivion Promethean: The Created Wraith: The Oblivion Vampire: The Masquerade Wraith: The Oblivion Wraith: The Oblivion Wraith: The Oblivion Vampire: The Masquerade Vampire: The Masquerade Wraith: The Oblivion Vampire: The Masquerade Vampire: The Masquerade Vampire: The Masquerade Exalted: The Emerald Dawn Werewolf: The Apocalypse Werewolf: The Apocalypse Vampire: The Masquerade Exalted: The Emerald Dawn Promethean: The Created 2nd Edition Vampire: The Masquerade Promethean: The Created Vampire: The Masquerade Exalted: The Emerald Dawn Wraith: The Oblivion Wraith: The Oblivion Vampire: The Masquerade Promethean: The Created 3rd Edition Werewolf: The Apocalypse Vampire: The Masquerade Wraith: The Oblivion Wraith: The Oblivion Vampire: The Masquerade Wraith: The Oblivion Wraith: The Oblivion Vampire: The Masquerade Vampire: The Masquerade Vampire: The Masquerade Vampire: The Masquerade Wraith: The Oblivion Wraith: The Oblivion Wraith: The Oblivion Wraith: The Oblivion Vampire: The Masquerade Vampire: The Masquerade Vampire: The Masquerade Wraith: The Oblivion Wraith: The Oblivion Wraith: The Oblivion Vampire: The Masquerade Vampire: The Masquerade Wraith: The Oblivion Wraith: The Oblivion Wraith: The Oblivion Wraith: The Oblivion Vampire]" time="0.092"><properties><property name="score" value="0.36553966993796194" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Each year, the Duluth News Tribune names a Great Duluthian. This year's Great Duluthian is Bernie Betz, founder and president of Betz Equipment.\n\nBetz, an avid hunter and fisherman, is not only a successful business man, he is a giver. His Betz Family Foundation has donated over $5 million to various organizations and projects, including Duluth's Discovery Center and the Betz Sportsman's Club, which has operated since 1948.\n\n&quot;It has been a real honor to know Bernie, work with him, and to benefit from his support,&quot; says John Stowe, president and CEO of the Duluth Entertainment Convention Center (DECC).\n\nIn his acceptance speech, Betz shared a quote from Henry Ford: &quot;If you think you can do a thing, or if you think you can't do a thing, you're right.&quot; Betz says he tries to follow this philosophy in all aspects of his life.\n\nBetz said that there are too many good people in Duluth to name one &quot;Great Duluthian,&quot; but he humbly accepted the award. &quot;This is just a thank you,&quot; Betz said. &quot;I thank you for this.&quot;]" time="0.023"><properties><property name="score" value="0.05591817960944323" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05591818&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05591818
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Beauty\n\nI came across this article in Style Magazine about Kim Kardashian being named a rising star by Teen Vogue. Really?! Kim Kardashian is being compared to Audrey Hepburn, and Grace Kelly? I am not quite sure what Teen Vogue was thinking, but this makes me think they are more concerned about the bottom line of a magazine than really focusing on their readers and what they want to read about. What kind of beauty standards does Teen Vogue endorse?]" time="0.021"><properties><property name="score" value="0.038638881879210496" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03863888&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03863888
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[This is the first time I\u2019ve made this recipe and I absolutely loved it! It is now one of my favorites. It had a lot of flavor and the ginger and garlic added great zing. I used 1 tsp of ginger and 1/2 tsp of garlic. This was great served over brown rice. My husband raved about this meal. Thanks for a great recipe!\n\nI have made this for years but never reviewed it. The longer it sits the better it tastes. I do make it on a regular basis. I often do half coconut milk and half veggie broth. One of my family's favorite meals. I am thinking of using the coconut milk and broth when I cook it instead of water and bullion cube.]" time="0.023"><properties><property name="score" value="0.043883247383991186" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04388325&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04388325
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Le lancement de Canon EOS M5 est imminent et pour ce faire, la soci\xe9t\xe9 nippone lance sur les r\xe9seaux sociaux une campagne marketing ax\xe9e autour de la qualit\xe9 des vid\xe9os. Un teaser se prom\xe8ne donc sur la toile d\u2019apr\xe8s laquelle le bo\xeetier serait capable de filmer dans une r\xe9solution 4K (passe de 2,7K).\n\nLes qualit\xe9s vid\xe9o d\u2019un bo\xeetier d\xe9termin\xe9es par la r\xe9solution de l\u2019enregistrement de la vid\xe9o. La baisse de la r\xe9solution ne signifie pas pour autant que la qualit\xe9 diminue : on peut avoir une r\xe9solution 4K pour une vid\xe9o en HD, ou inversement.\n\nEt de toute \xe9vidence, la r\xe9solution 4K de la vid\xe9o de Canon EOS M5 semble plus qu\u2019une simple rumeur. Et c\u2019est un scoop ? Pas vraiment puisque le constructeur avait annonc\xe9 que son prochain compact avait la capacit\xe9 de filmer en 4K il y a un an. Et pour une raison qui n\u2019est pas claire, la marque nippone a tout simplement gard\xe9 ce secret.\n\nUne r\xe9solution 4K a de quoi attirer les foules, et surtout les utilisateurs d\u2019appareils photo. En effet, la 4K constitue aujourd\u2019hui un format de r\xe9f\xe9rence, \xe0 l\u2019instar de la r\xe9solution Full HD. Cela dit, la 4K n\u2019est pas forc\xe9ment un avantage pour un bo\xeetier de ce type, mais plut\xf4t pour les bo\xeetiers d\u2019entr\xe9e de gamme ou de milieu de gamme.\n\nCanon EOS M5 : plusieurs versions ?\n\nLa r\xe9solution 4K, comme l\u2019explique Dpreview, est un atout dans le cas des fabricants de bo\xeetiers car ceux-ci peuvent d\xe8s lors proposer des machines \xe0 un prix bas et vendre ensuite des objectifs (propres \xe0 l\u2019appareil ou non) \xe0 prix plus \xe9lev\xe9.\n\nCanon EOS M5 a beau \xeatre attendu au tournant, il ne faudra toutefois pas s\u2019attendre \xe0 ce que l\u2019appareil soit dot\xe9 d\u2019une stabilisation des images. Le bo\xeetier ne semble en effet pas disposer d\u2019une stabilisation optique d\u2019image, alors que la concurrence (compacts compris) poss\xe8de d\xe9sormais un syst\xe8me de stabilisation optique.]" time="0.031"><properties><property name="score" value="0.1550005453274612" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.15500055&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.15500055
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[// Use of this source code is governed by a BSD-style license that can be\n\n// Check if the field is initialized or not\n\n// Determine the kind of path to use\n\n// Generate the path from the module name, directory and field name.\n\n// Module name comes first, to keep a consistant ordering between native\n\n// native method code and java/native method code, to allow for more readable\n\n// Remove 'this' parameter from the function declaration\n\n// If the arguments are references (ie. the passed in argument is a local\n\n// variable), mark the parameter as read only.\n\n// If there is a non-factory constructor, use it\n\n// Always use the default constructor\n\n// A non-null `type` value, or `value` with a type annotation,\n\n// means that we are constructing a type.\n\n// Use the native constructor if it exists, otherwise fall back to\n\n// Use the name of the class as the default constructor name.\n\n// Check if a constructor exists on the type\n\n// If the value is a reference, use the name of the field.\n\n// Use the default constructor name for a value.\n\n// If the value is a reference, use the name of the field.\n\n// If a type is specified (name or value), it takes precedence.\n\n// If we have a type annotation on the parameter, it must match the type\n\n// If we have a name annotation on the parameter, it must match the name\n\n// Use the name of the class as the constructor name\n\n// If we have a type annotation on the parameter, it must match the type\n\n// If we have a name annotation on the parameter, it must match the name\n\n// Don't check for annotation matches in constructors.\n\n// Try to use the field name as a constructor name\n\n// If there is a non-factory method, use it\n\n// Always use the default constructor\n\n// A non-null `type` value, or `value` with a type annotation,\n\n// means that we are constructing a type.\n\n// Use the native constructor if it exists, otherwise fall back to\n\n// Use the name of the class as the default constructor name.\n\n// Check if a constructor exists on the type\n\n// If the value is a reference, use the name of the field.\n\n// Use the default constructor name for a value.\n\n// If the value is a reference, use the name of the field.\n\n// If a type is specified (name or value), it takes precedence.\n\n// If we have a type annotation on the parameter, it must match the type\n\n// If we have a name annotation on the parameter, it must match the name\n\n// Use the name of the class as the constructor name\n\n// If we have a type annotation on the parameter, it must match the type\n\n// If we have a name annotation on the parameter, it must match the name\n\n// Don't check for annotation matches in constructors.\n\n// Try to use the field name as a constructor name\n\n// If there is a non-factory method, use it\n\n// Always use the default constructor\n\n// A non-null `type` value, or `value` with a type annotation,\n\n// means that we are constructing a type.\n\n// Use the native constructor if it exists, otherwise fall back to\n\n// Use the name of the class as the default constructor name.\n\n// Check if a constructor exists on the type\n\n// If the value is a reference, use the name of the field.\n\n// Use the default constructor name for a value.\n\n// If the value is a reference, use the name of the field.\n\n// If a type is specified (name or value), it takes precedence.\n\n// If we have a type annotation on the parameter, it must match the type\n\n// If we have a name annotation on the parameter, it must match the name\n\n// Use the name of the class as the constructor name\n\n// If we have a type annotation on the]" time="0.087"><properties><property name="score" value="0.20935602428403247" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Why won't anyone on the Senate Banking Committee acknowledge that debt restructuring is needed for the Puerto Rico debt crisis to be solved? I just spent about two hours on CSPAN watching the Committee's hearings about Puerto Rico. Not once was the possibility of debt restructuring, which would mean debt reduction, mentioned. All that was mentioned was higher taxes, pension cuts, and deep government spending cuts, on top of what Puerto Rico is already doing. A complete deadlock. Puerto Rico has been austerity-impaired for a long time now, and has experienced a 12% drop in GDP as a result. No one can grow their way out of debt. Yet, even with this depression-like economy, Puerto Rico is still paying a huge amount of its income to its creditors. A lot of that debt comes from bonds that the government of Puerto Rico sold to investors to get the money to cover the huge losses that its electricity monopoly incurred because of oil price spikes. The government still has no way to fix that, since electricity is now being generated by private companies, due to a privatization law passed in 2013. The government of Puerto Rico could offer a debt exchange to investors, swapping the current bonds for ones with a longer maturity date and a lower interest rate, and then reinvest the proceeds into infrastructure projects, where it has a big deficit. But that possibility is completely ruled out. On the other side, the creditors are as intransigent as ever. They claim they can sue the US government if it tries to change their contracts, which is not true. They also claim that Puerto Rico is defaulting on its debt already, which is also not true. The government of Puerto Rico has been paying its creditors 100 cents on the dollar. There is simply no default risk. This is all a huge waste of time. All that the Puerto Rico hearings did was show that the US Congress is not capable of doing what it should have done long ago: allow Puerto Rico to restructure its debts. That would have given Puerto Rico a fresh start, and the creditors would have ended up with their money back. And we would have avoided all this useless talking. Now, we are just wasting time and making the situation worse. Puerto Rico's economy continues to decline. Puerto Rico's economy is expected to contract again by at least 1% in 2016, and another 2% in 2017. Most of the business owners I interviewed in Puerto Rico last month said they expect business to get worse in 2017. So the longer the creditors keep Puerto Rico in this situation, the worse it will get. In the hearings today, several senators brought up a resolution that the House of Representatives passed yesterday in a partisan vote, saying that Puerto Rico is already in default. That resolution is just to make a political point. The Supreme Court has already ruled that Puerto Rico cannot declare bankruptcy like a US municipality can, and cannot restructure its debt in a court. This was because of a law that was passed in 1984. The US Congress could have changed this law a long time ago, but they did not. The Republicans in the House of Representatives, however, do not want to restructure Puerto Rico's debts. Instead, they are trying to force the government of Puerto Rico to do all the reforms it already is doing, without giving it the opportunity to do the debt restructuring that will put its economy back on track. This was not really mentioned in the Senate hearings, but I found out from reporters that the Democrats on the Senate Banking Committee are not going to move forward with the House resolution. They understand that it is not going to pass. We need Congress to pass the resolution anyway, just to show that it has bipartisan support. In the House hearings, Representative Sean Duffy of Wisconsin said, &quot;This is not an emergency.&quot; This is a statement that no one in Puerto Rico agrees with. By delaying the debt restructuring, the economy of Puerto Rico is getting worse. More people are moving to the US mainland, which will further reduce the income that Puerto Rico is able to collect in taxes. More people are being forced to do cash jobs in the underground economy, because they cannot find other jobs. And those jobs do not get reported to the government. The longer this process is delayed, the worse it will get. But the creditors and some members of Congress do not seem to understand that Puerto Rico needs a fresh start. So they will continue to waste time, talking about things that are already being done, without any hope of any positive outcome. Puerto Rico's creditors just don't want to take a loss. But we are not doing Puerto Rico any favors by letting this situation continue for so long.]" time="0.094"><properties><property name="score" value="0.027178798660224556" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0271788&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0271788
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[A 21-year-old man died early Tuesday morning after being struck by a car while riding his bike in Logan Square, police said.\n\nAt 1:26 a.m., authorities responded to the intersection of Fullerton and Armitage avenues where the bicyclist was hit by a car, according to Chicago Police. He was taken to Advocate Illinois Masonic Medical Center, where he was later pronounced dead.\n\nThe man was identified as Fares Khaled, of the 900 block of North Francisco Avenue in Logan Square, according to the Cook County medical examiner's office.\n\nThe driver of the vehicle that struck the cyclist stayed at the scene and cooperated with police, authorities said. The investigation is ongoing.\n\nrsobol@tribune.com\n\nTwitter: @RosemarySobol1]" time="0.016"><properties><property name="score" value="0.08078812617692832" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08078813&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.08078813
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[I love google chrome's ability to store passwords and usernames.\n\nBut it only works on &quot;http&quot; sites. I often use many other types of internet links, which would be useful to store credentials for.\n\nIs there a way to store and retrieve passwords on all types of web addresses?]" time="0.018"><properties><property name="score" value="0.021144425720526216" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02114443&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.02114443
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[President Trump got into a heated exchange with CNN White House correspondent Jim Acosta at his press conference on Thursday, calling the reporter \u201cfake news\u201d and challenging his network\u2019s reporting.\n\nBut Trump\u2019s son, Eric Trump, says the president will always be fair and give all reporters the time of day.\n\nIn an appearance on Fox News\u2019 \u201cMedia Buzz,\u201d Eric said Trump may attack CNN but always makes time for the network\u2019s reporters.\n\n\u201c[T]hey can still come in. The reporter that he did call on [at the press conference], we always give them an opportunity at every press conference. He\u2019s always going to be fair,\u201d he said.\n\nADVERTISEMENT\n\n\u201cHe\u2019s always going to give them an opportunity, but we\u2019re not going to let them take potshots at our president and our administration,\u201d Eric Trump added.\n\nEric Trump also predicted that eventually the president will win the fight against \u201cfake news\u201d in the mainstream media.\n\n\u201cMy father is a fighter, and he\u2019s an unbelievable person. I mean, he\u2019s built an unbelievable business, and he\u2019s built an unbelievable life. He\u2019s a person that\u2019s been elected to lead this country, and he\u2019s going to do it with or without the media. He\u2019s going to do it with or without certain networks. And it\u2019s sad because he\u2019s such a good man, and he\u2019s such a good father,\u201d Eric Trump said.\n\n\u201cBut at the same time, if they\u2019re going to put themselves in a position where they\u2019re going to be speaking so much, they\u2019re going to continue to make mistakes,\u201d he added.]" time="0.022"><properties><property name="score" value="0.0047436445164139585" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00474364&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00474364
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[This is not a bug or something that is wrong. The game is not detecting your Shield as something that it can block. The animation itself looks fine. I don't know how you can replicate it yourself to get a screenshot, but what you are seeing is the default Shieldblock animation without your Shield in your inventory.\n\n\n\nIf you are referring to the broken animation for Normal and Magic Shieldblock, the bug with that is fixed on 0.11.4. I'm not sure when we'll be able to patch that for Ascendancy. Last edited by Mark_GGG on Nov 11, 2015, 8:54:36 AM Posted by Mark_GGG\n\non Grinding Gear Games on Quote this Post\n\nI mean my shield is in my inventory and the character can block while using the skill.\n\n\n\nThat's what I mean about not being a bug.\n\n\n\nIf it's a game mechanic that can be prevented somehow that would be really helpful to know. Posted by Greendogo\n\non on Quote this Post\n\n&quot; Greendogo\n\n\n\nThat's what I mean about not being a bug.\n\n\n\nIf it's a game mechanic that can be prevented somehow that would be really helpful to know. I mean my shield is in my inventory and the character can block while using the skill.That's what I mean about not being a bug.If it's a game mechanic that can be prevented somehow that would be really helpful to know.\n\nThe game can't detect your shield. When you block, the shield is on your back. When you use your shield in the skill, it isn't. The game can't detect your shield. When you block, the shield is on your back. When you use your shield in the skill, it isn't. Posted by Mark_GGG\n\non Grinding Gear Games on Quote this Post\n\n&quot; Mark_GGG &quot; Greendogo\n\n\n\nThat's what I mean about not being a bug.\n\n\n\nIf it's a game mechanic that can be prevented somehow that would be really helpful to know. I mean my shield is in my inventory and the character can block while using the skill.That's what I mean about not being a bug.If it's a game mechanic that can be prevented somehow that would be really helpful to know.\n]" time="0.040"><properties><property name="score" value="0.12226578207324698" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Uninstall all Office add-ins in Office 365\n\nIn this article\n\nWhen you install an Office add-in, you will be able to see the add-in in the Add-ins page in Word, Excel, or PowerPoint. However, it is not the case that you can only install add-ins from this location. In fact, you can add any website to the browser toolbar, and install a specific add-in from there.\n\nThis article describes the steps to completely uninstall all add-ins from all Office applications that you have in your tenant. This will allow you to completely remove all Office add-ins from your Office 365 tenant. This can be useful in a tenant cleanup or a company migration.\n\nImportant Before following the instructions in this article, please verify that your Office 365 subscription has the add-in licensing model and that the Tenant admin has permissions to run PowerShell scripts.\n\nIf you have any questions or feedback, please let us know in the Q&amp;A section on this page.\n\nManual process\n\nUse the following steps to completely uninstall all Office add-ins from all Office applications that you have in your tenant.\n\nImportant Before following the instructions in this article, please verify that your Office 365 subscription has the add-in licensing model and that the Tenant admin has permissions to run PowerShell scripts.\n\nOpen the SharePoint Online Management Shell. Run the following command: Get-PnPUtilization | Remove-PnPUtilization To remove the add-ins from Word, Excel, and PowerPoint, run the following commands: Get-PnPUtilization | ? {$_.Product -like &quot;*Word*&quot;} | Remove-PnPUtilization Get-PnPUtilization | ? {$_.Product -like &quot;*Excel*&quot;} | Remove-PnPUtilization Get-PnPUtilization | ? {$_.Product -like &quot;*PowerPoint*&quot;} | Remove-PnPUtilization To remove the add-ins from Outlook, run the following commands: Get-PnPUtilization | ? {$_.Product -like &quot;*Outlook*&quot;} | Remove-PnPUtilization\n\nPowerShell script\n\nThe following PowerShell script will completely uninstall all Office add-ins from all Office applications that you have in your tenant.\n\nImportant Before following the instructions in this article, please verify that your Office 365 subscription has the add-in licensing model and that the Tenant admin has permissions to run PowerShell scripts.\n\n# run as SharePoint Administrator $spAdmin = [Microsoft.SharePoint.Administration.SPAdminService]::GetAdmin # Connect to the SharePoint Online Management Shell $webApp = Get-SPWebApplication &quot;https://&lt;your_tenant_name&gt;.sharepoint.com&quot; $w = $webApp.LoginPage $w.ReturnUrl = $w.Url.AbsoluteUri + &quot;/_layouts/15/ PowerShell -ApplicationPage.aspx&quot; $cred = Get-Credential $credential = $spAdmin.Credentials.GetNetworkCredential($cred.UserName, $cred.Password) $session = New-PSSession -ConfigurationName Microsoft.Exchange -ConnectionUri https://ps.outlook.com/powershell/ -Credential $cred -Authentication Basic -AllowRedirection Import-PSSession $session -CommandName &quot;Remove-PnPUtilization&quot; -AllowClobber $excel = Get-PnPUtilization | ? {$_.Product -like &quot;*Excel*&quot;} | Remove-PnPUtilization $word = Get-PnPUtilization | ? {$_.Product -like &quot;*Word*&quot;} | Remove-PnPUtilization $powerpoint = Get-PnPUtilization | ? {$_.Product -like &quot;*PowerPoint*&quot;} | Remove-PnPUtilization $outlook = Get-PnPUtilization | ? {$_.Product -like &quot;*Outlook*&quot;} | Remove-PnPUtilization Remove-PSSession $session\n\nNote In the PowerShell script, you can run the commands individually to remove a specific add-in for a specific Office application, or you can remove all Office add-ins by running the Remove-PnPUtilization cmdlet without any parameters.\n\nAdditional information\n\nLearn more about how to remove a specific add-in for a specific Office application\n\nLearn more about how to add or remove an Office add-in in Office 365\n\nLearn more about the SharePoint Online Management Shell]" time="0.084"><properties><property name="score" value="0.0684424007007797" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Shepard and his team have done it again. I've been a fan of &quot;American Chopper&quot; since day one. I used to get angry with Paul Sr. for the way he treated the guys but I guess that was part of the show's draw. I have never seen a father and son fight like that on TV before. I still hope he comes back someday to work with his son, and they can rebuild the relationship they had before. I can't say enough good things about this season of American Chopper, with all the good work the team did for cancer patients, this is one of the most memorable seasons of the series.\n\nThe creativity of the builds is incredible. I'm so glad that they were able to get the Ford back on the road after the fire. They did an amazing job on the bike itself and the amount of work they put into making sure everything was functional on the bike, even the brakes, was inspiring. The customization they did on the ambulance was pretty impressive as well, especially how they made sure all the hydraulic and electrical components worked correctly. I also loved the idea of doing a bike with a hunting theme and the amount of work they put into making the motorcycles look like they belonged to an entire hunting party was great. They did an amazing job on the bike that played homage to the military and even the Great American motorcycle show was awesome! The barber shop bike and &quot;Appalachian Outlaw&quot; motorcycle were just as awesome. The personal touches that were put on these bikes really made them so much better.\n\nI know a lot of people criticize the shows for not showing the whole building process, but in my opinion, I like to see the creativity that the builders put into the bikes. The shows are a lot more enjoyable when I see what goes on in their heads when they come up with their ideas.]" time="0.030"><properties><property name="score" value="0.049250705107414514" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04925071&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.04925071
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[1060\xb0\n\nFacebook Shows Off Most Personal and In-Depth Profiles\n\nTHR: Facebook has always been about connecting with the people in your life, but the social network is taking things to the next level with new profiles that put the emphasis on the people you care about the most.]" time="0.015"><properties><property name="score" value="0.015227058297340168" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01522706&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01522706
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The drama of the week is all about the inability of a local physician to get paid by insurance companies. It is not an uncommon story in that many physicians cannot get paid by insurance companies for a variety of reasons. The difference in this story is that this physician not only tried to get paid, but took the insurance company to court to try to get paid.\n\nIt started when a physician sent an invoice to an insurance company for services rendered. The physician was told that the insurance company did not have the funds to pay the invoice and that they would pay the bill at a later date. The physician sent the insurance company a bill every month for about a year until the insurance company stopped communicating. The physician was then advised by a collection agency that if he wanted to be paid, he would have to take the insurance company to court.\n\nThe physician then did just that. He filed a claim with the District Court seeking the unpaid bill plus interest and costs. The insurance company tried to have the case dismissed by claiming the physician did not have standing to sue. The insurance company argued that the physician only had a contractual right to be paid, but that right had not been breached. They also argued that he could not have been damaged by their failure to pay since it was their money.\n\nThe physician won the case. He was awarded the $10,600 plus interest and costs. The judge agreed that the insurance company breached the contract when it failed to pay the bill. The judge also found that the insurance company's withholding of payment did cause the physician to suffer damages in the form of legal fees to pursue payment.\n\nAs for the insurance company, it is probably a good idea that they got caught. Since their contract was breached and they caused damages, they may be liable for punitive damages which could double the amount of the doctor's claim.]" time="0.030"><properties><property name="score" value="0.06169664901658019" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Forgive me if this has been answered already.\n\n\n\nI bought a standard prepaid plan and activated it last week. I purchased the standard plan with 1 GB of data and 1 GB of minutes and unlimited texts.\n\n\n\nToday, I purchased a pre-paid plan with 1 GB of data and 3 GB of minutes and unlimited texts. When I added the new plan to my phone, it overwrote my previous plan with only 1 GB of data.\n\n\n\nWhen I called Koodo, I was told that I could not keep both plans.\n\n\n\nIs this correct? Can I not use 2 different pre-paid plans? I don't want to lose the 1 GB of data in my phone because it's a limited offer.\n\n\n\nThanks]" time="0.016"><properties><property name="score" value="0.009254033965216957" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00925403&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.00925403
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Land of the Sun\n\n10.03.17 / CD I Love You Like a Brother\n\n08.11.16 / CD I Love You Like A Brother\n\n08.11.16 / LP (analog) Grand Romantic\n\n08.11.16 / CD We Didn't Do It For Your Love\n\n03.08.14 / CD Why Did You Make Me Care?\n\n04.10.13 / CD Some Hearts\n\n02.12.12 / CD]" time="0.022"><properties><property name="score" value="0.04580869724929992" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0458087&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0458087
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Google+ on pakotettu entist\xe4 tiukempaan sis\xe4ll\xf6ntuotantoon, kertoo TechCrunch. My\xf6s yhti\xf6n taloudelliset resurssit k\xe4rsiv\xe4t j\xe4\xe4dytyksist\xe4.\n\nLue my\xf6s Google jatkaa keskittymist\xe4 Androidiin\n\nYhti\xf6 aikoo t\xe4sment\xe4\xe4, ett\xe4 j\xe4rjestelm\xe4 ei ole en\xe4\xe4 pelkk\xe4 verkkopalvelu, vaan tiettyyn sis\xe4ll\xf6ntuotantoa vaativaan verkkopakettiin. Verkkopalvelua varten tarvittavat julkaisuprofiilit, lis\xe4\xe4ntyneet palvelut ja Google Play -sovelluskaupan julkaisut ovat poistumassa.\n\nMy\xf6s Googlen lis\xe4\xe4ntyv\xe4t kulut verkkosivujen uudistamisesta kertoivat sis\xe4ll\xf6ntuotannon lis\xe4\xe4ntymisest\xe4. Tuotanto lis\xe4\xe4ntyi Googlen julkistaman ensimm\xe4isen raportin mukaan vuoden 2011 ensimm\xe4isest\xe4 nelj\xe4nneksest\xe4 vuoden 2013 ensimm\xe4iseen nelj\xe4nnekseen verrattuna 10 prosentilla, ja sis\xe4ll\xf6ntuotannon kustannukset kasvoivat samana aikana 48 prosenttia.\n\nGooglen taloudellisen kirjanpidon lukujen perusteella Google+ sai my\xf6s viime vuonna entist\xe4 korkeammat resurssit, mutta kulujen kasvattaminen vaikutti j\xe4\xe4dytykseen.\n\nGoogle on jo aiemmin keskitt\xe4nyt ty\xf6voimansa Android-k\xe4ytt\xf6j\xe4rjestelm\xe4\xe4n, ja samankaltainen kehitys voi my\xf6s Google+:n osalta olla tulossa.]" time="0.021"><properties><property name="score" value="0.181563821640629" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.18156382&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.18156382
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[the belief or doctrine that the universe is eternal and has no beginning or end.\n\nthe belief or doctrine that the universe is eternal and has no beginning or end.\n\na worldview, esp. in ancient India, that denies all distinctions between spirit and matter.\n\nthe doctrine that the universe is infinite and without beginning or end.\n\nthe doctrine that the universe is infinite and without beginning or end.\n\na form of pantheism that assumes the presence of god as a force or energy pervading the universe.\n\nthe belief that the universe is eternal and has always existed.\n\nthe belief that the universe is eternal and has always existed.\n\nthe doctrine that there is only one god.\n\na philosophy that stresses the unity and continuity of existence.\n\na philosophy that stresses the unity and continuity of existence.\n\nthe philosophical doctrine that there are no immaterial or transcendent entities, such as God or the soul, but that all phenomena of reality may be accounted for by material means.\n\na philosophy holding that truth is apprehended by intuition and not by reason.\n\nthe doctrine that God is not a person but an impersonal principle that accounts for the universe.\n\nthe doctrine that the universe is not a personal, creative agent but a blind, orderly, insensible machine.\n\nthe philosophical doctrine that the universe is not a personal, creative agent but a blind, orderly, insensible machine.\n\nthe doctrine that natural phenomena have no supernatural explanation.\n\na doctrine that is similar to pantheism but is regarded as more fundamental, asserting the essential unity of the universe.\n\nthe doctrine that God, nature, and humanity are not distinct entities but are one.\n\nthe belief that the universe is an organic unity, with every person and event affecting every other.\n\nthe belief that the universe is an organic unity, with every person and event affecting every other.\n\na worldview positing that all life and existence is inseparable from the interdependent parts of the whole.\n\nthe belief that there is an undying and immaterial part of man, esp. the soul.\n\nthe belief that there is an undying and immaterial part of man, esp. the soul.\n\nthe belief that there is a common and universal religious or moral experience in every person.\n\nthe belief that there is a common and universal religious or moral experience in every person.\n\nthe theory that all existing things in the world are formed from and constitute a unity with one fundamental substance, esp. matter or spirit.\n\nthe theory that all existing things in the world are formed from and constitute a unity with one fundamental substance, esp. matter or spirit.\n\nthe theory that only material things and properties exist.\n\nthe theory that only material things and properties exist.\n\nthe doctrine that the world and all its phenomena are merely the product of chance and necessity.\n\nthe doctrine that the world and all its phenomena are merely the product of chance and necessity.\n\nthe doctrine that man is no more than a natural product of evolution, with no spiritual aspect.\n\nthe doctrine that man is no more than a natural product of evolution, with no spiritual aspect.\n\na system of principles and practices based on the teachings of a spiritual leader.\n\nthe worship of idols or icons.\n\nthe worship of idols or icons.\n\na religious practice consisting of the use of some object as a focus of worship.\n\na religious practice consisting of the use of some object as a focus of worship.\n\nthe belief that ultimate reality cannot be expressed in ordinary language.\n\nthe belief that ultimate reality cannot be expressed in ordinary language.\n\nthe doctrine that all human beings possess a dual nature and can achieve a transcendent existence.\n\nthe doctrine that all human beings possess a dual nature and can achieve a transcendent existence.\n\nthe doctrine that only moral and spiritual values are real.\n\nthe doctrine that only moral and spiritual values are real.\n\na belief in the existence of a soul as an immaterial entity distinct from the body.\n\na belief in the existence of a soul as an immaterial entity distinct from the body.\n\nthe doctrine that knowledge is impossible.\n\nthe doctrine that knowledge is impossible.\n\nthe belief that everything is the result of chance.\n\nthe belief that everything is the result of chance.\n\nthe doctrine that all matter is intrinsically evil and that human existence is a burden imposed upon men by evil spirits.\n\nthe doctrine that all matter is intrinsically evil and that human existence is a burden imposed upon men by evil spirits.\n\nthe doctrine that knowledge is merely a by-product of learning and not a result of the processes of the mind.\n\nthe doctrine that knowledge is merely a by-product of learning and not a result of the processes of the mind.\n\nthe doctrine that God is unknowable and that religious truth can be attained only through intuition and inner experience.\n\nthe doctrine that God is unknowable and that religious truth can be attained only through intuition and inner experience.\n\nthe doctrine that life is an illusion and death is the reality.\n\nthe doctrine that life is an illusion and death is the reality.\n\nthe doctrine that the universe is composed of a single substance, such as matter or spirit, and that all things are modifications of this basic substance.\n\nthe doctrine that the universe is composed of a single substance, such as matter or spirit, and that all things are modifications of this basic substance.\n\nthe belief that all phenomena and events in the world are determined by necessary laws and have their causes in previous events.\n\nthe belief that all phenomena and events in the world are determined by necessary laws and have their causes in previous events.\n\nthe doctrine that matter and its motions are eternal and that God, if he exists, is wholly immanent in the world.\n\nthe doctrine that matter and its motions are eternal and that God, if he exists, is wholly immanent in the world.\n\na group of people sharing religious convictions.\n\na group of people sharing religious convictions.\n\nthe doctrine that the world is an aggregate of self-existent individuals.\n\nthe doctrine that the world is an aggregate of self-existent individuals.\n\na philosophy that identifies religion with a specific moral system.\n\nthe belief that knowledge of an eternal reality is gained through intuition.\n\nthe belief that knowledge of an eternal reality is gained through intuition.\n\nthe belief that a divine or other transcendent power has sovereignty over the universe.\n\nthe belief that a divine or other transcendent power has sovereignty over the universe.\n\na comprehensive system of political, economic, and social organization.\n\nthe practice of worshiping with songs, prayers, and other gestures in addition to rites and ceremonies.\n\nthe practice of worshiping with songs, prayers, and other gestures in addition to rites and ceremonies.\n\nthe doctrine that truth is gained through study and contemplation.\n\nthe doctrine that truth is gained through study and contemplation.\n\nthe theory that all qualities in the human mind are relative to the situation in which they are encountered.\n\nthe theory that all qualities in the human mind are relative to the situation in which they are encountered.\n\nthe belief that the divine spark of reason is in all human beings.\n\nthe belief that the divine spark of reason is in all human beings.\n\n]" time="0.088"><properties><property name="score" value="0.2587967162175529" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[The US Forest Service has announced that the Fern Lake Fire is fully contained and the evacuation advisory for North of Tofte has been lifted. This comes after firefighters spent the last week building up the fire lines around the blaze. It also means that Minnesota State Highway 61 is now open to traffic again, as are Highway 1 and Highway 61 (Forest Roads 18 and 55), which run through the fire area.\n\nAs of Monday morning, there are still seven aircraft and 380 firefighters assigned to the fire, which was estimated to be 70 percent contained. It has burned 6,900 acres and its cause remains under investigation.\n\nThe next update is expected by Friday morning.]" time="0.020"><properties><property name="score" value="0.039090062174146345" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03909006&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03909006
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[In my world it's important to be organized. Especially when you live with more than two people, which I do. Most of the time, everything is on a schedule. No one is allowed to eat at 11pm unless they're hungry. Everyone is not allowed to leave dirty dishes in the sink unless they're at work. If you're working out at a certain time, you have to shower. No one can go outside unless it's to throw out the garbage. And so on and so forth.\n\nFor some reason, though, the whole &quot;organization&quot; idea flies right out the window when I'm trying to schedule a trip. I tend to plan things to the day and then get annoyed if something goes wrong or someone else wants to do something that may interfere. I hate it when someone wants to do something at a time that I've already booked for another thing. I feel like I should have priority, especially since I've planned this whole thing and if it were cancelled, I would have to pay for everything anyway.\n\nBut, guess what? I've realized I'm wrong. I've realized that while I may have planned the whole thing, other people need to be included too. And that they deserve a say in what goes on too. Maybe that means they have to drive back a different route to get home or they need to leave an hour earlier to get to their house. Maybe they need to stay later than they would have planned. Or maybe they don't want to leave. And that's ok. It doesn't have to ruin anything.\n\nAs long as everyone is included and has a say, it's all good. It doesn't mean that there won't be stress involved. There will be. I'm not going to lie. We may end up fighting about some things. Or they might leave early, which makes us upset. But at the end of the day, we're still family. We still love each other. We still get to have fun.\n\nI guess what I'm saying is that sometimes, I need to let other people in. To let them have a say in what happens. Otherwise, it will just be me against them. And I'd rather be the bigger person. I'd rather be the one who says &quot;Fine, I'll do it this way&quot; instead of the one who gets in an argument. I don't want to be the one who stays up late crying over something that went wrong and couldn't be avoided.\n\nSo, to sum up my whole blog:]" time="0.036"><properties><property name="score" value="0.05680842392055515" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05680842&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05680842
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[&quot;Would You Accept A Ridiculously Generous Gift?\n\nThe book, Cults Inside Out, claims to reveal everything you ever wanted to know about cults, from recruiting and controlling techniques to the manipulation of the children of cult members. The book is written by Jeremy Spencer, who claims that his grandfather was a cult leader and that his father was in a cult. It is hoped that by publishing this book that Jeremy Spencer can stop this cycle and prevent any other children from being hurt in this way.\n\nA spokesperson for Jeremy Spencer has confirmed that he is desperately in need of some cash to fund the marketing and promotion of the book. The cost is simply too great to risk doing it alone. He hopes that if he can raise enough money to pay for advertising, then he will be able to share his story and help people.\n\nIf Jeremy Spencer can raise enough money for the book, then he will gift it to the person who donates the most money. The donation can be made via the crowdfunding site Kickstarter.com.\n\nDonating to this crowdfunding campaign is likely to have more benefits than simply receiving a copy of the book. You can be part of the book and even get a speaking part in it. This is a great opportunity to help someone with a great story.&quot;\n\nKickstarter is a crowdfunding website that allows you to donate money to people who are trying to raise money for something. In return for your donations, you can receive gifts from the person you are donating to.\n\nWould You Accept A Ridiculously Generous Gift?\n\nA friend of mine contacted me a few months ago to ask if I would consider reviewing a book for them. The book was called Cults Inside Out and the author is called Jeremy Spencer. He has offered to donate the book to the highest bidder.\n\nCults Inside Out is the first of a trilogy and Jeremy Spencer is trying to raise money to self-publish the second book. As part of his plan, Jeremy is offering to donate the book to the highest bidder. The highest bidder is whoever donates the most money to Jeremy's GoFundMe account.\n\nWould you accept a ridiculously generous gift? I decided that I would.\n\nJeremy Spencer is an interesting character. I had never heard of him before and I was sceptical about some of the things that he had to say. Jeremy is a bit of a controversial character and he is no stranger to the press. He is very controversial and he has a history of being sued for libel, but you can find out more about that on his GoFundMe page.\n\nThe book is about cults and how they use manipulation techniques to recruit people into their organisations. The book also talks about the psychology of the cult members and how they get manipulated. There is some fascinating stuff in there and it is an easy read.\n\nThe book starts with the stories of two children who were born into cults. One child was the son of a cult leader and the other child was the daughter of another cult leader. Both children were exposed to extreme forms of cult abuse and you can really feel the pain that they went through.\n\nI didn't really enjoy reading the stories of the two children because it was so sad. You could really feel their pain. I am glad that Jeremy was able to escape the cult, but I don't think that it is an easy thing to do. There are always going to be repercussions, no matter how many years later.\n\nAfter the stories of the two children, Jeremy goes on to talk about the psychology of the cult members. I didn't realise that there were so many different reasons why people join cults. There are a number of factors that can influence the decision to join a cult.\n\nThere are a number of strategies that cults use to recruit new members and Jeremy goes on to discuss the techniques that they use. He talks about how they use manipulation and persuasion techniques to encourage people to join them. Jeremy is not the first person to talk about these strategies, but he does explain them in a very clear way.\n\nJeremy also talks about the different kinds of cults and what they are like. There are religious cults and political cults and there are even business cults. There are all kinds of cults and it is fascinating to read about how they are structured and what they are all about.\n\nJeremy Spencer has also included information about his family and his time in the cult. I have never been in a cult, so I don't know what it is like, but I was impressed with how well Jeremy handled the subject. He didn't talk about the cult in a negative way and he didn't try to make it sound worse than it was.\n\nThere is a lot of information about cults in this book and Jeremy doesn't try to overstate any of the facts. He doesn't try to talk in hyperbole and he gives very factual descriptions of what it is like to be in a cult.\n\nIf you are a fan of cults, then this book is going to be right up your street. If you are just interested in psychology, then this book is also going to be right up your street. There are a lot of interesting theories and a lot of fascinating facts.\n\nThere are a number of interesting stories and you can really feel the emotion in this book. I am glad that I have read it, but I don't think that I will read the rest of the trilogy.\n\nWhat you will get if you make the highest bid\n\nIf you make the highest bid on Jeremy's GoFundMe page, then he will send you a copy of Cults Inside Out. The money that you donate will be used to help with the marketing and promotion of the book.\n\nI have seen people donate large amounts of money to crowdfunding projects, so it is definitely possible. Jeremy is even offering to sign your copy of the book, if you make the highest bid.\n\nThe book is available in paperback and as an e-book. You will also be able to get it in various different languages. It is likely to be available in a number of different countries.\n\nYou can get a signed copy of the book for just $30 and Jeremy has been very generous with his offers. There are some very generous rewards if you can make the highest bid. Jeremy is determined to get this book out and he will give you the opportunity to help him.\n\nIf you are interested in making the highest bid, then click here. If you are interested in reading more about the book, then click here.\n\nWhat are your thoughts about Jeremy Spencer's generosity? Please leave a comment below.\n\nSave\n\nSave\n\nSave\n\nSave]" time="0.080"><properties><property name="score" value="0.015866325528118945" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl['I'd only ever been with two girls'\n\nMany guys worry about their first time having sex \u2013 especially when it comes to having sex for the first time with a girl. How can you be sure she's comfortable? And what should you do if she says no?\n\nHere's how to handle your first time.\n\nGetting ready for the first time\n\nEven though your first time might be really exciting, you need to remember that having sex for the first time can be really awkward.\n\nThis is especially true if you're nervous, so to make it easier for both of you, try to relax, and make sure you're both feeling comfortable and happy to do it. You don't want to be worrying about whether your partner's enjoying it.\n\nThe best way to do this is to talk about it before you get started. Find out what she expects, and what she doesn't. Try to find out how much she knows about sex. It might seem strange to talk about this, but it will make it easier for you both when you get down to it.\n\nYou can talk about contraception too. Both of you should know how you're going to stop pregnancy happening and what to do if it does.\n\nA lot of guys worry about getting their first time over with, but in fact, the longer you take, the more you'll enjoy it. Most girls feel more comfortable the longer it takes. If you try to rush things, you're more likely to make her feel uncomfortable and she might not enjoy it.\n\nWhat if you're with someone new?\n\nIf you're having sex with someone for the first time, you may be nervous. You're not sure what to expect.\n\nYou're likely to be worried that you won't know how to do it properly or that you'll be rubbish in bed. But the fact is, it doesn't matter how good you are. All that matters is that you both enjoy it and that you're safe.\n\nYour worries might make you want to rush things and this might make you less sensitive and not as gentle as you should be. This is especially true if you're with a virgin. She'll be more nervous than you, so make sure you take your time and that you both enjoy it.\n\nYou'll know if she's enjoying it if she shows signs of pleasure, like:\n\nmoaning\n\ngetting very wet (this is the most obvious one!)\n\nshe may make noises or say things to let you know she's enjoying it\n\nremember to look for these signs!\n\nFirst time with a girl: the sex\n\nSex for the first time can be really exciting.\n\nWhat if she says no?\n\nIf your first time involves having sex with a girl, there's a chance she might say no. That's normal. It's important to make sure she's happy to do it. It's your responsibility to make sure she's okay with it and you don't want to do anything that might hurt her or make her feel uncomfortable.\n\nEven if she says yes, she might change her mind when the time comes. This can be a little awkward, especially if you've been going out with her for a while and you've had sex before.\n\nIf this happens, try to deal with it like a man. If you want to take it further, tell her you really want to, but you won't if she's not up for it.\n\nIf she wants to stop, you should respect that. It can happen, even if she's wanted to have sex for a while. If you think she might be pregnant, make sure you take care of that too.\n\nThe actual sex\n\nIt's important that you know how to please a girl, but this is especially important the first time. It will help you build your confidence, and it will also help her enjoy sex more.\n\nHow do you please a girl?\n\nWhat most people think about when it comes to pleasing a girl is what feels good for her. But in fact, what feels good for you is important too. It's the only way you can both enjoy yourselves and have a great time.\n\nBe careful to go slowly. If you go too fast, you might scare her or make her uncomfortable.\n\nWhat if you're not having sex with a virgin?\n\nIf you're not having sex with a virgin, it's unlikely that you'll have to worry about losing your virginity. It's more likely that she will lose hers!\n\nWhen you have sex for the first time, it's still important that you take it slow. It might be even more important because you'll need to show her that you can be gentle and caring. This will make sure she feels comfortable with you. It will also make sure that you have a great time too.\n\nIs she ready for sex?\n\nYou need to know when she's ready to have sex. If you rush it, you could end up doing something that she's not ready for. This could make her feel embarrassed or uncomfortable, which could make you feel embarrassed too.\n\nIf you don't think she's ready, you shouldn't go ahead. That's because it might make her feel like she's not good enough.\n\nJust be patient and make sure you're both happy and comfortable.\n\nYou can help her decide when she's ready for sex by asking her. Ask her what she thinks about having sex. If she says no, respect that, and don't push it.\n\nIf she says yes, talk about what you both think. Ask her if she's ready for it and talk about what you both expect. You'll both feel more relaxed if you know what's going to happen and you're both ready for it.\n\nDon't forget, it's important that you're both ready, not just her!\n\nWhat if you're with a virgin?\n\nSex for the first time can be really scary. But remember, you're both in the same boat and you can help each other through it.\n\nIf you're with a virgin, you need to remember that she's not an expert. It will take her a while to get used to sex and you need to give her time. It will help if you make sure you go slowly.\n\nAs you know, she'll be nervous, so you need to make sure she feels comfortable.\n\nSome girls are really confident and they're great at giving directions. But other girls aren't, so if she needs help, you should be there to help her.\n\nHow do you make sure she's ready?\n\nYou might have to talk about it. You might feel like she's pushing you into having sex, but she might not actually be ready for it.\n\nIf you don't feel comfortable talking to her, you can find out by looking at her. Does she seem relaxed? Is she making the right noises? Is she happy? If she seems happy, you're probably okay.\n\nMake sure you ask her about contraception too. It's important to keep yourself safe, so make sure you both know how you're going to prevent pregnancy.\n\nWhat if you have problems?\n\nSex for the first time can be really scary for some guys. But the good news is that it doesn't have to be.\n\nWhat if you're having trouble getting an erection?\n\nIf you're having trouble getting an erection, it could be that you're nervous. It could also be that you don't feel ready for sex.\n\nIf it's nerves, you should try to relax. Try to take it slowly. You don't want to scare your partner. If you're not ready for sex, then you need to tell her.\n\nIt might help to talk to her about it first. If you think you're ready, but she isn't, you should respect that. You don't want to make her feel bad or make her regret having sex with you.\n\nIt's important that you're both ready for it and that you both feel comfortable. If you're not ready, don't go through with it.\n\nWhat if it doesn't feel right?\n\nSome guys worry about whether they're doing it right. It can be a bit scary if you don't know what to do.\n\nIt's really important that you talk about it. If you think it's not right, you should ask her.\n\nAnd don't worry too much. There's no right way to have sex. The important thing is that both of you are enjoying it and that you're both safe.\n\nIt's really important that you both make sure you're ready for sex. You don't want to go through with it if she's not ready.\n\nIf you're having sex with a virgin, then you need to be more careful. She might not know what she's doing, so it's important that you tell her what you want. You should also tell her if it doesn't feel right.\n\nWhat if it hurts?\n\nIf you're having sex for the first time, there's a chance that it could hurt. But the good news is that this should only happen the first time. If it happens again, you should get help as soon as possible.\n\nYour partner should tell you if it hurts, so you need to listen to her. You should also be able to tell if it hurts. If you can't, then it's not right.\n\nIf it hurts her, then you need to stop and think about what you're doing. It's important that you both know what feels good and what doesn't.\n\nIf you're not sure what to do, then it might be best to practice with a toy. You can use it to find out what she likes and what she doesn't.\n\n]" time="0.096"><properties><property name="score" value="0.006374213274875792" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Public Bookmark *\n\nJames \u201cBucky\u201d Barnes was born in the dead of night, right at the edge of winter and the approaching storm. At least, that\u2019s how his mother likes to tell it. Her mother, a woman who had been born with the gift of sight, had foretold that James would be born during a storm, that his destiny would be caught up in a whirlwind, that he would live a life surrounded by shadows. And in the shadow of the war, there were few people more surrounded by them.\n\nHe lived in the shadow of Steve Rogers and it was only in the wake of his death that James could finally step out from it, could step out of the shadows.\n\nIt is a slow process, this transition. He tries to move forward, but it is difficult.\n\nSome habits are harder to break than others.]" time="0.016"><properties><property name="score" value="0.03991082078816932" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03991082&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03991082
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[This paper extends the framework of the intertemporal Tobit model in which explanatory variables are available only at the end of the sample to the multiple dependent variable case. The effects of endogeneity of these explanatory variables on the standard error of the estimated structural coefficients are quantified. A large endogeneity bias, whose size is not affected by the lag length, is found for all sets of variables, except the set with a positive effect on the last observed outcome, which is relatively immune to endogeneity bias. These results are used to analyze the effects of unobserved variables on the estimates of the true effects and on the standard errors of the estimated effects of observed variables. The impact of unobserved variables on the estimates is sometimes substantial.]" time="0.014"><properties><property name="score" value="0.04886765258917958" /></properties></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Post by Liz \xbb Fri Jan 03, 2018 10:08 am\n\nIt was a tough morning for the county road department this morning with all the road closures due to high water and downed trees. Public Works reports it was a busy day with well over 50 customers reporting issues with their service. The county was very busy all day responding to reported issues, working to clear downed trees, and repairing road damage.\n\nBelow are some pictures from this morning of some of the damage caused by this storm.\n\n\n\nOne of the downed trees on Lakeshore Road.\n\nPictured here is the removal of a large tree on Lakeshore Road.\n\nAnother downed tree on Lakeshore Road\n\nMore trees down on Lakeshore Road\n\nDebris on Lakeshore Road.\n\nTree down on Lewis Street.]" time="0.014"><properties><property name="score" value="0.002653797137058822" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0026538&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.0026538
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[Gary, I don't know who or what the new &quot;Oldsmobile&quot; will be. At this point I have not even seen the car. I think they are planning to unveil it in the spring.\n\n\n\nWhat I would like to see would be an Olds branded CUV with the attitude of the current Enclave. If they can do that, I would buy one. If not, they will be back at square one in the CUV market.\n\n\n\nThe other big issue is the financial status of GM. Is it just teetering on the edge, or has it already fallen into the abyss? If they go bankrupt, will the Olds brand survive? There are a lot of question marks surrounding the new Olds.]" time="0.015"><properties><property name="score" value="0.05680066803811501" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05680067&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05680067
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[North Fork is a great neighborhood for enjoying Long Island. It is right next to the Merrick and Seaford-Oyster Bay Expressways. You can head south to get to the beaches of Long Island, east for the Hamptons, or head west to enjoy all the city has to offer. This is a great location if you want to be able to spend time in the city or on the island, or both!\n\n\n\n\n\n\n\n]" time="0.014"><properties><property name="score" value="0.03059135380238387" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03059135&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.03059135
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_gpt3_jsonl[1. Recognition of the conflict by both sides: the de facto government and the organizations that support the peace process\n\nThis stage is important as it identifies the dynamics of the process and the relationship between the parties. If the process is not recognized by one of the parties, the process will be difficult and may not be sustainable.\n\n2. Opening channels: the expression of mutual understanding, for example through exchange of prisoners\n\n3. A period of dialogue: the exchange of prisoners or groups of prisoners in advance of dialogue, by agreement\n\n4. Dialogue: the willingness to talk\n\n5. Agreement: an agreement to implement a cease-fire, a political dialogue or to resolve any other issue\n\n6. Realization of an agreement: actual cessation of hostilities, the release of prisoners or a commitment to implement political reforms\n\n7. Stabilization: a return to normalcy\n\n8. The final stage: the long-term achievement of stability in which peace agreements are sustained.]" time="0.015"><properties><property name="score" value="0.011004464261371583" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i0]" time="0.028"><properties><property name="score" value="0.07459978036966051" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i1]" time="0.026"><properties><property name="score" value="0.07208471500038108" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i2]" time="0.017"><properties><property name="score" value="0.09370291310977219" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i3]" time="0.013"><properties><property name="score" value="0.029100712189922273" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i4]" time="0.025"><properties><property name="score" value="0.03754008407226027" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i5]" time="0.041"><properties><property name="score" value="0.055656761220208475" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i6]" time="0.031"><properties><property name="score" value="0.0942542493220258" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i7]" time="0.026"><properties><property name="score" value="0.06348605860240833" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i8]" time="0.026"><properties><property name="score" value="0.08304578580666408" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i9]" time="0.028"><properties><property name="score" value="0.01607919683533094" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.0160792&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.0160792
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i10]" time="0.015"><properties><property name="score" value="0.04140746664123007" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i11]" time="0.026"><properties><property name="score" value="0.06350385562562973" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i12]" time="0.015"><properties><property name="score" value="0.048072729947890025" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.04807273&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.04807273
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i13]" time="0.025"><properties><property name="score" value="0.04847196806158083" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i14]" time="0.012"><properties><property name="score" value="0.03811222266512093" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i15]" time="0.025"><properties><property name="score" value="0.10027660854823361" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i16]" time="0.024"><properties><property name="score" value="0.10586091625639901" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i17]" time="0.025"><properties><property name="score" value="0.05035323154371503" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i18]" time="0.026"><properties><property name="score" value="0.070640602440317" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i19]" time="0.027"><properties><property name="score" value="0.12020357904066303" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i20]" time="0.016"><properties><property name="score" value="0.05142703285571648" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i21]" time="0.025"><properties><property name="score" value="0.051982193787271175" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i22]" time="0.015"><properties><property name="score" value="0.08479035245128608" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i23]" time="0.025"><properties><property name="score" value="0.12950790756137354" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i24]" time="0.013"><properties><property name="score" value="0.042085609923192546" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i25]" time="0.028"><properties><property name="score" value="0.05127841587393869" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i26]" time="0.029"><properties><property name="score" value="0.057635271590979684" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i27]" time="0.026"><properties><property name="score" value="0.11126872462680122" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i28]" time="0.027"><properties><property name="score" value="0.058151301769146124" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.0581513&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.0581513
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i29]" time="0.015"><properties><property name="score" value="0.10259988124730057" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i30]" time="0.031"><properties><property name="score" value="0.0773459521284209" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i31]" time="0.016"><properties><property name="score" value="0.050899041123048105" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i32]" time="0.027"><properties><property name="score" value="0.09563490052501525" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i33]" time="0.086"><properties><property name="score" value="0.025151015706934943" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i34]" time="0.034"><properties><property name="score" value="0.11926102098362196" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i35]" time="0.016"><properties><property name="score" value="0.14347116369923807" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i36]" time="0.026"><properties><property name="score" value="0.0663485343055955" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i37]" time="0.031"><properties><property name="score" value="0.061820770482507614" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i38]" time="0.030"><properties><property name="score" value="0.010161813257591731" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i39]" time="0.032"><properties><property name="score" value="0.06751060242996579" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i40]" time="0.028"><properties><property name="score" value="0.06616433943800218" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i41]" time="0.014"><properties><property name="score" value="0.036452186173224765" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i42]" time="0.024"><properties><property name="score" value="0.10459172827663232" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i43]" time="0.027"><properties><property name="score" value="0.06837707509528561" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i44]" time="0.027"><properties><property name="score" value="0.07493619712978417" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i45]" time="0.026"><properties><property name="score" value="0.08488278149161402" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i46]" time="0.025"><properties><property name="score" value="0.003432957555599864" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.00343296&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.00343296
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i47]" time="0.027"><properties><property name="score" value="0.07478170836810936" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i48]" time="0.024"><properties><property name="score" value="0.04756074521686643" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i49]" time="0.025"><properties><property name="score" value="0.08669304459461656" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i50]" time="0.031"><properties><property name="score" value="0.06873733311018715" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i51]" time="0.027"><properties><property name="score" value="0.06324147334961336" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i52]" time="0.025"><properties><property name="score" value="0.0666549878706085" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i53]" time="0.023"><properties><property name="score" value="0.07048643037683403" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i54]" time="0.024"><properties><property name="score" value="0.11359484874714876" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i55]" time="0.081"><properties><property name="score" value="0.09175735486527349" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i56]" time="0.043"><properties><property name="score" value="0.09472221849902454" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i57]" time="0.019"><properties><property name="score" value="0.07905330642885566" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i58]" time="0.037"><properties><property name="score" value="0.06071184971470128" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i59]" time="0.030"><properties><property name="score" value="0.035977869346690027" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i60]" time="0.014"><properties><property name="score" value="0.04219086681575002" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i61]" time="0.036"><properties><property name="score" value="0.1347182784660036" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i62]" time="0.032"><properties><property name="score" value="0.05049886599618869" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i63]" time="0.027"><properties><property name="score" value="0.04241581708447806" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.04241582&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.04241582
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i64]" time="0.027"><properties><property name="score" value="0.09498109357761443" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i65]" time="0.027"><properties><property name="score" value="0.03056762503220303" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i66]" time="0.025"><properties><property name="score" value="0.05554436480680554" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i67]" time="0.025"><properties><property name="score" value="0.08250599019784666" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i68]" time="0.026"><properties><property name="score" value="0.07991330258161904" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i69]" time="0.028"><properties><property name="score" value="0.07333891092245062" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i70]" time="0.029"><properties><property name="score" value="0.08632330967923486" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i71]" time="0.029"><properties><property name="score" value="0.0634605769624194" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i72]" time="0.026"><properties><property name="score" value="0.06533317017980825" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i73]" time="0.025"><properties><property name="score" value="0.07639203400797699" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i74]" time="0.025"><properties><property name="score" value="0.044818249008168796" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i75]" time="0.025"><properties><property name="score" value="0.03869753072442188" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i76]" time="0.071"><properties><property name="score" value="0.056635719262508495" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i77]" time="0.026"><properties><property name="score" value="0.09958147658015284" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i78]" time="0.029"><properties><property name="score" value="0.07021318220808491" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i79]" time="0.027"><properties><property name="score" value="0.07587869720548346" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i80]" time="0.027"><properties><property name="score" value="0.0669121256882386" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i81]" time="0.026"><properties><property name="score" value="0.009613300168125831" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.0096133&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.0096133
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i82]" time="0.013"><properties><property name="score" value="0.055461547552621004" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i83]" time="0.013"><properties><property name="score" value="0.046680757029615716" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i84]" time="0.024"><properties><property name="score" value="0.12610787059521658" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i85]" time="0.025"><properties><property name="score" value="0.15031902261795427" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i86]" time="0.024"><properties><property name="score" value="0.004318124306796076" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.00431812&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.00431812
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i87]" time="0.013"><properties><property name="score" value="0.01843328242064901" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i88]" time="0.025"><properties><property name="score" value="0.05561462106022007" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.05561462&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.05561462
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i89]" time="0.023"><properties><property name="score" value="0.1081174046297574" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i90]" time="0.024"><properties><property name="score" value="0.0971956107046148" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i91]" time="0.024"><properties><property name="score" value="0.06258943576962887" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i92]" time="0.023"><properties><property name="score" value="0.008761616707996689" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.00876162&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.00876162
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i93]" time="0.025"><properties><property name="score" value="0.04950025856452099" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i94]" time="0.024"><properties><property name="score" value="0.09528059209571549" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i95]" time="0.027"><properties><property name="score" value="0.09307414929463714" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i96]" time="0.013"><properties><property name="score" value="0.05475143033691343" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i97]" time="0.023"><properties><property name="score" value="0.061680842370363316" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i98]" time="0.026"><properties><property name="score" value="0.07783972344157537" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i99]" time="0.025"><properties><property name="score" value="0.08700957226246649" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i100]" time="0.012"><properties><property name="score" value="0.027702891083314363" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i101]" time="0.024"><properties><property name="score" value="0.08323791654009373" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i102]" time="0.025"><properties><property name="score" value="0.07016814259782622" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i103]" time="0.024"><properties><property name="score" value="0.057120190010046416" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i104]" time="0.023"><properties><property name="score" value="0.10973533978337345" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i105]" time="0.024"><properties><property name="score" value="0.1046619361952388" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i106]" time="0.025"><properties><property name="score" value="0.019958202032611127" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i107]" time="0.023"><properties><property name="score" value="0.04744331566673393" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i108]" time="0.023"><properties><property name="score" value="0.12006991018471401" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i109]" time="0.014"><properties><property name="score" value="0.05903562710364296" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i110]" time="0.024"><properties><property name="score" value="0.02301442704543879" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i111]" time="0.024"><properties><property name="score" value="0.10987440996927388" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i112]" time="0.024"><properties><property name="score" value="0.09108156113759697" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i113]" time="0.027"><properties><property name="score" value="0.07267098807439101" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i114]" time="0.013"><properties><property name="score" value="0.04179323445700489" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i115]" time="0.024"><properties><property name="score" value="0.09057904159139336" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i116]" time="0.012"><properties><property name="score" value="0.10987591206398362" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i117]" time="0.024"><properties><property name="score" value="0.027052223163054643" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i118]" time="0.022"><properties><property name="score" value="0.10095917957497515" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i119]" time="0.022"><properties><property name="score" value="0.1454056614790181" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i120]" time="0.022"><properties><property name="score" value="0.07163208548504227" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i121]" time="0.022"><properties><property name="score" value="0.05146591871995905" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i122]" time="0.012"><properties><property name="score" value="0.06925480487450231" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i123]" time="0.022"><properties><property name="score" value="0.05234151799213404" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i124]" time="0.022"><properties><property name="score" value="0.06575137584925494" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i125]" time="0.022"><properties><property name="score" value="0.06573515834822002" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i126]" time="0.023"><properties><property name="score" value="0.07503645397552357" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i127]" time="0.024"><properties><property name="score" value="0.029694250956742895" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i128]" time="0.022"><properties><property name="score" value="0.07796756962819873" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i129]" time="0.024"><properties><property name="score" value="0.15043626578205233" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i130]" time="0.013"><properties><property name="score" value="0.0844482881677211" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i131]" time="0.023"><properties><property name="score" value="0.06546330510136122" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i132]" time="0.022"><properties><property name="score" value="0.11994142705745547" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i133]" time="0.023"><properties><property name="score" value="0.11293218965437324" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i134]" time="0.022"><properties><property name="score" value="0.07019069889233169" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i135]" time="0.012"><properties><property name="score" value="0.07047950439466044" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i136]" time="0.022"><properties><property name="score" value="0.02493804101550728" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i137]" time="0.011"><properties><property name="score" value="0.07382210561589497" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i138]" time="0.011"><properties><property name="score" value="0.03642900637821911" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i139]" time="0.022"><properties><property name="score" value="0.0687975871914883" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i140]" time="0.026"><properties><property name="score" value="0.07140607207581431" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i141]" time="0.022"><properties><property name="score" value="0.0753551110418306" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i142]" time="0.022"><properties><property name="score" value="0.015125218442238325" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i143]" time="0.023"><properties><property name="score" value="0.032453120255815426" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i144]" time="0.021"><properties><property name="score" value="0.08261163575509711" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i145]" time="0.022"><properties><property name="score" value="0.014996694311525666" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i146]" time="0.022"><properties><property name="score" value="0.046072100457340656" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i147]" time="0.022"><properties><property name="score" value="0.08521829795703063" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i148]" time="0.021"><properties><property name="score" value="0.04971684169496692" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i149]" time="0.023"><properties><property name="score" value="0.05646446677268986" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i150]" time="0.024"><properties><property name="score" value="0.051681024946537514" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i151]" time="0.021"><properties><property name="score" value="0.09350868059740407" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i152]" time="0.021"><properties><property name="score" value="0.009024495537510297" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i153]" time="0.020"><properties><property name="score" value="0.11076752593595367" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i154]" time="0.020"><properties><property name="score" value="0.10252703343351799" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i155]" time="0.021"><properties><property name="score" value="0.06786418720526753" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i156]" time="0.020"><properties><property name="score" value="0.06407468968474773" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i157]" time="0.022"><properties><property name="score" value="0.04886559544048348" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i158]" time="0.011"><properties><property name="score" value="0.1757565193468258" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i159]" time="0.020"><properties><property name="score" value="0.02803815998511423" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i160]" time="0.021"><properties><property name="score" value="0.10643111675756192" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i161]" time="0.010"><properties><property name="score" value="0.06614645112693607" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i162]" time="0.021"><properties><property name="score" value="0.030250539231662994" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.03025054&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.03025054
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i163]" time="0.011"><properties><property name="score" value="0.034111902120154136" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i164]" time="0.012"><properties><property name="score" value="0.028580070869321794" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i165]" time="0.020"><properties><property name="score" value="0.05204482675600519" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i166]" time="0.021"><properties><property name="score" value="0.06668898276969565" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i167]" time="0.021"><properties><property name="score" value="0.06470271329403858" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i168]" time="0.011"><properties><property name="score" value="0.09530749368652502" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i169]" time="0.011"><properties><property name="score" value="0.12664162770360532" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i170]" time="0.022"><properties><property name="score" value="0.0754365747102434" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i171]" time="0.022"><properties><property name="score" value="0.10512381976061103" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i172]" time="0.020"><properties><property name="score" value="0.017529408156943087" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i173]" time="0.020"><properties><property name="score" value="0.060999070590817314" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i174]" time="0.023"><properties><property name="score" value="0.02447017834629775" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.02447018&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.02447018
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i175]" time="0.058"><properties><property name="score" value="0.0845023188820373" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i176]" time="0.021"><properties><property name="score" value="0.11937370241853949" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i177]" time="0.011"><properties><property name="score" value="0.12736979449146646" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i178]" time="0.022"><properties><property name="score" value="0.07299564239744269" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i179]" time="0.021"><properties><property name="score" value="0.0929199260337315" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i180]" time="0.020"><properties><property name="score" value="0.08976560553841319" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i181]" time="0.021"><properties><property name="score" value="0.10888106103595066" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i182]" time="0.021"><properties><property name="score" value="0.07078871612104667" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.07078872&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.07078872
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i183]" time="0.063"><properties><property name="score" value="0.144089159375198" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i184]" time="0.021"><properties><property name="score" value="0.03193865121673445" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i185]" time="0.012"><properties><property name="score" value="0.054700150718767215" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i186]" time="0.021"><properties><property name="score" value="0.11000585132868278" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i187]" time="0.064"><properties><property name="score" value="0.04579181079621878" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i188]" time="0.021"><properties><property name="score" value="0.07102584266552868" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i189]" time="0.024"><properties><property name="score" value="0.08829692621575747" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.08829693&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.08829693
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i190]" time="0.011"><properties><property name="score" value="0.07604802528917709" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i191]" time="0.022"><properties><property name="score" value="0.08538573232484228" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i192]" time="0.022"><properties><property name="score" value="0.019782058558357837" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.01978206&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.01978206
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i193]" time="0.020"><properties><property name="score" value="0.10251736282484641" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i194]" time="0.021"><properties><property name="score" value="0.07747609139837762" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i195]" time="0.021"><properties><property name="score" value="0.008122575590825005" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.00812258&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.00812258
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i196]" time="0.021"><properties><property name="score" value="0.029216748850631268" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i197]" time="0.011"><properties><property name="score" value="0.037652818077837935" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i198]" time="0.021"><properties><property name="score" value="0.09667725105418429" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i199]" time="0.024"><properties><property name="score" value="0.035267493645902155" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i200]" time="0.011"><properties><property name="score" value="0.05815783693132737" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i201]" time="0.016"><properties><property name="score" value="0.03387632743913627" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i202]" time="0.022"><properties><property name="score" value="0.08223442454310316" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i203]" time="0.020"><properties><property name="score" value="0.12433105524729737" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i204]" time="0.021"><properties><property name="score" value="0.03171914289343318" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i205]" time="0.021"><properties><property name="score" value="0.05810673087889269" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i206]" time="0.021"><properties><property name="score" value="0.08861001611827593" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i207]" time="0.021"><properties><property name="score" value="0.07065997234514212" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i208]" time="0.020"><properties><property name="score" value="0.025556285687753953" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i209]" time="0.011"><properties><property name="score" value="0.0494915853698652" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i210]" time="0.019"><properties><property name="score" value="0.0971012138419254" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i211]" time="0.020"><properties><property name="score" value="0.06919342707273887" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i212]" time="0.010"><properties><property name="score" value="0.11039590604023375" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i213]" time="0.021"><properties><property name="score" value="0.10646451015590774" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i214]" time="0.021"><properties><property name="score" value="0.1364416138196025" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i215]" time="0.020"><properties><property name="score" value="0.07641750323764478" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i216]" time="0.012"><properties><property name="score" value="0.09058108287518518" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i217]" time="0.020"><properties><property name="score" value="0.08853439217341971" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i218]" time="0.021"><properties><property name="score" value="0.10355703781772001" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i219]" time="0.023"><properties><property name="score" value="0.05430092413461067" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i220]" time="0.020"><properties><property name="score" value="0.07031206527695516" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i221]" time="0.021"><properties><property name="score" value="0.0965782121517697" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i222]" time="0.010"><properties><property name="score" value="0.0848094368254726" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i223]" time="0.020"><properties><property name="score" value="0.021646597940221626" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i224]" time="0.020"><properties><property name="score" value="0.10942228143533712" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i225]" time="0.020"><properties><property name="score" value="0.02875243999364775" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i226]" time="0.020"><properties><property name="score" value="0.05977700691114283" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.05977701&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.05977701
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i227]" time="0.019"><properties><property name="score" value="0.11990330419409168" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i228]" time="0.025"><properties><property name="score" value="0.10261605858287057" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i229]" time="0.021"><properties><property name="score" value="0.09390020823636754" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i230]" time="0.022"><properties><property name="score" value="0.07980057342787406" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i231]" time="0.021"><properties><property name="score" value="0.04419531695268188" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.04419532&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.04419532
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i232]" time="0.015"><properties><property name="score" value="0.04056680564663018" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i233]" time="0.022"><properties><property name="score" value="0.06371182642743614" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i234]" time="0.023"><properties><property name="score" value="0.03485764987350025" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i235]" time="0.022"><properties><property name="score" value="0.12341203001112158" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i236]" time="0.011"><properties><property name="score" value="0.14997053743795452" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i237]" time="0.021"><properties><property name="score" value="0.026769510975191972" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i238]" time="0.011"><properties><property name="score" value="0.1031505007092981" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i239]" time="0.023"><properties><property name="score" value="0.08066460597683461" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i240]" time="0.061"><properties><property name="score" value="0.08002868241709583" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i241]" time="0.023"><properties><property name="score" value="0.10106193448721522" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i242]" time="0.012"><properties><property name="score" value="0.04692754578053404" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i243]" time="0.025"><properties><property name="score" value="0.07158454857630943" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i244]" time="0.022"><properties><property name="score" value="0.01137940387807042" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i245]" time="0.022"><properties><property name="score" value="0.08596855460829309" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i246]" time="0.023"><properties><property name="score" value="0.05769012723334414" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i247]" time="0.023"><properties><property name="score" value="0.02701230246395103" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i248]" time="0.023"><properties><property name="score" value="0.08369609169506918" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i249]" time="0.013"><properties><property name="score" value="0.024126425624382275" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i250]" time="0.012"><properties><property name="score" value="0.11675115576548212" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i251]" time="0.025"><properties><property name="score" value="0.03142919608341422" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i252]" time="0.023"><properties><property name="score" value="0.05738167283050601" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i253]" time="0.021"><properties><property name="score" value="0.01913715738023434" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i254]" time="0.015"><properties><property name="score" value="0.11505600499963076" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i255]" time="0.012"><properties><property name="score" value="0.04337488408067691" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i256]" time="0.023"><properties><property name="score" value="0.022796971808880177" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.02279697&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.02279697
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i257]" time="0.012"><properties><property name="score" value="0.05864751400150858" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i258]" time="0.022"><properties><property name="score" value="0.06981220613303754" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i259]" time="0.020"><properties><property name="score" value="0.056143510512474126" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i260]" time="0.021"><properties><property name="score" value="0.03332037750357356" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i261]" time="0.021"><properties><property name="score" value="0.03984681764700133" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i262]" time="0.020"><properties><property name="score" value="0.06368526435843436" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i263]" time="0.021"><properties><property name="score" value="0.04083197592686586" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i264]" time="0.021"><properties><property name="score" value="0.017742905393941525" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i265]" time="0.021"><properties><property name="score" value="0.02724667564057287" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i266]" time="0.023"><properties><property name="score" value="0.0897104452998638" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i267]" time="0.012"><properties><property name="score" value="0.05969647848931836" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i268]" time="0.025"><properties><property name="score" value="0.16607687120285208" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i269]" time="0.021"><properties><property name="score" value="0.05912470343937473" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i270]" time="0.064"><properties><property name="score" value="0.1145364786444098" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i271]" time="0.024"><properties><property name="score" value="0.13161927515791866" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i272]" time="0.020"><properties><property name="score" value="0.0846295572475908" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i273]" time="0.012"><properties><property name="score" value="0.10697394743228172" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i274]" time="0.020"><properties><property name="score" value="0.07821926470308212" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i275]" time="0.021"><properties><property name="score" value="0.06379039779256616" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i276]" time="0.021"><properties><property name="score" value="0.09963480601013038" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i277]" time="0.020"><properties><property name="score" value="0.07669177976539826" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i278]" time="0.011"><properties><property name="score" value="0.08058334319008265" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i279]" time="0.011"><properties><property name="score" value="0.12278726523054506" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i280]" time="0.020"><properties><property name="score" value="0.07909802961246283" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i281]" time="0.020"><properties><property name="score" value="0.059937473418152455" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i282]" time="0.021"><properties><property name="score" value="0.07503098283996418" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i283]" time="0.020"><properties><property name="score" value="0.05849409642427113" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i284]" time="0.022"><properties><property name="score" value="0.0070052934352059035" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.00700529&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.00700529
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i285]" time="0.025"><properties><property name="score" value="0.09365274627973666" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i286]" time="0.032"><properties><property name="score" value="0.10573078259244836" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i287]" time="0.015"><properties><property name="score" value="0.08624206232957787" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i288]" time="0.021"><properties><property name="score" value="0.11712379222528693" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i289]" time="0.021"><properties><property name="score" value="0.07330129852794302" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i290]" time="0.010"><properties><property name="score" value="0.09377485294930765" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i291]" time="0.022"><properties><property name="score" value="0.02857088570644962" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i292]" time="0.020"><properties><property name="score" value="0.060252751756308995" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i293]" time="0.020"><properties><property name="score" value="0.06287249391502245" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i294]" time="0.020"><properties><property name="score" value="0.041451674833142965" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i295]" time="0.021"><properties><property name="score" value="0.013868672149975905" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.01386867&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.01386867
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i296]" time="0.022"><properties><property name="score" value="0.03915885028413628" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i297]" time="0.014"><properties><property name="score" value="0.051352367741586136" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i298]" time="0.024"><properties><property name="score" value="0.03600683558862511" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.03600684&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.03600684
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i299]" time="0.020"><properties><property name="score" value="0.08738784178232284" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i300]" time="0.019"><properties><property name="score" value="0.04663590294625237" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i301]" time="0.020"><properties><property name="score" value="0.10066337639680811" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i302]" time="0.021"><properties><property name="score" value="0.029800468916657115" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i303]" time="0.019"><properties><property name="score" value="0.06773633471651785" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i304]" time="0.020"><properties><property name="score" value="0.0937698191773961" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i305]" time="0.020"><properties><property name="score" value="0.08793793548446617" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i306]" time="0.019"><properties><property name="score" value="0.03877482154124273" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i307]" time="0.020"><properties><property name="score" value="0.0542670285850243" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i308]" time="0.019"><properties><property name="score" value="0.05222254505669668" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i309]" time="0.020"><properties><property name="score" value="0.09828745903735034" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i310]" time="0.020"><properties><property name="score" value="0.07642446500974227" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i311]" time="0.019"><properties><property name="score" value="0.1028625269108236" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i312]" time="0.020"><properties><property name="score" value="0.17056110773679822" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i313]" time="0.019"><properties><property name="score" value="0.09465254929947328" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i314]" time="0.019"><properties><property name="score" value="0.10859029141946878" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i315]" time="0.019"><properties><property name="score" value="0.05824947737321317" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i316]" time="0.019"><properties><property name="score" value="0.09507824537017671" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i317]" time="0.019"><properties><property name="score" value="0.01783247469443616" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i318]" time="0.010"><properties><property name="score" value="0.042276558449338975" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i319]" time="0.020"><properties><property name="score" value="0.09549445315408028" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i320]" time="0.021"><properties><property name="score" value="0.1092294137660661" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i321]" time="0.020"><properties><property name="score" value="0.0028280969732771537" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.0028281&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.0028281
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i322]" time="0.014"><properties><property name="score" value="0.10779454055648285" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i323]" time="0.020"><properties><property name="score" value="0.06698397757640712" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i324]" time="0.020"><properties><property name="score" value="0.11380726776039145" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i325]" time="0.019"><properties><property name="score" value="0.07215659036085287" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i326]" time="0.020"><properties><property name="score" value="0.04442845666886052" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i327]" time="0.020"><properties><property name="score" value="0.04270926769326431" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i328]" time="0.021"><properties><property name="score" value="0.06837383547505288" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i329]" time="0.020"><properties><property name="score" value="0.08537355046574728" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i330]" time="0.020"><properties><property name="score" value="0.06445465193177409" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i331]" time="0.019"><properties><property name="score" value="0.016109856808520295" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i332]" time="0.020"><properties><property name="score" value="0.07319470961405997" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i333]" time="0.024"><properties><property name="score" value="0.056991042590971784" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i334]" time="0.019"><properties><property name="score" value="0.09460706979141502" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i335]" time="0.021"><properties><property name="score" value="0.09087074999185231" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i336]" time="0.019"><properties><property name="score" value="0.0928301706981719" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i337]" time="0.020"><properties><property name="score" value="0.01846284847424143" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i338]" time="0.020"><properties><property name="score" value="0.08828125773089321" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i339]" time="0.019"><properties><property name="score" value="0.03311573381909276" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i340]" time="0.020"><properties><property name="score" value="0.09766328589822248" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i341]" time="0.011"><properties><property name="score" value="0.09788312045044822" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i342]" time="0.021"><properties><property name="score" value="0.06195192166346161" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i343]" time="0.026"><properties><property name="score" value="0.11801361742808131" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i344]" time="0.021"><properties><property name="score" value="0.08113633653770504" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i345]" time="0.021"><properties><property name="score" value="0.04319336057842338" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i346]" time="0.016"><properties><property name="score" value="0.06432606435980426" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i347]" time="0.021"><properties><property name="score" value="0.05558836333595613" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i348]" time="0.025"><properties><property name="score" value="0.10278179137318078" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i349]" time="0.035"><properties><property name="score" value="0.0670446847153221" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i350]" time="0.055"><properties><property name="score" value="0.08383003084137564" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i351]" time="0.035"><properties><property name="score" value="0.05034715362126896" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i352]" time="0.020"><properties><property name="score" value="0.09112308416976966" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i353]" time="0.036"><properties><property name="score" value="0.10985927118255669" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i354]" time="0.031"><properties><property name="score" value="0.06911155176536128" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i355]" time="0.038"><properties><property name="score" value="0.055270866439063225" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i356]" time="0.036"><properties><property name="score" value="0.05432455469673076" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i357]" time="0.031"><properties><property name="score" value="0.009353495074961287" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i358]" time="0.034"><properties><property name="score" value="0.0914197002514098" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i359]" time="0.032"><properties><property name="score" value="0.003437610380325683" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i360]" time="0.027"><properties><property name="score" value="0.11050779356846534" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i361]" time="0.029"><properties><property name="score" value="0.02417899546545874" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i362]" time="0.027"><properties><property name="score" value="0.03546163737479552" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i363]" time="0.027"><properties><property name="score" value="0.0833566324465804" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i364]" time="0.026"><properties><property name="score" value="0.0948528551106034" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i365]" time="0.027"><properties><property name="score" value="0.10463914780321446" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i366]" time="0.056"><properties><property name="score" value="0.06583323171438815" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i367]" time="0.032"><properties><property name="score" value="0.08188630907593608" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i368]" time="0.031"><properties><property name="score" value="0.02894513311797281" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i369]" time="0.015"><properties><property name="score" value="0.08428540260209116" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i370]" time="0.028"><properties><property name="score" value="0.05789224189030047" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i371]" time="0.028"><properties><property name="score" value="0.06661807329136754" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i372]" time="0.026"><properties><property name="score" value="0.07693803983400938" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i373]" time="0.027"><properties><property name="score" value="0.09000915507169167" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i374]" time="0.012"><properties><property name="score" value="0.09523290567675091" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i375]" time="0.024"><properties><property name="score" value="0.07020952193602832" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i376]" time="0.029"><properties><property name="score" value="0.048510312213073226" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i377]" time="0.031"><properties><property name="score" value="0.08515994072978628" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i378]" time="0.014"><properties><property name="score" value="0.09256250370574537" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i379]" time="0.027"><properties><property name="score" value="0.08623056963996004" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i380]" time="0.027"><properties><property name="score" value="0.07836013585360797" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i381]" time="0.021"><properties><property name="score" value="0.11311611404334898" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i382]" time="0.080"><properties><property name="score" value="0.05575453139505269" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i383]" time="0.031"><properties><property name="score" value="0.020769558267615663" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i384]" time="0.042"><properties><property name="score" value="0.1356541619779074" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i385]" time="0.021"><properties><property name="score" value="0.09347504590960232" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i386]" time="0.042"><properties><property name="score" value="0.0664426541771269" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i387]" time="0.043"><properties><property name="score" value="0.07168194778362807" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i388]" time="0.102"><properties><property name="score" value="0.07973645069193283" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i389]" time="0.023"><properties><property name="score" value="0.018558657247941257" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i390]" time="0.070"><properties><property name="score" value="0.0556323764023623" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i391]" time="0.134"><properties><property name="score" value="0.04652619103309108" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i392]" time="0.041"><properties><property name="score" value="0.07997405481725459" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i393]" time="0.048"><properties><property name="score" value="0.09971170199760028" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i394]" time="0.048"><properties><property name="score" value="0.12397868249731296" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i395]" time="0.047"><properties><property name="score" value="0.11387369391094332" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i396]" time="0.035"><properties><property name="score" value="0.09567712466855197" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i397]" time="0.045"><properties><property name="score" value="0.052534499062360115" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i398]" time="0.043"><properties><property name="score" value="0.1803648042289558" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i399]" time="0.026"><properties><property name="score" value="0.11559660710040998" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i400]" time="0.035"><properties><property name="score" value="0.014794977221742257" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i401]" time="0.041"><properties><property name="score" value="0.07254252547823695" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i402]" time="0.040"><properties><property name="score" value="0.05831780114792762" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i403]" time="0.041"><properties><property name="score" value="0.06923220416634034" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i404]" time="0.037"><properties><property name="score" value="0.038449578081936255" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i405]" time="0.033"><properties><property name="score" value="0.07337492678997133" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i406]" time="0.042"><properties><property name="score" value="0.07891806255646132" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i407]" time="0.024"><properties><property name="score" value="0.13486798511452847" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i408]" time="0.082"><properties><property name="score" value="0.03927648290737252" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i409]" time="0.076"><properties><property name="score" value="0.039067702933518224" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i410]" time="0.062"><properties><property name="score" value="0.12491428575351446" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i411]" time="0.045"><properties><property name="score" value="0.061471042401101045" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i412]" time="0.046"><properties><property name="score" value="0.03997699766309496" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i413]" time="0.067"><properties><property name="score" value="0.02595622032093299" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i414]" time="0.053"><properties><property name="score" value="0.06170028850363763" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i415]" time="0.114"><properties><property name="score" value="0.07723356860087689" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i416]" time="0.103"><properties><property name="score" value="0.1525677534578973" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i417]" time="0.036"><properties><property name="score" value="0.12341629179294081" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i418]" time="0.034"><properties><property name="score" value="0.057781728383566405" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i419]" time="0.033"><properties><property name="score" value="0.036626578359978484" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i420]" time="0.033"><properties><property name="score" value="0.09401362421054629" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i421]" time="0.029"><properties><property name="score" value="0.04985253977501701" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i422]" time="0.028"><properties><property name="score" value="0.08256988910537498" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i423]" time="0.028"><properties><property name="score" value="0.08731978732478494" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i424]" time="0.029"><properties><property name="score" value="0.10461153993512205" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i425]" time="0.028"><properties><property name="score" value="0.07465398501591075" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i426]" time="0.014"><properties><property name="score" value="0.016110290583337683" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i427]" time="0.026"><properties><property name="score" value="0.11144773634820804" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i428]" time="0.026"><properties><property name="score" value="0.024036457555493194" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i429]" time="0.016"><properties><property name="score" value="0.09081344142402581" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i430]" time="0.031"><properties><property name="score" value="0.07879167647802757" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i431]" time="0.027"><properties><property name="score" value="0.21736930655872824" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i432]" time="0.031"><properties><property name="score" value="0.09391287687613553" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i433]" time="0.074"><properties><property name="score" value="0.08399408037681862" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i434]" time="0.034"><properties><property name="score" value="0.06957448828029204" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i435]" time="0.014"><properties><property name="score" value="0.006740943026578572" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i436]" time="0.029"><properties><property name="score" value="0.07557666090881095" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i437]" time="0.026"><properties><property name="score" value="0.052927143390398335" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i438]" time="0.014"><properties><property name="score" value="0.11076713231057966" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i439]" time="0.013"><properties><property name="score" value="0.10105632617506966" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i440]" time="0.025"><properties><property name="score" value="0.020135919583327645" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i441]" time="0.025"><properties><property name="score" value="0.09419398169899945" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i442]" time="0.025"><properties><property name="score" value="0.049350709875234405" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i443]" time="0.026"><properties><property name="score" value="0.08562043909403232" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i444]" time="0.026"><properties><property name="score" value="0.1676475627246426" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i445]" time="0.026"><properties><property name="score" value="0.09468149742483209" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i446]" time="0.027"><properties><property name="score" value="0.05350056224592829" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i447]" time="0.013"><properties><property name="score" value="0.10418347909160186" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i448]" time="0.024"><properties><property name="score" value="0.10213790378745748" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i449]" time="0.026"><properties><property name="score" value="0.09843995569061596" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i450]" time="0.024"><properties><property name="score" value="0.11940936949202463" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i451]" time="0.014"><properties><property name="score" value="0.10736312707292361" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i452]" time="0.024"><properties><property name="score" value="0.03939985352751565" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i453]" time="0.024"><properties><property name="score" value="0.04709659337629434" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i454]" time="0.025"><properties><property name="score" value="0.014908858117816939" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.01490886&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.01490886
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i455]" time="0.025"><properties><property name="score" value="0.05989173522864625" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i456]" time="0.024"><properties><property name="score" value="0.05132818867041412" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i457]" time="0.013"><properties><property name="score" value="0.06611680248657659" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i458]" time="0.013"><properties><property name="score" value="0.10417428487151659" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i459]" time="0.023"><properties><property name="score" value="0.0910689283780238" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i460]" time="0.023"><properties><property name="score" value="0.09857833304787991" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i461]" time="0.012"><properties><property name="score" value="0.061620747889284955" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i462]" time="0.023"><properties><property name="score" value="0.026122803698599824" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i463]" time="0.023"><properties><property name="score" value="0.08077662939208591" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i464]" time="0.023"><properties><property name="score" value="0.012988894072389913" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i465]" time="0.023"><properties><property name="score" value="0.07775524802727668" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i466]" time="0.026"><properties><property name="score" value="0.04813873366105356" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i467]" time="0.028"><properties><property name="score" value="0.08182953466860063" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i468]" time="0.026"><properties><property name="score" value="0.06457805091018463" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i469]" time="0.024"><properties><property name="score" value="0.0716279890796653" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i470]" time="0.023"><properties><property name="score" value="0.06973257569935871" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i471]" time="0.022"><properties><property name="score" value="0.06781645879358375" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i472]" time="0.011"><properties><property name="score" value="0.13725668495134236" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i473]" time="0.022"><properties><property name="score" value="0.0028359040163306788" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i474]" time="0.011"><properties><property name="score" value="0.06966067525813191" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i475]" time="0.022"><properties><property name="score" value="0.04064692617638725" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i476]" time="0.021"><properties><property name="score" value="0.08303973241454424" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i477]" time="0.021"><properties><property name="score" value="0.03798976540845614" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i478]" time="0.022"><properties><property name="score" value="0.043262635118560495" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i479]" time="0.022"><properties><property name="score" value="0.054241905921615874" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i480]" time="0.021"><properties><property name="score" value="0.05791194968070232" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i481]" time="0.021"><properties><property name="score" value="0.06175900966086617" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i482]" time="0.060"><properties><property name="score" value="0.09012530009076294" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i483]" time="0.013"><properties><property name="score" value="0.03562194391848461" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.03562194&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.03562194
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i484]" time="0.022"><properties><property name="score" value="0.05085529907304353" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i485]" time="0.022"><properties><property name="score" value="0.05423861107663812" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i486]" time="0.022"><properties><property name="score" value="0.11350670438138835" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i487]" time="0.013"><properties><property name="score" value="0.14484224029893134" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i488]" time="0.026"><properties><property name="score" value="0.014701504513512124" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.0147015&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.0147015
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i489]" time="0.022"><properties><property name="score" value="0.05591967466976461" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i490]" time="0.022"><properties><property name="score" value="0.0816717748080221" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i491]" time="0.021"><properties><property name="score" value="0.03829179154805801" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i492]" time="0.021"><properties><property name="score" value="0.08371425056254038" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i493]" time="0.021"><properties><property name="score" value="0.06709652987572094" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i494]" time="0.020"><properties><property name="score" value="0.06557817541276603" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i495]" time="0.010"><properties><property name="score" value="0.04921086786899287" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i496]" time="0.020"><properties><property name="score" value="0.04320321434265543" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i497]" time="0.020"><properties><property name="score" value="0.04738466998979453" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i498]" time="0.021"><properties><property name="score" value="0.09975276020999689" /></properties></testcase><testcase classname="test_lzma_detect" name="test_humannews_jsonl[i499]" time="0.021"><properties><property name="score" value="0.12377265206380217" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i0]" time="0.021"><properties><property name="score" value="0.03432566503081904" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i1]" time="0.022"><properties><property name="score" value="0.02506350160112071" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i2]" time="0.011"><properties><property name="score" value="0.011435184710473223" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i3]" time="0.023"><properties><property name="score" value="0.022285685617277166" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i4]" time="0.022"><properties><property name="score" value="0.005414082529356" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i5]" time="0.022"><properties><property name="score" value="0.0002800864767085161" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00028009&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00028009
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i6]" time="0.022"><properties><property name="score" value="0.03990266637650108" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03990267&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03990267
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i7]" time="0.024"><properties><property name="score" value="0.004287297615021191" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i8]" time="0.020"><properties><property name="score" value="0.009614749888470265" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00961475&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00961475
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i9]" time="0.021"><properties><property name="score" value="0.005488672025356724" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00548867&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00548867
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i10]" time="0.012"><properties><property name="score" value="0.01135305542536158" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01135306&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01135306
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i11]" time="0.021"><properties><property name="score" value="0.03779012360411607" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03779012&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03779012
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i12]" time="0.012"><properties><property name="score" value="0.028766738314528473" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i13]" time="0.020"><properties><property name="score" value="0.00185708998885116" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i14]" time="0.021"><properties><property name="score" value="0.12154569460183764" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i15]" time="0.020"><properties><property name="score" value="0.0018944918693233614" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00189449&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00189449
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i16]" time="0.025"><properties><property name="score" value="0.031731041026554496" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03173104&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03173104
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i17]" time="0.022"><properties><property name="score" value="0.016332221957197756" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i18]" time="0.026"><properties><property name="score" value="0.0004233315513716097" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i19]" time="0.068"><properties><property name="score" value="0.039464706502947464" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03946471&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03946471
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i20]" time="0.013"><properties><property name="score" value="0.048613423133853084" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04861342&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04861342
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i21]" time="0.023"><properties><property name="score" value="0.009415245658526052" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00941525&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00941525
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i22]" time="0.014"><properties><property name="score" value="0.018762842013930747" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01876284&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01876284
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i23]" time="0.025"><properties><property name="score" value="0.022298392211365158" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i24]" time="0.015"><properties><property name="score" value="0.011612387022941206" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01161239&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01161239
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i25]" time="0.030"><properties><property name="score" value="0.01916992887852525" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01916993&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01916993
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i26]" time="0.030"><properties><property name="score" value="0.025083666376257592" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02508367&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02508367
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i27]" time="0.027"><properties><property name="score" value="0.06858107771275773" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06858108&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06858108
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i28]" time="0.029"><properties><property name="score" value="0.06507742056301813" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i29]" time="0.014"><properties><property name="score" value="0.04585402390246407" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04585402&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04585402
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i30]" time="0.030"><properties><property name="score" value="0.044546921124439955" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04454692&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04454692
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i31]" time="0.023"><properties><property name="score" value="0.0600475408451262" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06004754&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06004754
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i32]" time="0.028"><properties><property name="score" value="0.035420367450225775" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i33]" time="0.024"><properties><property name="score" value="0.02641250841827657" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i34]" time="0.025"><properties><property name="score" value="0.009777318153743183" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00977732&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00977732
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i35]" time="0.012"><properties><property name="score" value="0.05270928837968758" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05270929&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05270929
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i36]" time="0.022"><properties><property name="score" value="0.08237481738815057" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i37]" time="0.023"><properties><property name="score" value="0.012118810054293789" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i38]" time="0.023"><properties><property name="score" value="0.05190847513240256" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05190848&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05190848
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i39]" time="0.023"><properties><property name="score" value="0.01464971683826916" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01464972&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01464972
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i40]" time="0.070"><properties><property name="score" value="0.011749975272924623" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i41]" time="0.013"><properties><property name="score" value="0.004292406785916247" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i42]" time="0.023"><properties><property name="score" value="0.036421624451735884" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i43]" time="0.023"><properties><property name="score" value="0.006228716362281841" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00622872&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00622872
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i44]" time="0.023"><properties><property name="score" value="0.028351812038653258" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i45]" time="0.025"><properties><property name="score" value="0.0030481284450260236" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00304813&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00304813
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i46]" time="0.022"><properties><property name="score" value="0.05191264625993128" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i47]" time="0.021"><properties><property name="score" value="0.03589904013748092" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i48]" time="0.023"><properties><property name="score" value="0.030344643309225205" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03034464&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03034464
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i49]" time="0.022"><properties><property name="score" value="0.02129348889454594" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02129349&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02129349
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i50]" time="0.022"><properties><property name="score" value="0.008123061366219847" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00812306&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00812306
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i51]" time="0.022"><properties><property name="score" value="0.021627094200283158" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02162709&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02162709
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i52]" time="0.025"><properties><property name="score" value="0.041342081505171824" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04134208&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04134208
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i53]" time="0.084"><properties><property name="score" value="0.003674929644796326" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i54]" time="0.061"><properties><property name="score" value="0.0261953427184354" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i55]" time="0.022"><properties><property name="score" value="0.015499929851222904" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i56]" time="0.024"><properties><property name="score" value="0.06075342671754648" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06075343&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06075343
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i57]" time="0.013"><properties><property name="score" value="0.04205309142546465" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04205309&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04205309
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i58]" time="0.022"><properties><property name="score" value="0.016120506285489455" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01612051&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01612051
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i59]" time="0.021"><properties><property name="score" value="0.0381277569619894" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i60]" time="0.011"><properties><property name="score" value="0.037039005956124305" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i61]" time="0.022"><properties><property name="score" value="0.060312748418841794" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06031275&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06031275
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i62]" time="0.021"><properties><property name="score" value="0.04564684490300093" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i63]" time="0.021"><properties><property name="score" value="0.014199280768503039" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i64]" time="0.020"><properties><property name="score" value="0.007107876792630616" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00710788&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00710788
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i65]" time="0.021"><properties><property name="score" value="0.024958768639840048" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02495877&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02495877
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i66]" time="0.021"><properties><property name="score" value="0.020172121825265447" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i67]" time="0.020"><properties><property name="score" value="0.030135791802712437" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i68]" time="0.021"><properties><property name="score" value="0.0017110029439671216" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.001711&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.001711
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i69]" time="0.021"><properties><property name="score" value="0.0533652379716879" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05336524&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05336524
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i70]" time="0.021"><properties><property name="score" value="0.0016728664709750483" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00167287&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00167287
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i71]" time="0.021"><properties><property name="score" value="0.004593373670636426" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00459337&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00459337
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i72]" time="0.062"><properties><property name="score" value="0.004033306964425438" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00403331&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00403331
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i73]" time="0.060"><properties><property name="score" value="0.027260991070093787" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i74]" time="0.022"><properties><property name="score" value="0.03275553402335369" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i75]" time="0.022"><properties><property name="score" value="0.016991397585250007" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0169914&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0169914
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i76]" time="0.021"><properties><property name="score" value="0.03858845537197712" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i77]" time="0.022"><properties><property name="score" value="0.029442857061089234" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02944286&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02944286
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i78]" time="0.022"><properties><property name="score" value="0.028844751084793785" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02884475&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02884475
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i79]" time="0.029"><properties><property name="score" value="0.025682548338415906" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02568255&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02568255
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i80]" time="0.024"><properties><property name="score" value="0.05547493594882062" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i81]" time="0.027"><properties><property name="score" value="0.015107464209078148" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01510746&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01510746
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i82]" time="0.011"><properties><property name="score" value="0.03550679682758351" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i83]" time="0.011"><properties><property name="score" value="0.03157055461684988" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i84]" time="0.025"><properties><property name="score" value="0.00825623815334503" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00825624&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00825624
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i85]" time="0.023"><properties><property name="score" value="0.033691226119838116" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03369123&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03369123
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i86]" time="0.024"><properties><property name="score" value="0.016563315876168716" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i87]" time="0.017"><properties><property name="score" value="0.007596345525018311" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i88]" time="0.026"><properties><property name="score" value="0.030374832114327212" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i89]" time="0.038"><properties><property name="score" value="0.07047223979479851" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07047224&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07047224
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i90]" time="0.032"><properties><property name="score" value="0.049408028268464674" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i91]" time="0.024"><properties><property name="score" value="0.026514818500542117" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i92]" time="0.026"><properties><property name="score" value="0.06342957699705147" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i93]" time="0.034"><properties><property name="score" value="0.026390079904236446" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02639008&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02639008
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i94]" time="0.043"><properties><property name="score" value="0.0027754403746822164" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i95]" time="0.048"><properties><property name="score" value="0.057018585197759444" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05701859&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05701859
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i96]" time="0.024"><properties><property name="score" value="0.026635370991656693" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02663537&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02663537
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i97]" time="0.057"><properties><property name="score" value="0.030068738226558842" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i98]" time="0.049"><properties><property name="score" value="0.016104852510984244" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01610485&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01610485
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i99]" time="0.050"><properties><property name="score" value="0.005439481252079108" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i100]" time="0.032"><properties><property name="score" value="0.0035181321835386914" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00351813&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00351813
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i101]" time="0.055"><properties><property name="score" value="0.014953882330735012" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01495388&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01495388
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i102]" time="0.053"><properties><property name="score" value="0.04142712036231755" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i103]" time="0.057"><properties><property name="score" value="0.02690530373368414" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i104]" time="0.059"><properties><property name="score" value="0.050174630095047294" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05017463&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05017463
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i105]" time="0.063"><properties><property name="score" value="0.014537834407392172" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01453783&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01453783
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i106]" time="0.045"><properties><property name="score" value="0.008295612874105806" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i107]" time="0.047"><properties><property name="score" value="0.008864475955661814" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i108]" time="0.040"><properties><property name="score" value="0.06977047212060801" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06977047&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06977047
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i109]" time="0.027"><properties><property name="score" value="0.021060642041742517" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i110]" time="0.049"><properties><property name="score" value="0.020771724123894126" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02077172&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02077172
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i111]" time="0.039"><properties><property name="score" value="0.09427785432263094" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.09427785&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.09427785
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i112]" time="0.042"><properties><property name="score" value="0.02005833373731563" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02005833&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02005833
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i113]" time="0.039"><properties><property name="score" value="0.006260364852444409" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00626036&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00626036
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i114]" time="0.039"><properties><property name="score" value="0.019713403030450383" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i115]" time="0.048"><properties><property name="score" value="0.023257859854485877" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02325786&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02325786
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i116]" time="0.021"><properties><property name="score" value="0.013559913471394847" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01355991&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01355991
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i117]" time="0.040"><properties><property name="score" value="0.0029357627747014893" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00293576&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00293576
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i118]" time="0.033"><properties><property name="score" value="0.003940552027322217" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i119]" time="0.038"><properties><property name="score" value="0.0793885583796515" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07938856&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07938856
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i120]" time="0.041"><properties><property name="score" value="0.0014333759016156738" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00143338&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00143338
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i121]" time="0.035"><properties><property name="score" value="0.009679521296504456" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i122]" time="0.019"><properties><property name="score" value="0.09791233544890199" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.09791234&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.09791234
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i123]" time="0.045"><properties><property name="score" value="0.013174226666836578" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i124]" time="0.039"><properties><property name="score" value="0.010080113201787484" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01008011&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01008011
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i125]" time="0.042"><properties><property name="score" value="0.04258220155064854" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i126]" time="0.036"><properties><property name="score" value="0.030153241596108722" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03015324&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03015324
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i127]" time="0.041"><properties><property name="score" value="0.03657969669948424" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0365797&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0365797
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i128]" time="0.050"><properties><property name="score" value="0.04039147949345449" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i129]" time="0.044"><properties><property name="score" value="0.010785233841430058" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i130]" time="0.023"><properties><property name="score" value="0.02448169151754165" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i131]" time="0.041"><properties><property name="score" value="0.023390996394097488" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.023391&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.023391
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i132]" time="0.046"><properties><property name="score" value="0.042013089758109814" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04201309&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04201309
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i133]" time="0.048"><properties><property name="score" value="0.03150735601442918" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03150736&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03150736
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i134]" time="0.046"><properties><property name="score" value="0.019789665313381437" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i135]" time="0.022"><properties><property name="score" value="0.07213707645761236" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07213708&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07213708
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i136]" time="0.022"><properties><property name="score" value="0.056677491269085944" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i137]" time="0.024"><properties><property name="score" value="0.06716209401944617" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06716209&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06716209
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i138]" time="0.021"><properties><property name="score" value="0.0265125502720176" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i139]" time="0.038"><properties><property name="score" value="0.0009279162023789301" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00092792&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00092792
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i140]" time="0.037"><properties><property name="score" value="0.047536127338128016" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i141]" time="0.038"><properties><property name="score" value="0.030575099122563194" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0305751&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0305751
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i142]" time="0.043"><properties><property name="score" value="0.013360534597680888" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i143]" time="0.056"><properties><property name="score" value="0.01951283352620826" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i144]" time="0.051"><properties><property name="score" value="0.03257010887469378" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i145]" time="0.048"><properties><property name="score" value="0.04315781939487004" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i146]" time="0.064"><properties><property name="score" value="0.03060114182940982" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03060114&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03060114
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i147]" time="0.069"><properties><property name="score" value="0.018572150923795414" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i148]" time="0.060"><properties><property name="score" value="0.020190119464228627" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02019012&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02019012
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i149]" time="0.136"><properties><property name="score" value="0.04304098007435418" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04304098&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04304098
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i150]" time="0.065"><properties><property name="score" value="0.01974006175018061" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01974006&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01974006
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i151]" time="0.089"><properties><property name="score" value="0.05115537873821529" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05115538&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05115538
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i152]" time="0.048"><properties><property name="score" value="0.050938382811748555" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i153]" time="0.057"><properties><property name="score" value="0.034003809216728687" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03400381&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03400381
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i154]" time="0.045"><properties><property name="score" value="0.012069109547892665" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i155]" time="0.053"><properties><property name="score" value="0.017209802907503322" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0172098&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0172098
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i156]" time="0.044"><properties><property name="score" value="0.017229348070939808" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01722935&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01722935
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i157]" time="0.041"><properties><property name="score" value="0.03404623713593629" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i158]" time="0.052"><properties><property name="score" value="0.04559814723395288" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04559815&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04559815
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i159]" time="0.053"><properties><property name="score" value="0.057939666723084496" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05793967&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05793967
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i160]" time="0.050"><properties><property name="score" value="0.03711572025840607" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03711572&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03711572
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i161]" time="0.026"><properties><property name="score" value="0.06551829657228114" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0655183&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0655183
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i162]" time="0.049"><properties><property name="score" value="0.0341063079856912" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i163]" time="0.030"><properties><property name="score" value="0.015262641256663656" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i164]" time="0.031"><properties><property name="score" value="0.015654575875090337" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i165]" time="0.068"><properties><property name="score" value="0.0020801861947658705" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00208019&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00208019
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i166]" time="0.046"><properties><property name="score" value="0.031369487919361706" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03136949&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03136949
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i167]" time="0.054"><properties><property name="score" value="0.0023018287313667663" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00230183&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00230183
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i168]" time="0.035"><properties><property name="score" value="0.034741679655814206" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03474168&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03474168
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i169]" time="0.032"><properties><property name="score" value="0.058268609476885835" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05826861&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05826861
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i170]" time="0.049"><properties><property name="score" value="0.06963313851649848" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i171]" time="0.055"><properties><property name="score" value="0.04417329157994031" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04417329&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04417329
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i172]" time="0.051"><properties><property name="score" value="0.0016240176622610422" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i173]" time="0.065"><properties><property name="score" value="0.010970500133665632" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i174]" time="0.054"><properties><property name="score" value="0.016574363510009474" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i175]" time="0.045"><properties><property name="score" value="0.0026160410295054826" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i176]" time="0.042"><properties><property name="score" value="0.0029433181907440216" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i177]" time="0.024"><properties><property name="score" value="0.05398965475937323" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05398965&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05398965
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i178]" time="0.038"><properties><property name="score" value="0.005429118653652331" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00542912&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00542912
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i179]" time="0.042"><properties><property name="score" value="0.020997527959464208" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02099753&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02099753
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i180]" time="0.042"><properties><property name="score" value="0.019493020418723545" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01949302&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01949302
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i181]" time="0.052"><properties><property name="score" value="0.07736047212580964" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07736047&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07736047
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i182]" time="0.054"><properties><property name="score" value="0.06278999034872848" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i183]" time="0.066"><properties><property name="score" value="0.029873287667974004" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02987329&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02987329
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i184]" time="0.058"><properties><property name="score" value="0.032789189343217943" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i185]" time="0.024"><properties><property name="score" value="0.0180489350859403" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i186]" time="0.057"><properties><property name="score" value="0.021304778711203745" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i187]" time="0.044"><properties><property name="score" value="0.07457648085152702" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i188]" time="0.050"><properties><property name="score" value="0.0540908553960755" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05409086&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05409086
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i189]" time="0.051"><properties><property name="score" value="0.030356965816296773" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i190]" time="0.066"><properties><property name="score" value="0.02775014795812931" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i191]" time="0.054"><properties><property name="score" value="0.017183145374559113" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01718315&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01718315
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i192]" time="0.051"><properties><property name="score" value="0.006607404750194347" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i193]" time="0.043"><properties><property name="score" value="0.021502030267332325" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02150203&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02150203
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i194]" time="0.052"><properties><property name="score" value="0.008255909697185773" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i195]" time="0.043"><properties><property name="score" value="0.013466156612897448" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01346616&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01346616
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i196]" time="0.121"><properties><property name="score" value="0.05287723352555931" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i197]" time="0.060"><properties><property name="score" value="0.0048427209528485015" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00484272&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00484272
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i198]" time="0.052"><properties><property name="score" value="0.017500480570767318" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01750048&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01750048
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i199]" time="0.078"><properties><property name="score" value="0.01703221493004623" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i200]" time="0.041"><properties><property name="score" value="0.0027772126572511713" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00277721&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00277721
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i201]" time="0.031"><properties><property name="score" value="0.020395428279124106" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i202]" time="0.055"><properties><property name="score" value="0.026232495690950635" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0262325&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0262325
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i203]" time="0.048"><properties><property name="score" value="0.08085098752798077" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.08085099&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.08085099
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i204]" time="0.056"><properties><property name="score" value="0.04390620740170253" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i205]" time="0.045"><properties><property name="score" value="0.021146621731611526" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i206]" time="0.143"><properties><property name="score" value="0.08015271936203915" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i207]" time="0.046"><properties><property name="score" value="0.008543228294733897" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00854323&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00854323
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i208]" time="0.047"><properties><property name="score" value="0.049385895978252485" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i209]" time="0.020"><properties><property name="score" value="0.010439440294812696" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01043944&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01043944
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i210]" time="0.142"><properties><property name="score" value="0.02409213636650171" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02409214&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02409214
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i211]" time="0.075"><properties><property name="score" value="0.0561499396010262" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05614994&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05614994
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i212]" time="0.035"><properties><property name="score" value="0.002650796803238542" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i213]" time="0.067"><properties><property name="score" value="0.016290005518831217" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i214]" time="0.083"><properties><property name="score" value="0.0249598066288563" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02495981&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02495981
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i215]" time="0.347"><properties><property name="score" value="0.005677903830970298" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0056779&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0056779
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i216]" time="0.056"><properties><property name="score" value="0.019489451476711883" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01948945&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01948945
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i217]" time="0.086"><properties><property name="score" value="0.032106870655407876" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03210687&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03210687
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i218]" time="0.065"><properties><property name="score" value="0.06672362011367261" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06672362&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06672362
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i219]" time="0.047"><properties><property name="score" value="0.044989957943922174" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i220]" time="0.054"><properties><property name="score" value="0.015062997327137676" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.015063&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.015063
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i221]" time="0.059"><properties><property name="score" value="0.014713341127922424" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i222]" time="0.028"><properties><property name="score" value="0.04231699365081498" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04231699&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04231699
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i223]" time="0.056"><properties><property name="score" value="0.03752050587400639" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i224]" time="0.057"><properties><property name="score" value="0.016830890762056842" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01683089&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01683089
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i225]" time="0.045"><properties><property name="score" value="0.0192172402928689" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01921724&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01921724
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i226]" time="0.055"><properties><property name="score" value="0.0031389270799443295" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00313893&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00313893
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i227]" time="0.064"><properties><property name="score" value="0.04317387715370625" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04317388&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04317388
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i228]" time="0.074"><properties><property name="score" value="0.06498562250959451" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06498562&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06498562
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i229]" time="0.065"><properties><property name="score" value="0.00954549947699757" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0095455&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0095455
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i230]" time="0.058"><properties><property name="score" value="0.022955213758598887" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i231]" time="0.050"><properties><property name="score" value="0.028215519759133625" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i232]" time="0.027"><properties><property name="score" value="0.011807376253614166" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01180738&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01180738
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i233]" time="0.046"><properties><property name="score" value="0.010447450093043941" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i234]" time="0.051"><properties><property name="score" value="0.02858858001826438" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i235]" time="0.052"><properties><property name="score" value="0.07804318929355455" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07804319&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07804319
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i236]" time="0.061"><properties><property name="score" value="0.07659465608192577" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07659466&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07659466
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i237]" time="0.039"><properties><property name="score" value="0.010141311621388031" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i238]" time="0.022"><properties><property name="score" value="0.06596694601934128" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06596695&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06596695
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i239]" time="0.038"><properties><property name="score" value="0.017574313554588473" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01757431&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01757431
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i240]" time="0.051"><properties><property name="score" value="0.024148063366916425" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02414806&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02414806
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i241]" time="0.037"><properties><property name="score" value="0.002035562086591417" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00203556&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00203556
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i242]" time="0.026"><properties><property name="score" value="0.05976807585060806" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05976808&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05976808
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i243]" time="0.039"><properties><property name="score" value="0.006801856060918564" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i244]" time="0.044"><properties><property name="score" value="0.0443572960697637" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0443573&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0443573
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i245]" time="0.038"><properties><property name="score" value="0.023574013223395588" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02357401&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02357401
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i246]" time="0.036"><properties><property name="score" value="0.04614084034502426" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04614084&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04614084
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i247]" time="0.039"><properties><property name="score" value="0.02419321491192237" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i248]" time="0.043"><properties><property name="score" value="0.0009509763436480953" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i249]" time="0.026"><properties><property name="score" value="0.043901375065608095" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04390138&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04390138
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i250]" time="0.024"><properties><property name="score" value="0.03220753669433973" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03220754&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03220754
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i251]" time="0.043"><properties><property name="score" value="0.009013648073468935" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00901365&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00901365
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i252]" time="0.047"><properties><property name="score" value="0.0404230956993018" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0404231&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0404231
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i253]" time="0.044"><properties><property name="score" value="0.052193291466040466" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i254]" time="0.026"><properties><property name="score" value="0.011223373196794784" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01122337&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01122337
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i255]" time="0.023"><properties><property name="score" value="0.08314283485489349" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.08314283&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.08314283
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i256]" time="0.039"><properties><property name="score" value="0.008840423475750243" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i257]" time="0.017"><properties><property name="score" value="0.01665102694138887" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01665103&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01665103
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i258]" time="0.042"><properties><property name="score" value="0.008487588703620373" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i259]" time="0.034"><properties><property name="score" value="0.015787440951821937" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01578744&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01578744
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i260]" time="0.032"><properties><property name="score" value="0.0033685451885377736" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00336855&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00336855
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i261]" time="0.039"><properties><property name="score" value="0.01519633749908633" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i262]" time="0.035"><properties><property name="score" value="0.00800150852290782" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00800151&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00800151
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i263]" time="0.037"><properties><property name="score" value="0.0067910041607805085" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i264]" time="0.039"><properties><property name="score" value="0.002888992372483816" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00288899&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00288899
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i265]" time="0.038"><properties><property name="score" value="0.0012256581338751255" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00122566&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00122566
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i266]" time="0.033"><properties><property name="score" value="0.04236615506798176" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i267]" time="0.020"><properties><property name="score" value="0.03312822721591724" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03312823&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03312823
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i268]" time="0.035"><properties><property name="score" value="0.07168631711128823" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07168632&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07168632
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i269]" time="0.095"><properties><property name="score" value="0.06327015999098012" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i270]" time="0.036"><properties><property name="score" value="0.002615055514067056" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i271]" time="0.030"><properties><property name="score" value="0.040123032914737615" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04012303&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04012303
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i272]" time="0.032"><properties><property name="score" value="0.0850593618869673" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.08505936&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.08505936
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i273]" time="0.033"><properties><property name="score" value="0.02896681786142651" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02896682&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02896682
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i274]" time="0.040"><properties><property name="score" value="0.07145942604334976" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07145943&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07145943
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i275]" time="0.034"><properties><property name="score" value="0.04577696328781197" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04577696&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04577696
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i276]" time="0.033"><properties><property name="score" value="0.008300395180988428" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i277]" time="0.041"><properties><property name="score" value="0.033622271101347107" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i278]" time="0.018"><properties><property name="score" value="0.04854234930918855" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04854235&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04854235
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i279]" time="0.036"><properties><property name="score" value="0.022901949697984796" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i280]" time="0.032"><properties><property name="score" value="0.002144589487761474" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00214459&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00214459
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i281]" time="0.044"><properties><property name="score" value="0.056584793974028005" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i282]" time="0.032"><properties><property name="score" value="0.04202677708892328" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04202678&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04202678
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i283]" time="0.030"><properties><property name="score" value="0.027314824167286957" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02731482&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02731482
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i284]" time="0.034"><properties><property name="score" value="0.026352807308353498" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i285]" time="0.101"><properties><property name="score" value="0.027352291020710358" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02735229&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02735229
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i286]" time="0.034"><properties><property name="score" value="0.03737911547005109" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03737912&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03737912
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i287]" time="0.033"><properties><property name="score" value="0.03392913877175996" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03392914&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03392914
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i288]" time="0.083"><properties><property name="score" value="0.05682835067128093" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05682835&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05682835
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i289]" time="0.143"><properties><property name="score" value="0.04200369804027561" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i290]" time="0.041"><properties><property name="score" value="0.00987959376497973" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i291]" time="0.033"><properties><property name="score" value="0.05813860827318111" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i292]" time="0.040"><properties><property name="score" value="0.005400416615779155" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00540042&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00540042
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i293]" time="0.039"><properties><property name="score" value="0.012582305499154012" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01258231&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01258231
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i294]" time="0.038"><properties><property name="score" value="0.03241102798751139" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i295]" time="0.031"><properties><property name="score" value="0.026876658001620583" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02687666&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02687666
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i296]" time="0.032"><properties><property name="score" value="0.04442540633108221" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04442541&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04442541
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i297]" time="0.021"><properties><property name="score" value="0.015006970180458934" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i298]" time="0.035"><properties><property name="score" value="0.013562281537124144" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i299]" time="0.033"><properties><property name="score" value="0.01270686585592586" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i300]" time="0.035"><properties><property name="score" value="0.05171153535966945" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i301]" time="0.035"><properties><property name="score" value="0.0018753313799362206" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i302]" time="0.129"><properties><property name="score" value="0.040079833282079605" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i303]" time="0.073"><properties><property name="score" value="0.04460314244452772" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04460314&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04460314
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i304]" time="0.059"><properties><property name="score" value="0.004061877950824745" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i305]" time="0.038"><properties><property name="score" value="0.0620021883503494" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06200219&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06200219
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i306]" time="0.031"><properties><property name="score" value="0.058906738192121555" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i307]" time="0.097"><properties><property name="score" value="0.05064231724301855" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i308]" time="0.034"><properties><property name="score" value="0.003397174004532595" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i309]" time="0.036"><properties><property name="score" value="0.00956370171585358" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i310]" time="0.034"><properties><property name="score" value="0.0447397080790001" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04473971&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04473971
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i311]" time="0.031"><properties><property name="score" value="0.040766172449050335" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04076617&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04076617
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i312]" time="0.031"><properties><property name="score" value="0.02833144990028972" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02833145&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02833145
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i313]" time="0.032"><properties><property name="score" value="0.0822897306224078" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i314]" time="0.037"><properties><property name="score" value="0.07129781653573038" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07129782&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07129782
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i315]" time="0.030"><properties><property name="score" value="0.029455106376238424" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02945511&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02945511
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i316]" time="0.033"><properties><property name="score" value="0.04122321217800676" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04122321&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04122321
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i317]" time="0.034"><properties><property name="score" value="0.04068447225883383" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i318]" time="0.032"><properties><property name="score" value="0.026199947032144264" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02619995&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02619995
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i319]" time="0.047"><properties><property name="score" value="0.013644584122572103" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01364458&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01364458
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i320]" time="0.084"><properties><property name="score" value="0.055455663046147" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05545566&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05545566
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i321]" time="0.041"><properties><property name="score" value="0.05904293375553755" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i322]" time="0.032"><properties><property name="score" value="0.014109537075584462" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i323]" time="0.032"><properties><property name="score" value="0.029709061257570828" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02970906&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02970906
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i324]" time="0.034"><properties><property name="score" value="0.054862090982666104" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05486209&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05486209
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i325]" time="0.030"><properties><property name="score" value="0.014729943420559005" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01472994&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01472994
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i326]" time="0.032"><properties><property name="score" value="0.0173387356537548" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i327]" time="0.030"><properties><property name="score" value="0.03342558918771622" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i328]" time="0.034"><properties><property name="score" value="0.0066815725958939386" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i329]" time="0.037"><properties><property name="score" value="0.03277044530743636" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03277045&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03277045
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i330]" time="0.029"><properties><property name="score" value="0.005362176468244172" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i331]" time="0.028"><properties><property name="score" value="0.09787396530029265" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i332]" time="0.030"><properties><property name="score" value="0.05042651897872393" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05042652&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05042652
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i333]" time="0.080"><properties><property name="score" value="0.014366052267890735" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i334]" time="0.032"><properties><property name="score" value="0.02143980001364737" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i335]" time="0.031"><properties><property name="score" value="0.05823401281693227" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05823401&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05823401
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i336]" time="0.029"><properties><property name="score" value="0.043086040199032095" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04308604&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04308604
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i337]" time="0.028"><properties><property name="score" value="0.05566473010235916" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i338]" time="0.030"><properties><property name="score" value="0.05031926131983828" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05031926&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05031926
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i339]" time="0.027"><properties><property name="score" value="0.009770438946639998" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i340]" time="0.027"><properties><property name="score" value="0.046375528124488544" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04637553&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04637553
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i341]" time="0.015"><properties><property name="score" value="0.004871392138477981" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00487139&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00487139
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i342]" time="0.026"><properties><property name="score" value="0.01825405946015391" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i343]" time="0.027"><properties><property name="score" value="0.043256598519048756" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0432566&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0432566
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i344]" time="0.026"><properties><property name="score" value="0.0061242027577199055" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0061242&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0061242
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i345]" time="0.027"><properties><property name="score" value="0.04152621779632723" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04152622&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04152622
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i346]" time="0.016"><properties><property name="score" value="0.07461886810273266" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07461887&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07461887
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i347]" time="0.027"><properties><property name="score" value="0.020173007308137172" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02017301&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02017301
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i348]" time="0.027"><properties><property name="score" value="0.023487403133855233" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0234874&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0234874
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i349]" time="0.027"><properties><property name="score" value="0.09878371240747502" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i350]" time="0.034"><properties><property name="score" value="0.0651498635727309" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06514986&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06514986
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i351]" time="0.025"><properties><property name="score" value="0.018277947673390655" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01827795&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01827795
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i352]" time="0.027"><properties><property name="score" value="0.005286947161530977" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00528695&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00528695
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i353]" time="0.026"><properties><property name="score" value="0.03722509379094133" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03722509&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03722509
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i354]" time="0.028"><properties><property name="score" value="0.06949413394503504" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i355]" time="0.032"><properties><property name="score" value="0.02147282426811272" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i356]" time="0.033"><properties><property name="score" value="0.0018330219218815893" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00183302&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00183302
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i357]" time="0.038"><properties><property name="score" value="0.028881029718166105" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i358]" time="0.029"><properties><property name="score" value="0.012185009200271211" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i359]" time="0.026"><properties><property name="score" value="0.01581932119198695" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i360]" time="0.026"><properties><property name="score" value="0.04975646328831701" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04975646&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04975646
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i361]" time="0.034"><properties><property name="score" value="0.0009873755159287548" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00098738&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00098738
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i362]" time="0.034"><properties><property name="score" value="0.00063901212594285" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00063901&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00063901
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i363]" time="0.030"><properties><property name="score" value="0.030922614570398826" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03092261&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03092261
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i364]" time="0.033"><properties><property name="score" value="0.0630544713402996" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06305447&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06305447
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i365]" time="0.028"><properties><property name="score" value="0.013210259606413694" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i366]" time="0.031"><properties><property name="score" value="0.02147583076725293" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02147583&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02147583
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i367]" time="0.031"><properties><property name="score" value="0.021768432666373197" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i368]" time="0.026"><properties><property name="score" value="0.012554501568012622" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0125545&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0125545
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i369]" time="0.015"><properties><property name="score" value="0.016063622281986545" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i370]" time="0.024"><properties><property name="score" value="0.02672149667339628" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i371]" time="0.024"><properties><property name="score" value="0.06982424082465633" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i372]" time="0.027"><properties><property name="score" value="0.08467978601655546" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i373]" time="0.026"><properties><property name="score" value="0.006701074341036938" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i374]" time="0.015"><properties><property name="score" value="0.07380929602386899" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0738093&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0738093
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i375]" time="0.027"><properties><property name="score" value="0.01132120408358068" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0113212&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0113212
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i376]" time="0.024"><properties><property name="score" value="0.022902731104176878" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i377]" time="0.025"><properties><property name="score" value="0.0049804062889179995" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i378]" time="0.024"><properties><property name="score" value="0.02268253764312179" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i379]" time="0.025"><properties><property name="score" value="0.0668260924477872" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06682609&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06682609
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i380]" time="0.025"><properties><property name="score" value="0.012113227465416647" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i381]" time="0.016"><properties><property name="score" value="0.0332438917538036" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03324389&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03324389
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i382]" time="0.077"><properties><property name="score" value="0.005492867037820827" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00549287&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00549287
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i383]" time="0.028"><properties><property name="score" value="0.01359652412453467" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i384]" time="0.029"><properties><property name="score" value="0.042948977928572406" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04294898&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04294898
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i385]" time="0.015"><properties><property name="score" value="0.05407424959650853" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05407425&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05407425
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i386]" time="0.027"><properties><property name="score" value="0.03798895127756183" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03798895&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03798895
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i387]" time="0.033"><properties><property name="score" value="0.04555791393540165" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04555791&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04555791
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i388]" time="0.045"><properties><property name="score" value="0.002902518652579366" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00290252&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00290252
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i389]" time="0.031"><properties><property name="score" value="0.06438383099308821" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i390]" time="0.031"><properties><property name="score" value="0.03429084610979118" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i391]" time="0.095"><properties><property name="score" value="0.02011890793725517" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i392]" time="0.026"><properties><property name="score" value="0.07096003067337699" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07096003&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07096003
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i393]" time="0.034"><properties><property name="score" value="0.026482432611733928" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02648243&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02648243
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i394]" time="0.025"><properties><property name="score" value="0.0277903353805258" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02779034&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02779034
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i395]" time="0.028"><properties><property name="score" value="0.04797071479650978" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04797071&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04797071
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i396]" time="0.028"><properties><property name="score" value="0.006762184115269226" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00676218&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00676218
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i397]" time="0.040"><properties><property name="score" value="0.0036056981141405564" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i398]" time="0.030"><properties><property name="score" value="0.04481835648962838" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04481836&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04481836
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i399]" time="0.020"><properties><property name="score" value="0.052792732357509475" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05279273&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05279273
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i400]" time="0.029"><properties><property name="score" value="0.006467558223990881" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i401]" time="0.026"><properties><property name="score" value="0.01918700383983092" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.019187&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.019187
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i402]" time="0.024"><properties><property name="score" value="0.02109668910484705" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i403]" time="0.028"><properties><property name="score" value="0.04319342455365363" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i404]" time="0.029"><properties><property name="score" value="0.07965340677243259" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07965341&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07965341
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i405]" time="0.024"><properties><property name="score" value="0.03230863354657208" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03230863&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03230863
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i406]" time="0.023"><properties><property name="score" value="0.033610579813442454" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03361058&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03361058
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i407]" time="0.012"><properties><property name="score" value="0.0015161096650728734" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i408]" time="0.025"><properties><property name="score" value="0.009447202577640753" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0094472&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0094472
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i409]" time="0.027"><properties><property name="score" value="0.040434482712511544" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i410]" time="0.026"><properties><property name="score" value="0.006153970111897883" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00615397&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00615397
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i411]" time="0.077"><properties><property name="score" value="0.021969452487328293" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02196945&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02196945
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i412]" time="0.030"><properties><property name="score" value="0.010989987999278233" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i413]" time="0.028"><properties><property name="score" value="0.04283248813361633" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i414]" time="0.030"><properties><property name="score" value="0.0067303960854033455" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0067304&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0067304
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i415]" time="0.028"><properties><property name="score" value="0.031678591307555085" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i416]" time="0.034"><properties><property name="score" value="0.09232579793381802" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0923258&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0923258
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i417]" time="0.035"><properties><property name="score" value="0.05686195264372045" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05686195&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05686195
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i418]" time="0.033"><properties><property name="score" value="0.025919149631519013" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02591915&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02591915
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i419]" time="0.025"><properties><property name="score" value="0.00014668122655073744" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00014668&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00014668
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i420]" time="0.025"><properties><property name="score" value="0.003147015950927595" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00314702&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00314702
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i421]" time="0.025"><properties><property name="score" value="5.718463024447819e-07" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 5.7e-07&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 5.7e-07
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i422]" time="0.025"><properties><property name="score" value="0.02226579714416232" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i423]" time="0.027"><properties><property name="score" value="0.044642992262073564" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i424]" time="0.037"><properties><property name="score" value="0.014302655077110428" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01430266&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01430266
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i425]" time="0.031"><properties><property name="score" value="0.0072266521502207335" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i426]" time="0.013"><properties><property name="score" value="0.004855979616441575" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i427]" time="0.022"><properties><property name="score" value="0.05460810856059892" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05460811&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05460811
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i428]" time="0.023"><properties><property name="score" value="0.07627568322972002" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i429]" time="0.014"><properties><property name="score" value="0.04536172943914174" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04536173&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04536173
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i430]" time="0.023"><properties><property name="score" value="0.0200399402071392" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02003994&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02003994
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i431]" time="0.023"><properties><property name="score" value="0.024394115267806462" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02439412&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02439412
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i432]" time="0.023"><properties><property name="score" value="0.0031886633074636552" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00318866&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00318866
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i433]" time="0.022"><properties><property name="score" value="0.008811320553207054" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00881132&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00881132
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i434]" time="0.023"><properties><property name="score" value="0.004320087309712917" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00432009&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00432009
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i435]" time="0.012"><properties><property name="score" value="0.0020340146377306745" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i436]" time="0.021"><properties><property name="score" value="0.01222654625712027" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i437]" time="0.022"><properties><property name="score" value="0.02000969188486659" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i438]" time="0.012"><properties><property name="score" value="0.018096913181298024" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i439]" time="0.024"><properties><property name="score" value="0.0030496664548653007" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00304967&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00304967
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i440]" time="0.023"><properties><property name="score" value="0.032772727002531976" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i441]" time="0.022"><properties><property name="score" value="0.054814554900392706" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05481455&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05481455
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i442]" time="0.022"><properties><property name="score" value="0.043732544582458774" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04373254&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04373254
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i443]" time="0.023"><properties><property name="score" value="0.04633646673384721" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04633647&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04633647
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i444]" time="0.023"><properties><property name="score" value="0.07099807841234135" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07099808&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07099808
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i445]" time="0.077"><properties><property name="score" value="0.010670697604182123" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i446]" time="0.015"><properties><property name="score" value="0.01686239113836141" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01686239&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01686239
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i447]" time="0.024"><properties><property name="score" value="0.04264590587235629" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04264591&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04264591
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i448]" time="0.025"><properties><property name="score" value="0.05993189938326047" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0599319&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0599319
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i449]" time="0.028"><properties><property name="score" value="0.023585127949187643" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02358513&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02358513
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i450]" time="0.024"><properties><property name="score" value="0.023939871384036963" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i451]" time="0.012"><properties><property name="score" value="0.02210919378114462" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02210919&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02210919
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i452]" time="0.023"><properties><property name="score" value="0.005104978881662734" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00510498&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00510498
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i453]" time="0.023"><properties><property name="score" value="0.05035577517424143" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i454]" time="0.022"><properties><property name="score" value="0.05444375278668556" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i455]" time="0.026"><properties><property name="score" value="0.02214186288533299" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02214186&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02214186
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i456]" time="0.029"><properties><property name="score" value="0.040107202883861146" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i457]" time="0.023"><properties><property name="score" value="0.023202742925668995" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i458]" time="0.012"><properties><property name="score" value="0.0283729092477758" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i459]" time="0.023"><properties><property name="score" value="0.0700813335877187" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07008133&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07008133
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i460]" time="0.023"><properties><property name="score" value="0.025454492506418576" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i461]" time="0.022"><properties><property name="score" value="0.00021806874460340766" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00021807&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00021807
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i462]" time="0.069"><properties><property name="score" value="0.022252050516734224" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i463]" time="0.024"><properties><property name="score" value="0.03385324504200031" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03385325&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03385325
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i464]" time="0.023"><properties><property name="score" value="0.013633905156351166" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i465]" time="0.023"><properties><property name="score" value="0.03856315882866612" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03856316&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03856316
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i466]" time="0.023"><properties><property name="score" value="0.019617021996010522" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01961702&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01961702
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i467]" time="0.067"><properties><property name="score" value="0.03043373434608187" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i468]" time="0.023"><properties><property name="score" value="0.044377473369700646" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i469]" time="0.024"><properties><property name="score" value="0.036540775288898125" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03654078&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03654078
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i470]" time="0.024"><properties><property name="score" value="0.06407205800643573" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06407206&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06407206
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i471]" time="0.025"><properties><property name="score" value="0.016192455137674405" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i472]" time="0.015"><properties><property name="score" value="0.0011790097799668775" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00117901&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00117901
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i473]" time="0.023"><properties><property name="score" value="0.040886842305763234" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i474]" time="0.012"><properties><property name="score" value="0.024841417289028778" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i475]" time="0.025"><properties><property name="score" value="0.027297114588244048" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i476]" time="0.032"><properties><property name="score" value="0.05718252524757069" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05718253&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05718253
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i477]" time="0.026"><properties><property name="score" value="0.01779902143569554" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01779902&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01779902
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i478]" time="0.022"><properties><property name="score" value="0.0020015877701879668" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i479]" time="0.024"><properties><property name="score" value="0.0005525250677483294" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i480]" time="0.030"><properties><property name="score" value="0.013227918140218713" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01322792&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01322792
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i481]" time="0.027"><properties><property name="score" value="0.009633054030332584" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i482]" time="0.026"><properties><property name="score" value="0.028289065972975924" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02828907&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02828907
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i483]" time="0.037"><properties><property name="score" value="0.013533745273389883" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i484]" time="0.046"><properties><property name="score" value="0.010636463045513706" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01063646&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01063646
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i485]" time="0.038"><properties><property name="score" value="0.0043972964787555" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i486]" time="0.036"><properties><property name="score" value="0.004780173521193853" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i487]" time="0.018"><properties><property name="score" value="0.03203875053945371" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i488]" time="0.028"><properties><property name="score" value="0.02058645389141325" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02058645&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02058645
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i489]" time="0.026"><properties><property name="score" value="0.06630362090259243" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06630362&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.06630362
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i490]" time="0.025"><properties><property name="score" value="0.0424844284036777" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i491]" time="0.026"><properties><property name="score" value="0.02291620445907142" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0229162&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0229162
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i492]" time="0.029"><properties><property name="score" value="0.024560542834387935" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i493]" time="0.026"><properties><property name="score" value="0.0321181629792816" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03211816&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03211816
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i494]" time="0.027"><properties><property name="score" value="0.03901213564404257" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i495]" time="0.015"><properties><property name="score" value="0.00958407203143774" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00958407&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00958407
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i496]" time="0.040"><properties><property name="score" value="0.023059867250132447" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i497]" time="0.032"><properties><property name="score" value="0.08055154650871588" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.08055155&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.08055155
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i498]" time="0.089"><properties><property name="score" value="0.0665880305855379" /></properties></testcase><testcase classname="test_lzma_detect" name="test_chatgptnews_jsonl[i499]" time="0.036"><properties><property name="score" value="0.023457785446249324" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02345779&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.02345779
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i0]" time="0.013"><properties><property name="score" value="0.004446183419393179" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600003 [856] (title: An Improved Var) is a human-generated sample, misclassified as AI-generated with confidence 0.00444618&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600003 [856] (title: An Improved Var) is a human-generated sample, misclassified as AI-generated with confidence 0.00444618
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i1]" time="0.014"><properties><property name="score" value="0.031374053791916046" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600004 [909] (title: Mobile Robot Lo) is a human-generated sample, misclassified as AI-generated with confidence 0.03137405&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600004 [909] (title: Mobile Robot Lo) is a human-generated sample, misclassified as AI-generated with confidence 0.03137405
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i2]" time="0.013"><properties><property name="score" value="0.017291718094625264" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600008 [885] (title: Vertical Handof) is a human-generated sample, misclassified as AI-generated with confidence 0.01729172&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600008 [885] (title: Vertical Handof) is a human-generated sample, misclassified as AI-generated with confidence 0.01729172
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i3]" time="0.014"><properties><property name="score" value="0.0011093553392071964" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i4]" time="0.015"><properties><property name="score" value="0.017714595599843674" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600014 [800] (title: Gaussian MAC wi) is a human-generated sample, misclassified as AI-generated with confidence 0.0177146&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600014 [800] (title: Gaussian MAC wi) is a human-generated sample, misclassified as AI-generated with confidence 0.0177146
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i5]" time="0.021"><properties><property name="score" value="0.00461853911387089" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600018 [642] (title: Efficient Detec) is a human-generated sample, misclassified as AI-generated with confidence 0.00461854&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600018 [642] (title: Efficient Detec) is a human-generated sample, misclassified as AI-generated with confidence 0.00461854
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i6]" time="0.022"><properties><property name="score" value="0.09329819403008366" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600029 [1139] (title: A Social Bots D) is a human-generated sample, misclassified as AI-generated with confidence 0.09329819&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600029 [1139] (title: A Social Bots D) is a human-generated sample, misclassified as AI-generated with confidence 0.09329819
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i7]" time="0.014"><properties><property name="score" value="0.0014640775594709776" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600031 [727] (title: Energy-Efficien) is a human-generated sample, misclassified as AI-generated with confidence 0.00146408&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600031 [727] (title: Energy-Efficien) is a human-generated sample, misclassified as AI-generated with confidence 0.00146408
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i8]" time="0.015"><properties><property name="score" value="0.009518968339550238" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i9]" time="0.014"><properties><property name="score" value="0.052952460533689516" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600034 [1255] (title: Secrecy Anti-ja) is a human-generated sample, misclassified as AI-generated with confidence 0.05295246&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600034 [1255] (title: Secrecy Anti-ja) is a human-generated sample, misclassified as AI-generated with confidence 0.05295246
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i10]" time="0.016"><properties><property name="score" value="0.0007846309974857135" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600036 [1316] (title: Pedestrian Dete) is a human-generated sample, misclassified as AI-generated with confidence 0.00078463&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600036 [1316] (title: Pedestrian Dete) is a human-generated sample, misclassified as AI-generated with confidence 0.00078463
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i11]" time="0.016"><properties><property name="score" value="0.03258209841475179" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600037 [962] (title: Nonlinear State) is a human-generated sample, misclassified as AI-generated with confidence 0.0325821&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600037 [962] (title: Nonlinear State) is a human-generated sample, misclassified as AI-generated with confidence 0.0325821
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i12]" time="0.017"><properties><property name="score" value="0.008224822818475586" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600039 [886] (title: Unsupervised Eg) is a human-generated sample, misclassified as AI-generated with confidence 0.00822482&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600039 [886] (title: Unsupervised Eg) is a human-generated sample, misclassified as AI-generated with confidence 0.00822482
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i13]" time="0.014"><properties><property name="score" value="0.0098866243825102" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600042 [1259] (title: Large-Scale Ima) is a human-generated sample, misclassified as AI-generated with confidence 0.00988662&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600042 [1259] (title: Large-Scale Ima) is a human-generated sample, misclassified as AI-generated with confidence 0.00988662
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i14]" time="0.014"><properties><property name="score" value="0.007893446577184049" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600045 [897] (title: A Method Based ) is a human-generated sample, misclassified as AI-generated with confidence 0.00789345&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600045 [897] (title: A Method Based ) is a human-generated sample, misclassified as AI-generated with confidence 0.00789345
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i15]" time="0.017"><properties><property name="score" value="0.0822984144599248" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600048 [820] (title: Improved Data-A) is a human-generated sample, misclassified as AI-generated with confidence 0.08229841&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600048 [820] (title: Improved Data-A) is a human-generated sample, misclassified as AI-generated with confidence 0.08229841
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i16]" time="0.021"><properties><property name="score" value="0.07120075836218875" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600052 [1130] (title: Deep Neural Net) is a human-generated sample, misclassified as AI-generated with confidence 0.07120076&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600052 [1130] (title: Deep Neural Net) is a human-generated sample, misclassified as AI-generated with confidence 0.07120076
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i17]" time="0.015"><properties><property name="score" value="0.025474473747005844" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600055 [1394] (title: An Effective Se) is a human-generated sample, misclassified as AI-generated with confidence 0.02547447&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600055 [1394] (title: An Effective Se) is a human-generated sample, misclassified as AI-generated with confidence 0.02547447
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i18]" time="0.014"><properties><property name="score" value="0.007445006186956871" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i19]" time="0.014"><properties><property name="score" value="0.03944495287311889" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600060 [731] (title: Efficient Chann) is a human-generated sample, misclassified as AI-generated with confidence 0.03944495&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600060 [731] (title: Efficient Chann) is a human-generated sample, misclassified as AI-generated with confidence 0.03944495
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i20]" time="0.014"><properties><property name="score" value="0.0015401816787385947" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600061 [657] (title: Metric Learning) is a human-generated sample, misclassified as AI-generated with confidence 0.00154018&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600061 [657] (title: Metric Learning) is a human-generated sample, misclassified as AI-generated with confidence 0.00154018
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i21]" time="0.013"><properties><property name="score" value="0.04109832494342269" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600062 [699] (title: A Novel Route R) is a human-generated sample, misclassified as AI-generated with confidence 0.04109832&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600062 [699] (title: A Novel Route R) is a human-generated sample, misclassified as AI-generated with confidence 0.04109832
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i22]" time="0.018"><properties><property name="score" value="0.04162125494923119" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600065 [798] (title: Identification ) is a human-generated sample, misclassified as AI-generated with confidence 0.04162125&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600065 [798] (title: Identification ) is a human-generated sample, misclassified as AI-generated with confidence 0.04162125
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i23]" time="0.013"><properties><property name="score" value="0.015591817697896415" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i24]" time="0.035"><properties><property name="score" value="0.03922110837953117" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600072 [1639] (title: Analysis of Com) is a human-generated sample, misclassified as AI-generated with confidence 0.03922111&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600072 [1639] (title: Analysis of Com) is a human-generated sample, misclassified as AI-generated with confidence 0.03922111
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i25]" time="0.022"><properties><property name="score" value="0.03014489780779339" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600074 [1015] (title: Human Activity ) is a human-generated sample, misclassified as AI-generated with confidence 0.0301449&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600074 [1015] (title: Human Activity ) is a human-generated sample, misclassified as AI-generated with confidence 0.0301449
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i26]" time="0.015"><properties><property name="score" value="0.008081045529412245" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600079 [936] (title: Noise Suppressi) is a human-generated sample, misclassified as AI-generated with confidence 0.00808105&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600079 [936] (title: Noise Suppressi) is a human-generated sample, misclassified as AI-generated with confidence 0.00808105
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i27]" time="0.016"><properties><property name="score" value="0.004026507650395361" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600080 [1331] (title: Identification ) is a human-generated sample, misclassified as AI-generated with confidence 0.00402651&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600080 [1331] (title: Identification ) is a human-generated sample, misclassified as AI-generated with confidence 0.00402651
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i28]" time="0.013"><properties><property name="score" value="0.025263595346525225" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600081 [864] (title: Fast Projection) is a human-generated sample, misclassified as AI-generated with confidence 0.0252636&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600081 [864] (title: Fast Projection) is a human-generated sample, misclassified as AI-generated with confidence 0.0252636
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i29]" time="0.014"><properties><property name="score" value="0.016115892061119297" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600082 [1085] (title: Matching Game B) is a human-generated sample, misclassified as AI-generated with confidence 0.01611589&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600082 [1085] (title: Matching Game B) is a human-generated sample, misclassified as AI-generated with confidence 0.01611589
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i30]" time="0.013"><properties><property name="score" value="0.012462940021573665" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i31]" time="0.014"><properties><property name="score" value="0.02794198346454113" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600085 [855] (title: The Residual An) is a human-generated sample, misclassified as AI-generated with confidence 0.02794198&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600085 [855] (title: The Residual An) is a human-generated sample, misclassified as AI-generated with confidence 0.02794198
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i32]" time="0.014"><properties><property name="score" value="0.021397301200559715" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i33]" time="0.014"><properties><property name="score" value="0.029272482623865503" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600089 [934] (title: Modeling MME Re) is a human-generated sample, misclassified as AI-generated with confidence 0.02927248&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600089 [934] (title: Modeling MME Re) is a human-generated sample, misclassified as AI-generated with confidence 0.02927248
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i34]" time="0.013"><properties><property name="score" value="0.07677264798085504" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600090 [791] (title: Research on Int) is a human-generated sample, misclassified as AI-generated with confidence 0.07677265&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600090 [791] (title: Research on Int) is a human-generated sample, misclassified as AI-generated with confidence 0.07677265
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i35]" time="0.025"><properties><property name="score" value="0.023682981669389754" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600091 [1737] (title: Improved Epoch ) is a human-generated sample, misclassified as AI-generated with confidence 0.02368298&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600091 [1737] (title: Improved Epoch ) is a human-generated sample, misclassified as AI-generated with confidence 0.02368298
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i36]" time="0.018"><properties><property name="score" value="0.0010613146363802617" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600092 [1205] (title: Training-Based ) is a human-generated sample, misclassified as AI-generated with confidence 0.00106131&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600092 [1205] (title: Training-Based ) is a human-generated sample, misclassified as AI-generated with confidence 0.00106131
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i37]" time="0.019"><properties><property name="score" value="0.021000250011432575" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600094 [1397] (title: NLOS Detection ) is a human-generated sample, misclassified as AI-generated with confidence 0.02100025&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600094 [1397] (title: NLOS Detection ) is a human-generated sample, misclassified as AI-generated with confidence 0.02100025
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i38]" time="0.023"><properties><property name="score" value="0.05000264065796456" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600097 [1216] (title: Exploiting Part) is a human-generated sample, misclassified as AI-generated with confidence 0.05000264&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600097 [1216] (title: Exploiting Part) is a human-generated sample, misclassified as AI-generated with confidence 0.05000264
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i39]" time="0.023"><properties><property name="score" value="0.005028892632746862" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600098 [870] (title: Performance Com) is a human-generated sample, misclassified as AI-generated with confidence 0.00502889&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600098 [870] (title: Performance Com) is a human-generated sample, misclassified as AI-generated with confidence 0.00502889
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i40]" time="0.017"><properties><property name="score" value="0.09365074835105958" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600099 [1118] (title: Ranking-based C) is a human-generated sample, misclassified as AI-generated with confidence 0.09365075&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600099 [1118] (title: Ranking-based C) is a human-generated sample, misclassified as AI-generated with confidence 0.09365075
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i41]" time="0.013"><properties><property name="score" value="0.020399897366624176" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i42]" time="0.018"><properties><property name="score" value="0.07148905216446955" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600104 [907] (title: Network Intrusi) is a human-generated sample, misclassified as AI-generated with confidence 0.07148905&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600104 [907] (title: Network Intrusi) is a human-generated sample, misclassified as AI-generated with confidence 0.07148905
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i43]" time="0.019"><properties><property name="score" value="0.022500575123374222" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600106 [608] (title: Distributed Com) is a human-generated sample, misclassified as AI-generated with confidence 0.02250058&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600106 [608] (title: Distributed Com) is a human-generated sample, misclassified as AI-generated with confidence 0.02250058
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i44]" time="0.020"><properties><property name="score" value="0.04517370826938594" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600109 [1131] (title: Predicting Anch) is a human-generated sample, misclassified as AI-generated with confidence 0.04517371&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600109 [1131] (title: Predicting Anch) is a human-generated sample, misclassified as AI-generated with confidence 0.04517371
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i45]" time="0.018"><properties><property name="score" value="0.02773148737218456" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600111 [489] (title: A GMM-UBM Based) is a human-generated sample, misclassified as AI-generated with confidence 0.02773149&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600111 [489] (title: A GMM-UBM Based) is a human-generated sample, misclassified as AI-generated with confidence 0.02773149
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i46]" time="0.023"><properties><property name="score" value="0.06796581435831173" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600112 [1118] (title: Co-Prime Sampli) is a human-generated sample, misclassified as AI-generated with confidence 0.06796581&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600112 [1118] (title: Co-Prime Sampli) is a human-generated sample, misclassified as AI-generated with confidence 0.06796581
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i47]" time="0.023"><properties><property name="score" value="0.0011581746270916682" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600114 [590] (title: An Analysis of ) is a human-generated sample, misclassified as AI-generated with confidence 0.00115817&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600114 [590] (title: An Analysis of ) is a human-generated sample, misclassified as AI-generated with confidence 0.00115817
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i48]" time="0.019"><properties><property name="score" value="0.04952715750922154" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i49]" time="0.015"><properties><property name="score" value="0.015615005578883956" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600122 [829] (title: A Refined Phase) is a human-generated sample, misclassified as AI-generated with confidence 0.01561501&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600122 [829] (title: A Refined Phase) is a human-generated sample, misclassified as AI-generated with confidence 0.01561501
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i50]" time="0.014"><properties><property name="score" value="0.015191045117657342" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i51]" time="0.013"><properties><property name="score" value="0.009395461914263326" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i52]" time="0.013"><properties><property name="score" value="0.021488595792634735" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600126 [995] (title: A SDN-Based Hyb) is a human-generated sample, misclassified as AI-generated with confidence 0.0214886&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600126 [995] (title: A SDN-Based Hyb) is a human-generated sample, misclassified as AI-generated with confidence 0.0214886
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i53]" time="0.014"><properties><property name="score" value="0.01340762692203512" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600131 [1001] (title: Cell-Phone Iden) is a human-generated sample, misclassified as AI-generated with confidence 0.01340763&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600131 [1001] (title: Cell-Phone Iden) is a human-generated sample, misclassified as AI-generated with confidence 0.01340763
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i54]" time="0.019"><properties><property name="score" value="0.00894439176605255" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i55]" time="0.023"><properties><property name="score" value="0.01938803793047339" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600136 [694] (title: Mixing Matrix E) is a human-generated sample, misclassified as AI-generated with confidence 0.01938804&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600136 [694] (title: Mixing Matrix E) is a human-generated sample, misclassified as AI-generated with confidence 0.01938804
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i56]" time="0.022"><properties><property name="score" value="0.037711364927761615" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600137 [998] (title: Review of TDMA-) is a human-generated sample, misclassified as AI-generated with confidence 0.03771136&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600137 [998] (title: Review of TDMA-) is a human-generated sample, misclassified as AI-generated with confidence 0.03771136
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i57]" time="0.020"><properties><property name="score" value="0.022097231345202806" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600139 [942] (title: Lightweight Gro) is a human-generated sample, misclassified as AI-generated with confidence 0.02209723&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600139 [942] (title: Lightweight Gro) is a human-generated sample, misclassified as AI-generated with confidence 0.02209723
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i58]" time="0.017"><properties><property name="score" value="0.04291788807118668" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i59]" time="0.021"><properties><property name="score" value="0.04183262630516604" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600143 [1271] (title: Balanced Use of) is a human-generated sample, misclassified as AI-generated with confidence 0.04183263&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600143 [1271] (title: Balanced Use of) is a human-generated sample, misclassified as AI-generated with confidence 0.04183263
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i60]" time="0.019"><properties><property name="score" value="0.04026070903122014" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600145 [887] (title: Comparative Ana) is a human-generated sample, misclassified as AI-generated with confidence 0.04026071&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600145 [887] (title: Comparative Ana) is a human-generated sample, misclassified as AI-generated with confidence 0.04026071
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i61]" time="0.016"><properties><property name="score" value="0.001553544881960356" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600146 [974] (title: Machine Fault D) is a human-generated sample, misclassified as AI-generated with confidence 0.00155354&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600146 [974] (title: Machine Fault D) is a human-generated sample, misclassified as AI-generated with confidence 0.00155354
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i62]" time="0.016"><properties><property name="score" value="0.00888239541774194" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i63]" time="0.014"><properties><property name="score" value="0.006201329166355407" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i64]" time="0.014"><properties><property name="score" value="0.04151297902577067" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600150 [867] (title: User Associatio) is a human-generated sample, misclassified as AI-generated with confidence 0.04151298&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600150 [867] (title: User Associatio) is a human-generated sample, misclassified as AI-generated with confidence 0.04151298
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i65]" time="0.013"><properties><property name="score" value="0.03566630086911937" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600153 [1380] (title: A New Combinato) is a human-generated sample, misclassified as AI-generated with confidence 0.0356663&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600153 [1380] (title: A New Combinato) is a human-generated sample, misclassified as AI-generated with confidence 0.0356663
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i66]" time="0.015"><properties><property name="score" value="0.055664287511386634" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600154 [719] (title: Prediction of H) is a human-generated sample, misclassified as AI-generated with confidence 0.05566429&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600154 [719] (title: Prediction of H) is a human-generated sample, misclassified as AI-generated with confidence 0.05566429
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i67]" time="0.018"><properties><property name="score" value="0.021728614408517144" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600156 [1201] (title: A Novel Cluster) is a human-generated sample, misclassified as AI-generated with confidence 0.02172861&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600156 [1201] (title: A Novel Cluster) is a human-generated sample, misclassified as AI-generated with confidence 0.02172861
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i68]" time="0.015"><properties><property name="score" value="0.01710922055184194" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i69]" time="0.016"><properties><property name="score" value="0.062097165177715796" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600165 [1193] (title: Adaptive Unequa) is a human-generated sample, misclassified as AI-generated with confidence 0.06209717&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600165 [1193] (title: Adaptive Unequa) is a human-generated sample, misclassified as AI-generated with confidence 0.06209717
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i70]" time="0.014"><properties><property name="score" value="0.0012325577435967627" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i71]" time="0.014"><properties><property name="score" value="0.009596360217561939" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600169 [1121] (title: LayerOS: Schedu) is a human-generated sample, misclassified as AI-generated with confidence 0.00959636&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600169 [1121] (title: LayerOS: Schedu) is a human-generated sample, misclassified as AI-generated with confidence 0.00959636
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i72]" time="0.014"><properties><property name="score" value="0.029604067750382823" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i73]" time="0.016"><properties><property name="score" value="0.030699668190165008" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600178 [956] (title: Cost Effective ) is a human-generated sample, misclassified as AI-generated with confidence 0.03069967&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600178 [956] (title: Cost Effective ) is a human-generated sample, misclassified as AI-generated with confidence 0.03069967
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i74]" time="0.015"><properties><property name="score" value="0.01942202543304683" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600186 [954] (title: Atmospheric Abs) is a human-generated sample, misclassified as AI-generated with confidence 0.01942203&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600186 [954] (title: Atmospheric Abs) is a human-generated sample, misclassified as AI-generated with confidence 0.01942203
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i75]" time="0.014"><properties><property name="score" value="0.03775793802612837" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600189 [924] (title: A Design of Cro) is a human-generated sample, misclassified as AI-generated with confidence 0.03775794&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600189 [924] (title: A Design of Cro) is a human-generated sample, misclassified as AI-generated with confidence 0.03775794
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i76]" time="0.027"><properties><property name="score" value="0.03890583898620769" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600190 [1608] (title: Manner of Artic) is a human-generated sample, misclassified as AI-generated with confidence 0.03890584&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600190 [1608] (title: Manner of Artic) is a human-generated sample, misclassified as AI-generated with confidence 0.03890584
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i77]" time="0.012"><properties><property name="score" value="0.0119833531782354" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600195 [745] (title: Dictionary lear) is a human-generated sample, misclassified as AI-generated with confidence 0.01198335&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600195 [745] (title: Dictionary lear) is a human-generated sample, misclassified as AI-generated with confidence 0.01198335
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i78]" time="0.013"><properties><property name="score" value="0.016764310768802027" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i79]" time="0.012"><properties><property name="score" value="0.003905390787706242" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i80]" time="0.014"><properties><property name="score" value="0.027867318350410786" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600199 [1016] (title: Range Free Loca) is a human-generated sample, misclassified as AI-generated with confidence 0.02786732&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600199 [1016] (title: Range Free Loca) is a human-generated sample, misclassified as AI-generated with confidence 0.02786732
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i81]" time="0.015"><properties><property name="score" value="0.0026282717651648912" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600200 [1162] (title: Exploiting Clas) is a human-generated sample, misclassified as AI-generated with confidence 0.00262827&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600200 [1162] (title: Exploiting Clas) is a human-generated sample, misclassified as AI-generated with confidence 0.00262827
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i82]" time="0.021"><properties><property name="score" value="0.004055658943441509" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i83]" time="0.018"><properties><property name="score" value="0.01782979543537011" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i84]" time="0.014"><properties><property name="score" value="0.03497825640810248" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600205 [993] (title: Vehicles to Ped) is a human-generated sample, misclassified as AI-generated with confidence 0.03497826&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600205 [993] (title: Vehicles to Ped) is a human-generated sample, misclassified as AI-generated with confidence 0.03497826
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i85]" time="0.012"><properties><property name="score" value="0.020136093270067823" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i86]" time="0.014"><properties><property name="score" value="0.027778006131584654" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i87]" time="0.012"><properties><property name="score" value="0.05872871912186439" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600209 [885] (title: Real-time evalu) is a human-generated sample, misclassified as AI-generated with confidence 0.05872872&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600209 [885] (title: Real-time evalu) is a human-generated sample, misclassified as AI-generated with confidence 0.05872872
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i88]" time="0.014"><properties><property name="score" value="0.014565450528031265" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600210 [1049] (title: A Fingerprint L) is a human-generated sample, misclassified as AI-generated with confidence 0.01456545&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600210 [1049] (title: A Fingerprint L) is a human-generated sample, misclassified as AI-generated with confidence 0.01456545
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i89]" time="0.012"><properties><property name="score" value="0.010441887643558578" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i90]" time="0.014"><properties><property name="score" value="0.016317557814105044" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600214 [873] (title: Study on a Deci) is a human-generated sample, misclassified as AI-generated with confidence 0.01631756&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600214 [873] (title: Study on a Deci) is a human-generated sample, misclassified as AI-generated with confidence 0.01631756
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i91]" time="0.013"><properties><property name="score" value="0.030223419384300954" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600216 [609] (title: Research on the) is a human-generated sample, misclassified as AI-generated with confidence 0.03022342&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600216 [609] (title: Research on the) is a human-generated sample, misclassified as AI-generated with confidence 0.03022342
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i92]" time="0.013"><properties><property name="score" value="0.06059477828301296" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600218 [985] (title: A Reinforcement) is a human-generated sample, misclassified as AI-generated with confidence 0.06059478&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600218 [985] (title: A Reinforcement) is a human-generated sample, misclassified as AI-generated with confidence 0.06059478
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i93]" time="0.018"><properties><property name="score" value="0.020516401740372503" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600219 [985] (title: Deep Learning I) is a human-generated sample, misclassified as AI-generated with confidence 0.0205164&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600219 [985] (title: Deep Learning I) is a human-generated sample, misclassified as AI-generated with confidence 0.0205164
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i94]" time="0.020"><properties><property name="score" value="0.019510294923336113" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600223 [717] (title: Coverage Hole D) is a human-generated sample, misclassified as AI-generated with confidence 0.01951029&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600223 [717] (title: Coverage Hole D) is a human-generated sample, misclassified as AI-generated with confidence 0.01951029
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i95]" time="0.019"><properties><property name="score" value="0.05017697994837422" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600225 [969] (title: Quantum Key Int) is a human-generated sample, misclassified as AI-generated with confidence 0.05017698&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600225 [969] (title: Quantum Key Int) is a human-generated sample, misclassified as AI-generated with confidence 0.05017698
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i96]" time="0.018"><properties><property name="score" value="0.008135510824325287" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i97]" time="0.021"><properties><property name="score" value="0.050282203533053016" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600231 [1432] (title: Uplink Asynchro) is a human-generated sample, misclassified as AI-generated with confidence 0.0502822&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600231 [1432] (title: Uplink Asynchro) is a human-generated sample, misclassified as AI-generated with confidence 0.0502822
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i98]" time="0.020"><properties><property name="score" value="0.06093952101392519" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600233 [961] (title: Power System Fr) is a human-generated sample, misclassified as AI-generated with confidence 0.06093952&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600233 [961] (title: Power System Fr) is a human-generated sample, misclassified as AI-generated with confidence 0.06093952
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i99]" time="0.021"><properties><property name="score" value="0.0623202870941697" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600234 [880] (title: Association Ana) is a human-generated sample, misclassified as AI-generated with confidence 0.06232029&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600234 [880] (title: Association Ana) is a human-generated sample, misclassified as AI-generated with confidence 0.06232029
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i100]" time="0.016"><properties><property name="score" value="0.022087895462674645" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600237 [1432] (title: The Adaptive Co) is a human-generated sample, misclassified as AI-generated with confidence 0.0220879&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600237 [1432] (title: The Adaptive Co) is a human-generated sample, misclassified as AI-generated with confidence 0.0220879
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i101]" time="0.012"><properties><property name="score" value="0.040153641246964966" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600239 [953] (title: Towards Multi-F) is a human-generated sample, misclassified as AI-generated with confidence 0.04015364&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600239 [953] (title: Towards Multi-F) is a human-generated sample, misclassified as AI-generated with confidence 0.04015364
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i102]" time="0.013"><properties><property name="score" value="0.004857951450609743" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i103]" time="0.024"><properties><property name="score" value="0.01900120712547415" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600243 [1197] (title: A Frequency Ass) is a human-generated sample, misclassified as AI-generated with confidence 0.01900121&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600243 [1197] (title: A Frequency Ass) is a human-generated sample, misclassified as AI-generated with confidence 0.01900121
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i104]" time="0.013"><properties><property name="score" value="0.0027519101154172" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i105]" time="0.012"><properties><property name="score" value="0.06623984450789822" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600246 [938] (title: Dynamic Spectru) is a human-generated sample, misclassified as AI-generated with confidence 0.06623984&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600246 [938] (title: Dynamic Spectru) is a human-generated sample, misclassified as AI-generated with confidence 0.06623984
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i106]" time="0.016"><properties><property name="score" value="0.025279354649521135" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600252 [1114] (title: Multi-objective) is a human-generated sample, misclassified as AI-generated with confidence 0.02527935&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600252 [1114] (title: Multi-objective) is a human-generated sample, misclassified as AI-generated with confidence 0.02527935
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i107]" time="0.015"><properties><property name="score" value="0.01083047851999086" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i108]" time="0.015"><properties><property name="score" value="0.01387403719585456" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600255 [758] (title: Vehicular Clust) is a human-generated sample, misclassified as AI-generated with confidence 0.01387404&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600255 [758] (title: Vehicular Clust) is a human-generated sample, misclassified as AI-generated with confidence 0.01387404
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i109]" time="0.014"><properties><property name="score" value="0.08814013110148289" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600263 [1009] (title: Two-Layer Resou) is a human-generated sample, misclassified as AI-generated with confidence 0.08814013&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600263 [1009] (title: Two-Layer Resou) is a human-generated sample, misclassified as AI-generated with confidence 0.08814013
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i110]" time="0.018"><properties><property name="score" value="0.006164263284119675" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600265 [752] (title: A Network Selec) is a human-generated sample, misclassified as AI-generated with confidence 0.00616426&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600265 [752] (title: A Network Selec) is a human-generated sample, misclassified as AI-generated with confidence 0.00616426
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i111]" time="0.014"><properties><property name="score" value="0.004082391569681931" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600267 [707] (title: Approaches to C) is a human-generated sample, misclassified as AI-generated with confidence 0.00408239&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600267 [707] (title: Approaches to C) is a human-generated sample, misclassified as AI-generated with confidence 0.00408239
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i112]" time="0.014"><properties><property name="score" value="0.007053742606732971" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i113]" time="0.015"><properties><property name="score" value="0.02163464182277086" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600272 [1430] (title: Time-Sensitive ) is a human-generated sample, misclassified as AI-generated with confidence 0.02163464&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600272 [1430] (title: Time-Sensitive ) is a human-generated sample, misclassified as AI-generated with confidence 0.02163464
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i114]" time="0.015"><properties><property name="score" value="0.004950180852913988" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600275 [936] (title: Hellinger Dista) is a human-generated sample, misclassified as AI-generated with confidence 0.00495018&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600275 [936] (title: Hellinger Dista) is a human-generated sample, misclassified as AI-generated with confidence 0.00495018
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i115]" time="0.024"><properties><property name="score" value="0.02612925810506212" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600298 [1545] (title: Toward Secure a) is a human-generated sample, misclassified as AI-generated with confidence 0.02612926&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600298 [1545] (title: Toward Secure a) is a human-generated sample, misclassified as AI-generated with confidence 0.02612926
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i116]" time="0.016"><properties><property name="score" value="0.034321525714420265" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600301 [1456] (title: Coastal SLAM Wi) is a human-generated sample, misclassified as AI-generated with confidence 0.03432153&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600301 [1456] (title: Coastal SLAM Wi) is a human-generated sample, misclassified as AI-generated with confidence 0.03432153
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i117]" time="0.026"><properties><property name="score" value="0.011660517896032802" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600305 [1650] (title: Data Security S) is a human-generated sample, misclassified as AI-generated with confidence 0.01166052&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600305 [1650] (title: Data Security S) is a human-generated sample, misclassified as AI-generated with confidence 0.01166052
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i118]" time="0.033"><properties><property name="score" value="0.07196882757110928" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600306 [1572] (title: A Two-Stage Att) is a human-generated sample, misclassified as AI-generated with confidence 0.07196883&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600306 [1572] (title: A Two-Stage Att) is a human-generated sample, misclassified as AI-generated with confidence 0.07196883
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i119]" time="0.015"><properties><property name="score" value="0.010898685017335374" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i120]" time="0.026"><properties><property name="score" value="0.028075648169472472" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600316 [1592] (title: Deep Learning f) is a human-generated sample, misclassified as AI-generated with confidence 0.02807565&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600316 [1592] (title: Deep Learning f) is a human-generated sample, misclassified as AI-generated with confidence 0.02807565
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i121]" time="0.031"><properties><property name="score" value="0.0524792725409922" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600317 [1781] (title: Enabling Reprod) is a human-generated sample, misclassified as AI-generated with confidence 0.05247927&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600317 [1781] (title: Enabling Reprod) is a human-generated sample, misclassified as AI-generated with confidence 0.05247927
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i122]" time="0.021"><properties><property name="score" value="0.13785627851398696" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600320 [1497] (title: Multiview Trans) is a human-generated sample, misclassified as AI-generated with confidence 0.13785628&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600320 [1497] (title: Multiview Trans) is a human-generated sample, misclassified as AI-generated with confidence 0.13785628
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i123]" time="0.018"><properties><property name="score" value="0.05839827646823004" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600326 [1262] (title: Uplink Time Sch) is a human-generated sample, misclassified as AI-generated with confidence 0.05839828&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600326 [1262] (title: Uplink Time Sch) is a human-generated sample, misclassified as AI-generated with confidence 0.05839828
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i124]" time="0.018"><properties><property name="score" value="0.030283874108993913" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600329 [1382] (title: Texture Feature) is a human-generated sample, misclassified as AI-generated with confidence 0.03028387&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600329 [1382] (title: Texture Feature) is a human-generated sample, misclassified as AI-generated with confidence 0.03028387
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i125]" time="0.017"><properties><property name="score" value="0.05888464982650721" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600330 [1005] (title: Finite-State Ap) is a human-generated sample, misclassified as AI-generated with confidence 0.05888465&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600330 [1005] (title: Finite-State Ap) is a human-generated sample, misclassified as AI-generated with confidence 0.05888465
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i126]" time="0.024"><properties><property name="score" value="0.005216553315818066" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i127]" time="0.016"><properties><property name="score" value="0.047671074170663186" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600335 [922] (title: A Deep Learning) is a human-generated sample, misclassified as AI-generated with confidence 0.04767107&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600335 [922] (title: A Deep Learning) is a human-generated sample, misclassified as AI-generated with confidence 0.04767107
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i128]" time="0.014"><properties><property name="score" value="0.05888562522270435" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600338 [1456] (title: Jointly Learnin) is a human-generated sample, misclassified as AI-generated with confidence 0.05888563&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600338 [1456] (title: Jointly Learnin) is a human-generated sample, misclassified as AI-generated with confidence 0.05888563
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i129]" time="0.027"><properties><property name="score" value="0.03581920785870586" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i130]" time="0.013"><properties><property name="score" value="0.0050959714547382795" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i131]" time="0.013"><properties><property name="score" value="0.03206188148174305" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600351 [1202] (title: Machine Learnin) is a human-generated sample, misclassified as AI-generated with confidence 0.03206188&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600351 [1202] (title: Machine Learnin) is a human-generated sample, misclassified as AI-generated with confidence 0.03206188
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i132]" time="0.013"><properties><property name="score" value="0.015122343537299354" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i133]" time="0.013"><properties><property name="score" value="0.08750841649953212" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600359 [1248] (title: Supervised Spar) is a human-generated sample, misclassified as AI-generated with confidence 0.08750842&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600359 [1248] (title: Supervised Spar) is a human-generated sample, misclassified as AI-generated with confidence 0.08750842
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i134]" time="0.015"><properties><property name="score" value="0.009720638662502612" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600361 [1144] (title: Communication N) is a human-generated sample, misclassified as AI-generated with confidence 0.00972064&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600361 [1144] (title: Communication N) is a human-generated sample, misclassified as AI-generated with confidence 0.00972064
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i135]" time="0.031"><properties><property name="score" value="0.030632150414196483" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600363 [2064] (title: A Hierarchical ) is a human-generated sample, misclassified as AI-generated with confidence 0.03063215&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600363 [2064] (title: A Hierarchical ) is a human-generated sample, misclassified as AI-generated with confidence 0.03063215
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i136]" time="0.014"><properties><property name="score" value="0.07983137570218846" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600369 [1149] (title: Local Discrimin) is a human-generated sample, misclassified as AI-generated with confidence 0.07983138&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600369 [1149] (title: Local Discrimin) is a human-generated sample, misclassified as AI-generated with confidence 0.07983138
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i137]" time="0.012"><properties><property name="score" value="0.07150351366755348" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600370 [1482] (title: On Low-Resoluti) is a human-generated sample, misclassified as AI-generated with confidence 0.07150351&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600370 [1482] (title: On Low-Resoluti) is a human-generated sample, misclassified as AI-generated with confidence 0.07150351
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i138]" time="0.014"><properties><property name="score" value="0.020599989845371258" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i139]" time="0.014"><properties><property name="score" value="0.006712636052097487" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i140]" time="0.019"><properties><property name="score" value="0.012343520265462038" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600373 [1293] (title: Lightweight Fea) is a human-generated sample, misclassified as AI-generated with confidence 0.01234352&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600373 [1293] (title: Lightweight Fea) is a human-generated sample, misclassified as AI-generated with confidence 0.01234352
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i141]" time="0.027"><properties><property name="score" value="0.06310960824935828" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i142]" time="0.028"><properties><property name="score" value="0.003474156592407505" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i143]" time="0.017"><properties><property name="score" value="0.04546712336778169" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600382 [1363] (title: A Deep Reinforc) is a human-generated sample, misclassified as AI-generated with confidence 0.04546712&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600382 [1363] (title: A Deep Reinforc) is a human-generated sample, misclassified as AI-generated with confidence 0.04546712
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i144]" time="0.019"><properties><property name="score" value="0.04488696612067189" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600384 [1076] (title: A Generative Di) is a human-generated sample, misclassified as AI-generated with confidence 0.04488697&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600384 [1076] (title: A Generative Di) is a human-generated sample, misclassified as AI-generated with confidence 0.04488697
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i145]" time="0.014"><properties><property name="score" value="0.019019977587075854" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i146]" time="0.018"><properties><property name="score" value="0.006608657269674412" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600392 [1388] (title: Low Rank and St) is a human-generated sample, misclassified as AI-generated with confidence 0.00660866&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600392 [1388] (title: Low Rank and St) is a human-generated sample, misclassified as AI-generated with confidence 0.00660866
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i147]" time="0.015"><properties><property name="score" value="0.006364325815039207" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i148]" time="0.015"><properties><property name="score" value="0.020821180504798642" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600395 [711] (title: Sensor Localiza) is a human-generated sample, misclassified as AI-generated with confidence 0.02082118&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600395 [711] (title: Sensor Localiza) is a human-generated sample, misclassified as AI-generated with confidence 0.02082118
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i149]" time="0.015"><properties><property name="score" value="0.04909494806107917" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600396 [1029] (title: Real-Time Capac) is a human-generated sample, misclassified as AI-generated with confidence 0.04909495&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600396 [1029] (title: Real-Time Capac) is a human-generated sample, misclassified as AI-generated with confidence 0.04909495
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i150]" time="0.018"><properties><property name="score" value="0.00103225185625333" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i151]" time="0.013"><properties><property name="score" value="0.0030868778040149536" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i152]" time="0.014"><properties><property name="score" value="0.01835116185719965" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i153]" time="0.014"><properties><property name="score" value="0.0001327134919903994" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i154]" time="0.014"><properties><property name="score" value="0.032808680226231735" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i155]" time="0.015"><properties><property name="score" value="0.01185440327120535" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600425 [969] (title: An Efficient Es) is a human-generated sample, misclassified as AI-generated with confidence 0.0118544&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600425 [969] (title: An Efficient Es) is a human-generated sample, misclassified as AI-generated with confidence 0.0118544
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i156]" time="0.017"><properties><property name="score" value="0.02829160488652116" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600428 [1340] (title: An Adaptive Rat) is a human-generated sample, misclassified as AI-generated with confidence 0.0282916&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600428 [1340] (title: An Adaptive Rat) is a human-generated sample, misclassified as AI-generated with confidence 0.0282916
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i157]" time="0.016"><properties><property name="score" value="0.022976508434707776" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600431 [1019] (title: Rotation and Pr) is a human-generated sample, misclassified as AI-generated with confidence 0.02297651&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600431 [1019] (title: Rotation and Pr) is a human-generated sample, misclassified as AI-generated with confidence 0.02297651
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i158]" time="0.015"><properties><property name="score" value="0.00030551029609471314" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600432 [604] (title: The Two-Directi) is a human-generated sample, misclassified as AI-generated with confidence 0.00030551&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600432 [604] (title: The Two-Directi) is a human-generated sample, misclassified as AI-generated with confidence 0.00030551
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i159]" time="0.020"><properties><property name="score" value="0.006980609726731091" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600437 [904] (title: Visualization o) is a human-generated sample, misclassified as AI-generated with confidence 0.00698061&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600437 [904] (title: Visualization o) is a human-generated sample, misclassified as AI-generated with confidence 0.00698061
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i160]" time="0.018"><properties><property name="score" value="0.03748403032539649" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i161]" time="0.015"><properties><property name="score" value="0.01580451326174681" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i162]" time="0.017"><properties><property name="score" value="0.01092486348409083" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600443 [558] (title: Review of Resea) is a human-generated sample, misclassified as AI-generated with confidence 0.01092486&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600443 [558] (title: Review of Resea) is a human-generated sample, misclassified as AI-generated with confidence 0.01092486
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i163]" time="0.019"><properties><property name="score" value="0.012655474894152086" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i164]" time="0.013"><properties><property name="score" value="0.017832551000573904" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i165]" time="0.015"><properties><property name="score" value="0.055011627901147" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600447 [1036] (title: Tone Correction) is a human-generated sample, misclassified as AI-generated with confidence 0.05501163&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600447 [1036] (title: Tone Correction) is a human-generated sample, misclassified as AI-generated with confidence 0.05501163
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i166]" time="0.013"><properties><property name="score" value="0.04102105349190199" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i167]" time="0.019"><properties><property name="score" value="0.03475948033856424" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600457 [1373] (title: A Generic Appro) is a human-generated sample, misclassified as AI-generated with confidence 0.03475948&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600457 [1373] (title: A Generic Appro) is a human-generated sample, misclassified as AI-generated with confidence 0.03475948
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i168]" time="0.034"><properties><property name="score" value="0.14122976458451866" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600458 [1695] (title: A Fast CTU Dept) is a human-generated sample, misclassified as AI-generated with confidence 0.14122976&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600458 [1695] (title: A Fast CTU Dept) is a human-generated sample, misclassified as AI-generated with confidence 0.14122976
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i169]" time="0.012"><properties><property name="score" value="0.012524840192995601" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i170]" time="0.015"><properties><property name="score" value="0.03853052601738849" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i171]" time="0.017"><properties><property name="score" value="0.008936481653498696" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i172]" time="0.016"><properties><property name="score" value="0.015183679206359768" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600471 [751] (title: An Optimized Me) is a human-generated sample, misclassified as AI-generated with confidence 0.01518368&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600471 [751] (title: An Optimized Me) is a human-generated sample, misclassified as AI-generated with confidence 0.01518368
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i173]" time="0.013"><properties><property name="score" value="0.022044288122763378" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i174]" time="0.016"><properties><property name="score" value="0.023713026905952672" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600475 [826] (title: A Cascade Calib) is a human-generated sample, misclassified as AI-generated with confidence 0.02371303&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600475 [826] (title: A Cascade Calib) is a human-generated sample, misclassified as AI-generated with confidence 0.02371303
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i175]" time="0.013"><properties><property name="score" value="0.07291098478124924" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600476 [1118] (title: An Improved MMS) is a human-generated sample, misclassified as AI-generated with confidence 0.07291098&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600476 [1118] (title: An Improved MMS) is a human-generated sample, misclassified as AI-generated with confidence 0.07291098
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i176]" time="0.015"><properties><property name="score" value="0.029925938025504584" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600483 [813] (title: Abnormal Human ) is a human-generated sample, misclassified as AI-generated with confidence 0.02992594&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600483 [813] (title: Abnormal Human ) is a human-generated sample, misclassified as AI-generated with confidence 0.02992594
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i177]" time="0.013"><properties><property name="score" value="0.03505180349340775" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600484 [1102] (title: A Modeling Atta) is a human-generated sample, misclassified as AI-generated with confidence 0.0350518&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600484 [1102] (title: A Modeling Atta) is a human-generated sample, misclassified as AI-generated with confidence 0.0350518
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i178]" time="0.014"><properties><property name="score" value="0.02789757145402194" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i179]" time="0.015"><properties><property name="score" value="0.07568577857872127" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600486 [1312] (title: A Fast Multi-Ma) is a human-generated sample, misclassified as AI-generated with confidence 0.07568578&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600486 [1312] (title: A Fast Multi-Ma) is a human-generated sample, misclassified as AI-generated with confidence 0.07568578
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i180]" time="0.016"><properties><property name="score" value="0.027605383242051973" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600489 [882] (title: Research on Fun) is a human-generated sample, misclassified as AI-generated with confidence 0.02760538&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600489 [882] (title: Research on Fun) is a human-generated sample, misclassified as AI-generated with confidence 0.02760538
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i181]" time="0.015"><properties><property name="score" value="0.002899873938488895" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600492 [735] (title: A Two-frequency) is a human-generated sample, misclassified as AI-generated with confidence 0.00289987&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600492 [735] (title: A Two-frequency) is a human-generated sample, misclassified as AI-generated with confidence 0.00289987
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i182]" time="0.014"><properties><property name="score" value="0.00287701448547506" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i183]" time="0.014"><properties><property name="score" value="0.07170446169473266" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600496 [952] (title: Reliability Eva) is a human-generated sample, misclassified as AI-generated with confidence 0.07170446&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600496 [952] (title: Reliability Eva) is a human-generated sample, misclassified as AI-generated with confidence 0.07170446
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i184]" time="0.014"><properties><property name="score" value="0.018809274069392368" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i185]" time="0.014"><properties><property name="score" value="0.01657916816215077" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600501 [1252] (title: Multi-frame Ima) is a human-generated sample, misclassified as AI-generated with confidence 0.01657917&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600501 [1252] (title: Multi-frame Ima) is a human-generated sample, misclassified as AI-generated with confidence 0.01657917
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i186]" time="0.014"><properties><property name="score" value="0.003392033712812381" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i187]" time="0.016"><properties><property name="score" value="0.01213262904850021" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i188]" time="0.032"><properties><property name="score" value="0.0018621345154466603" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600505 [1747] (title: The Event-Drive) is a human-generated sample, misclassified as AI-generated with confidence 0.00186213&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600505 [1747] (title: The Event-Drive) is a human-generated sample, misclassified as AI-generated with confidence 0.00186213
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i189]" time="0.015"><properties><property name="score" value="0.0019492668965614968" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i190]" time="0.029"><properties><property name="score" value="0.0208374501427198" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i191]" time="0.017"><properties><property name="score" value="0.041040326136698724" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600509 [1294] (title: The Detection o) is a human-generated sample, misclassified as AI-generated with confidence 0.04104033&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600509 [1294] (title: The Detection o) is a human-generated sample, misclassified as AI-generated with confidence 0.04104033
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i192]" time="0.015"><properties><property name="score" value="0.011416086818349935" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i193]" time="0.015"><properties><property name="score" value="0.00963652852521979" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600514 [975] (title: Patch-Based App) is a human-generated sample, misclassified as AI-generated with confidence 0.00963653&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600514 [975] (title: Patch-Based App) is a human-generated sample, misclassified as AI-generated with confidence 0.00963653
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i194]" time="0.020"><properties><property name="score" value="0.012112999955560255" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600515 [527] (title: Image Deblurrin) is a human-generated sample, misclassified as AI-generated with confidence 0.012113&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600515 [527] (title: Image Deblurrin) is a human-generated sample, misclassified as AI-generated with confidence 0.012113
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i195]" time="0.019"><properties><property name="score" value="0.009987670064554033" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600516 [630] (title: Detecting Adver) is a human-generated sample, misclassified as AI-generated with confidence 0.00998767&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600516 [630] (title: Detecting Adver) is a human-generated sample, misclassified as AI-generated with confidence 0.00998767
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i196]" time="0.019"><properties><property name="score" value="0.03173411592650166" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600517 [728] (title: A Pre-Compensat) is a human-generated sample, misclassified as AI-generated with confidence 0.03173412&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600517 [728] (title: A Pre-Compensat) is a human-generated sample, misclassified as AI-generated with confidence 0.03173412
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i197]" time="0.023"><properties><property name="score" value="0.008623769283083726" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600519 [1124] (title: Outage Performa) is a human-generated sample, misclassified as AI-generated with confidence 0.00862377&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600519 [1124] (title: Outage Performa) is a human-generated sample, misclassified as AI-generated with confidence 0.00862377
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i198]" time="0.022"><properties><property name="score" value="0.027875016652784135" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600520 [1264] (title: Spacecraft Dete) is a human-generated sample, misclassified as AI-generated with confidence 0.02787502&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600520 [1264] (title: Spacecraft Dete) is a human-generated sample, misclassified as AI-generated with confidence 0.02787502
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i199]" time="0.013"><properties><property name="score" value="0.03373407132777895" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i200]" time="0.012"><properties><property name="score" value="0.018656917983667087" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i201]" time="0.013"><properties><property name="score" value="0.07186186976039788" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600526 [800] (title: Research on Eme) is a human-generated sample, misclassified as AI-generated with confidence 0.07186187&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600526 [800] (title: Research on Eme) is a human-generated sample, misclassified as AI-generated with confidence 0.07186187
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i202]" time="0.012"><properties><property name="score" value="0.017312556933485324" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600528 [646] (title: Mobile Data Min) is a human-generated sample, misclassified as AI-generated with confidence 0.01731256&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600528 [646] (title: Mobile Data Min) is a human-generated sample, misclassified as AI-generated with confidence 0.01731256
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i203]" time="0.017"><properties><property name="score" value="0.01491166850550174" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600529 [790] (title: Real-time Multi) is a human-generated sample, misclassified as AI-generated with confidence 0.01491167&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600529 [790] (title: Real-time Multi) is a human-generated sample, misclassified as AI-generated with confidence 0.01491167
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i204]" time="0.016"><properties><property name="score" value="0.1022686347556756" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600530 [1241] (title: Joint Temporal-) is a human-generated sample, misclassified as AI-generated with confidence 0.10226863&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600530 [1241] (title: Joint Temporal-) is a human-generated sample, misclassified as AI-generated with confidence 0.10226863
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i205]" time="0.021"><properties><property name="score" value="0.08012393138121521" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600533 [830] (title: Radar Target Di) is a human-generated sample, misclassified as AI-generated with confidence 0.08012393&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600533 [830] (title: Radar Target Di) is a human-generated sample, misclassified as AI-generated with confidence 0.08012393
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i206]" time="0.020"><properties><property name="score" value="0.019967578273982767" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600535 [722] (title: Text Recognitio) is a human-generated sample, misclassified as AI-generated with confidence 0.01996758&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600535 [722] (title: Text Recognitio) is a human-generated sample, misclassified as AI-generated with confidence 0.01996758
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i207]" time="0.014"><properties><property name="score" value="0.03793327875512009" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600536 [956] (title: Convolution Neu) is a human-generated sample, misclassified as AI-generated with confidence 0.03793328&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600536 [956] (title: Convolution Neu) is a human-generated sample, misclassified as AI-generated with confidence 0.03793328
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i208]" time="0.019"><properties><property name="score" value="0.0385880216581147" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i209]" time="0.017"><properties><property name="score" value="0.03674791853818249" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600542 [1409] (title: Optimal Smart I) is a human-generated sample, misclassified as AI-generated with confidence 0.03674792&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600542 [1409] (title: Optimal Smart I) is a human-generated sample, misclassified as AI-generated with confidence 0.03674792
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i210]" time="0.013"><properties><property name="score" value="0.014083902830758221" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i211]" time="0.012"><properties><property name="score" value="0.02285565104880871" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600545 [1001] (title: A Search Space ) is a human-generated sample, misclassified as AI-generated with confidence 0.02285565&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600545 [1001] (title: A Search Space ) is a human-generated sample, misclassified as AI-generated with confidence 0.02285565
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i212]" time="0.012"><properties><property name="score" value="0.026945633068448105" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600547 [757] (title: Visualization o) is a human-generated sample, misclassified as AI-generated with confidence 0.02694563&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600547 [757] (title: Visualization o) is a human-generated sample, misclassified as AI-generated with confidence 0.02694563
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i213]" time="0.012"><properties><property name="score" value="0.01741888227479449" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i214]" time="0.016"><properties><property name="score" value="0.012919175242143144" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i215]" time="0.018"><properties><property name="score" value="0.0014281839079977043" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600555 [1347] (title: SDP Based Unit ) is a human-generated sample, misclassified as AI-generated with confidence 0.00142818&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600555 [1347] (title: SDP Based Unit ) is a human-generated sample, misclassified as AI-generated with confidence 0.00142818
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i216]" time="0.015"><properties><property name="score" value="0.05888464982650721" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600567 [1003] (title: Probabilistic E) is a human-generated sample, misclassified as AI-generated with confidence 0.05888465&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600567 [1003] (title: Probabilistic E) is a human-generated sample, misclassified as AI-generated with confidence 0.05888465
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i217]" time="0.013"><properties><property name="score" value="0.06308418727344334" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600571 [1393] (title: Appropriate Tol) is a human-generated sample, misclassified as AI-generated with confidence 0.06308419&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600571 [1393] (title: Appropriate Tol) is a human-generated sample, misclassified as AI-generated with confidence 0.06308419
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i218]" time="0.012"><properties><property name="score" value="0.007376381205864702" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i219]" time="0.014"><properties><property name="score" value="0.006227162076438697" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600578 [764] (title: DC State Estima) is a human-generated sample, misclassified as AI-generated with confidence 0.00622716&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600578 [764] (title: DC State Estima) is a human-generated sample, misclassified as AI-generated with confidence 0.00622716
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i220]" time="0.014"><properties><property name="score" value="0.00043313472110617646" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600580 [626] (title: Bilevel Program) is a human-generated sample, misclassified as AI-generated with confidence 0.00043313&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600580 [626] (title: Bilevel Program) is a human-generated sample, misclassified as AI-generated with confidence 0.00043313
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i221]" time="0.013"><properties><property name="score" value="0.00472873900148052" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i222]" time="0.013"><properties><property name="score" value="0.012450520194218218" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600584 [1150] (title: Minimization of) is a human-generated sample, misclassified as AI-generated with confidence 0.01245052&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600584 [1150] (title: Minimization of) is a human-generated sample, misclassified as AI-generated with confidence 0.01245052
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i223]" time="0.016"><properties><property name="score" value="0.010240351442764029" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i224]" time="0.014"><properties><property name="score" value="0.062316920352306626" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600595 [1239] (title: An Artificial N) is a human-generated sample, misclassified as AI-generated with confidence 0.06231692&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600595 [1239] (title: An Artificial N) is a human-generated sample, misclassified as AI-generated with confidence 0.06231692
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i225]" time="0.016"><properties><property name="score" value="0.009592302511945983" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600596 [644] (title: ANDES: A Python) is a human-generated sample, misclassified as AI-generated with confidence 0.0095923&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600596 [644] (title: ANDES: A Python) is a human-generated sample, misclassified as AI-generated with confidence 0.0095923
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i226]" time="0.013"><properties><property name="score" value="0.014494468419645745" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i227]" time="0.019"><properties><property name="score" value="0.00955420518177208" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600612 [1034] (title: The Potential f) is a human-generated sample, misclassified as AI-generated with confidence 0.00955421&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600612 [1034] (title: The Potential f) is a human-generated sample, misclassified as AI-generated with confidence 0.00955421
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i228]" time="0.016"><properties><property name="score" value="0.043464514722757475" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600617 [1062] (title: Geometrically M) is a human-generated sample, misclassified as AI-generated with confidence 0.04346451&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600617 [1062] (title: Geometrically M) is a human-generated sample, misclassified as AI-generated with confidence 0.04346451
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i229]" time="0.013"><properties><property name="score" value="0.057072698927801424" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600623 [907] (title: A Probabilistic) is a human-generated sample, misclassified as AI-generated with confidence 0.0570727&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600623 [907] (title: A Probabilistic) is a human-generated sample, misclassified as AI-generated with confidence 0.0570727
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i230]" time="0.012"><properties><property name="score" value="0.03565451712920775" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600628 [949] (title: Optimal Operati) is a human-generated sample, misclassified as AI-generated with confidence 0.03565452&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600628 [949] (title: Optimal Operati) is a human-generated sample, misclassified as AI-generated with confidence 0.03565452
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i231]" time="0.012"><properties><property name="score" value="0.006671006916036948" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600635 [1401] (title: MILP Model for ) is a human-generated sample, misclassified as AI-generated with confidence 0.00667101&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600635 [1401] (title: MILP Model for ) is a human-generated sample, misclassified as AI-generated with confidence 0.00667101
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i232]" time="0.012"><properties><property name="score" value="0.012069408609305388" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i233]" time="0.013"><properties><property name="score" value="0.019119944588441573" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600639 [1054] (title: Cyber physical ) is a human-generated sample, misclassified as AI-generated with confidence 0.01911994&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600639 [1054] (title: Cyber physical ) is a human-generated sample, misclassified as AI-generated with confidence 0.01911994
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i234]" time="0.017"><properties><property name="score" value="0.016655670110932164" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i235]" time="0.015"><properties><property name="score" value="0.016423721660813184" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600647 [569] (title: Energy Manageme) is a human-generated sample, misclassified as AI-generated with confidence 0.01642372&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600647 [569] (title: Energy Manageme) is a human-generated sample, misclassified as AI-generated with confidence 0.01642372
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i236]" time="0.015"><properties><property name="score" value="0.04783417259044964" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i237]" time="0.014"><properties><property name="score" value="0.020175710655184753" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i238]" time="0.014"><properties><property name="score" value="0.0012242887328206553" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600654 [994] (title: Application of ) is a human-generated sample, misclassified as AI-generated with confidence 0.00122429&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600654 [994] (title: Application of ) is a human-generated sample, misclassified as AI-generated with confidence 0.00122429
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i239]" time="0.014"><properties><property name="score" value="0.005879370709621634" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i240]" time="0.016"><properties><property name="score" value="0.022799744529078403" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600661 [822] (title: Synchrophasor-B) is a human-generated sample, misclassified as AI-generated with confidence 0.02279974&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600661 [822] (title: Synchrophasor-B) is a human-generated sample, misclassified as AI-generated with confidence 0.02279974
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i241]" time="0.013"><properties><property name="score" value="0.042185398560828924" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600662 [1428] (title: Distributed Qua) is a human-generated sample, misclassified as AI-generated with confidence 0.0421854&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600662 [1428] (title: Distributed Qua) is a human-generated sample, misclassified as AI-generated with confidence 0.0421854
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i242]" time="0.015"><properties><property name="score" value="0.03303782670306399" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600668 [967] (title: Performance of ) is a human-generated sample, misclassified as AI-generated with confidence 0.03303783&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600668 [967] (title: Performance of ) is a human-generated sample, misclassified as AI-generated with confidence 0.03303783
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i243]" time="0.025"><properties><property name="score" value="0.03948640626656438" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600675 [1559] (title: A Dynamic State) is a human-generated sample, misclassified as AI-generated with confidence 0.03948641&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600675 [1559] (title: A Dynamic State) is a human-generated sample, misclassified as AI-generated with confidence 0.03948641
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i244]" time="0.013"><properties><property name="score" value="0.04392269177082575" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i245]" time="0.014"><properties><property name="score" value="0.022154567376331258" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600681 [1118] (title: Least-Cost Join) is a human-generated sample, misclassified as AI-generated with confidence 0.02215457&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600681 [1118] (title: Least-Cost Join) is a human-generated sample, misclassified as AI-generated with confidence 0.02215457
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i246]" time="0.013"><properties><property name="score" value="0.006513875883012998" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i247]" time="0.018"><properties><property name="score" value="0.007471766582684003" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600713 [1298] (title: Using PPG Signa) is a human-generated sample, misclassified as AI-generated with confidence 0.00747177&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600713 [1298] (title: Using PPG Signa) is a human-generated sample, misclassified as AI-generated with confidence 0.00747177
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i248]" time="0.013"><properties><property name="score" value="0.022004854670559792" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600717 [1009] (title: Deterministic P) is a human-generated sample, misclassified as AI-generated with confidence 0.02200485&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600717 [1009] (title: Deterministic P) is a human-generated sample, misclassified as AI-generated with confidence 0.02200485
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i249]" time="0.013"><properties><property name="score" value="0.013852827084587238" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600725 [1240] (title: Supporting the ) is a human-generated sample, misclassified as AI-generated with confidence 0.01385283&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600725 [1240] (title: Supporting the ) is a human-generated sample, misclassified as AI-generated with confidence 0.01385283
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i250]" time="0.015"><properties><property name="score" value="0.024473796231549994" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600727 [1458] (title: Three-Dimension) is a human-generated sample, misclassified as AI-generated with confidence 0.0244738&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600727 [1458] (title: Three-Dimension) is a human-generated sample, misclassified as AI-generated with confidence 0.0244738
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i251]" time="0.018"><properties><property name="score" value="0.05859872753415374" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600730 [1483] (title: High-Performanc) is a human-generated sample, misclassified as AI-generated with confidence 0.05859873&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600730 [1483] (title: High-Performanc) is a human-generated sample, misclassified as AI-generated with confidence 0.05859873
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i252]" time="0.020"><properties><property name="score" value="0.015241359748363603" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600732 [1170] (title: A Novel Generat) is a human-generated sample, misclassified as AI-generated with confidence 0.01524136&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600732 [1170] (title: A Novel Generat) is a human-generated sample, misclassified as AI-generated with confidence 0.01524136
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i253]" time="0.042"><properties><property name="score" value="0.041015251763834464" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i254]" time="0.024"><properties><property name="score" value="0.008312065461518747" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600735 [1051] (title: Weather-Classif) is a human-generated sample, misclassified as AI-generated with confidence 0.00831207&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600735 [1051] (title: Weather-Classif) is a human-generated sample, misclassified as AI-generated with confidence 0.00831207
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i255]" time="0.028"><properties><property name="score" value="0.022790975245926954" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600736 [1625] (title: Forecasting Tra) is a human-generated sample, misclassified as AI-generated with confidence 0.02279098&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600736 [1625] (title: Forecasting Tra) is a human-generated sample, misclassified as AI-generated with confidence 0.02279098
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i256]" time="0.014"><properties><property name="score" value="0.01450421164383986" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600737 [634] (title: Disturbance Obs) is a human-generated sample, misclassified as AI-generated with confidence 0.01450421&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600737 [634] (title: Disturbance Obs) is a human-generated sample, misclassified as AI-generated with confidence 0.01450421
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i257]" time="0.014"><properties><property name="score" value="0.12733124142977892" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600738 [1441] (title: Enhancing Cloud) is a human-generated sample, misclassified as AI-generated with confidence 0.12733124&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600738 [1441] (title: Enhancing Cloud) is a human-generated sample, misclassified as AI-generated with confidence 0.12733124
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i258]" time="0.013"><properties><property name="score" value="0.023585932180880675" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600748 [973] (title: Deploying the F) is a human-generated sample, misclassified as AI-generated with confidence 0.02358593&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600748 [973] (title: Deploying the F) is a human-generated sample, misclassified as AI-generated with confidence 0.02358593
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i259]" time="0.015"><properties><property name="score" value="0.010636532233249119" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i260]" time="0.014"><properties><property name="score" value="0.025529380367916854" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600751 [992] (title: Blockchain-Base) is a human-generated sample, misclassified as AI-generated with confidence 0.02552938&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600751 [992] (title: Blockchain-Base) is a human-generated sample, misclassified as AI-generated with confidence 0.02552938
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i261]" time="0.013"><properties><property name="score" value="0.04432185959129531" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600752 [1145] (title: Active Learning) is a human-generated sample, misclassified as AI-generated with confidence 0.04432186&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600752 [1145] (title: Active Learning) is a human-generated sample, misclassified as AI-generated with confidence 0.04432186
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i262]" time="0.015"><properties><property name="score" value="0.04420209599141911" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600753 [1036] (title: Hybrid-Augmente) is a human-generated sample, misclassified as AI-generated with confidence 0.0442021&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600753 [1036] (title: Hybrid-Augmente) is a human-generated sample, misclassified as AI-generated with confidence 0.0442021
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i263]" time="0.014"><properties><property name="score" value="0.048471046026232756" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i264]" time="0.013"><properties><property name="score" value="0.05524264734162454" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600755 [999] (title: Covert Timing C) is a human-generated sample, misclassified as AI-generated with confidence 0.05524265&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600755 [999] (title: Covert Timing C) is a human-generated sample, misclassified as AI-generated with confidence 0.05524265
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i265]" time="0.014"><properties><property name="score" value="0.021396769066345245" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600757 [1062] (title: Covert Wireless) is a human-generated sample, misclassified as AI-generated with confidence 0.02139677&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600757 [1062] (title: Covert Wireless) is a human-generated sample, misclassified as AI-generated with confidence 0.02139677
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i266]" time="0.017"><properties><property name="score" value="0.023673311912408845" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600762 [1070] (title: Latent Factor-B) is a human-generated sample, misclassified as AI-generated with confidence 0.02367331&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600762 [1070] (title: Latent Factor-B) is a human-generated sample, misclassified as AI-generated with confidence 0.02367331
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i267]" time="0.017"><properties><property name="score" value="0.020272612584015803" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600766 [777] (title: A Disturbance R) is a human-generated sample, misclassified as AI-generated with confidence 0.02027261&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600766 [777] (title: A Disturbance R) is a human-generated sample, misclassified as AI-generated with confidence 0.02027261
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i268]" time="0.020"><properties><property name="score" value="0.026069481299667574" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600778 [1093] (title: Securing Consum) is a human-generated sample, misclassified as AI-generated with confidence 0.02606948&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600778 [1093] (title: Securing Consum) is a human-generated sample, misclassified as AI-generated with confidence 0.02606948
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i269]" time="0.022"><properties><property name="score" value="0.06079000445390892" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600779 [1202] (title: Privacy-Preserv) is a human-generated sample, misclassified as AI-generated with confidence 0.06079&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600779 [1202] (title: Privacy-Preserv) is a human-generated sample, misclassified as AI-generated with confidence 0.06079
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i270]" time="0.019"><properties><property name="score" value="0.08465848275004406" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600780 [1410] (title: Distributed Dat) is a human-generated sample, misclassified as AI-generated with confidence 0.08465848&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600780 [1410] (title: Distributed Dat) is a human-generated sample, misclassified as AI-generated with confidence 0.08465848
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i271]" time="0.014"><properties><property name="score" value="0.04057773247310159" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600781 [1205] (title: Proactive Cache) is a human-generated sample, misclassified as AI-generated with confidence 0.04057773&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600781 [1205] (title: Proactive Cache) is a human-generated sample, misclassified as AI-generated with confidence 0.04057773
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i272]" time="0.013"><properties><property name="score" value="0.004890389211248403" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i273]" time="0.013"><properties><property name="score" value="0.019364341854055667" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i274]" time="0.015"><properties><property name="score" value="0.10303606454281633" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600784 [1119] (title: Privacy-Preserv) is a human-generated sample, misclassified as AI-generated with confidence 0.10303606&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600784 [1119] (title: Privacy-Preserv) is a human-generated sample, misclassified as AI-generated with confidence 0.10303606
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i275]" time="0.026"><properties><property name="score" value="0.07801010536822947" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600788 [1815] (title: Communication s) is a human-generated sample, misclassified as AI-generated with confidence 0.07801011&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600788 [1815] (title: Communication s) is a human-generated sample, misclassified as AI-generated with confidence 0.07801011
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i276]" time="0.013"><properties><property name="score" value="0.016849462667239967" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i277]" time="0.029"><properties><property name="score" value="0.01846781795028019" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600790 [1539] (title: Wireless acoust) is a human-generated sample, misclassified as AI-generated with confidence 0.01846782&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600790 [1539] (title: Wireless acoust) is a human-generated sample, misclassified as AI-generated with confidence 0.01846782
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i278]" time="0.016"><properties><property name="score" value="0.010092518397664119" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600791 [972] (title: Event-triggered) is a human-generated sample, misclassified as AI-generated with confidence 0.01009252&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600791 [972] (title: Event-triggered) is a human-generated sample, misclassified as AI-generated with confidence 0.01009252
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i279]" time="0.014"><properties><property name="score" value="0.03158903460293572" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600797 [1115] (title: Template-based ) is a human-generated sample, misclassified as AI-generated with confidence 0.03158903&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600797 [1115] (title: Template-based ) is a human-generated sample, misclassified as AI-generated with confidence 0.03158903
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i280]" time="0.016"><properties><property name="score" value="0.0239144601179353" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i281]" time="0.029"><properties><property name="score" value="0.01745292380793325" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600824 [1634] (title: IoT Device Fing) is a human-generated sample, misclassified as AI-generated with confidence 0.01745292&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600824 [1634] (title: IoT Device Fing) is a human-generated sample, misclassified as AI-generated with confidence 0.01745292
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i282]" time="0.014"><properties><property name="score" value="0.05818658994978243" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600826 [1136] (title: Personal Authen) is a human-generated sample, misclassified as AI-generated with confidence 0.05818659&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600826 [1136] (title: Personal Authen) is a human-generated sample, misclassified as AI-generated with confidence 0.05818659
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i283]" time="0.013"><properties><property name="score" value="0.011155257049660916" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i284]" time="0.014"><properties><property name="score" value="0.012652903139903637" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i285]" time="0.013"><properties><property name="score" value="0.02875685020536367" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600836 [925] (title: A Soft Ring-sha) is a human-generated sample, misclassified as AI-generated with confidence 0.02875685&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600836 [925] (title: A Soft Ring-sha) is a human-generated sample, misclassified as AI-generated with confidence 0.02875685
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i286]" time="0.016"><properties><property name="score" value="0.02361301990676523" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600837 [1051] (title: IoT Based Poult) is a human-generated sample, misclassified as AI-generated with confidence 0.02361302&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600837 [1051] (title: IoT Based Poult) is a human-generated sample, misclassified as AI-generated with confidence 0.02361302
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i287]" time="0.013"><properties><property name="score" value="0.03441756522640782" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i288]" time="0.016"><properties><property name="score" value="0.04259981592122375" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i289]" time="0.019"><properties><property name="score" value="0.0025946225820305813" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i290]" time="0.023"><properties><property name="score" value="0.0011592018049011443" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i291]" time="0.022"><properties><property name="score" value="0.03377230805370335" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600844 [940] (title: Digital Twins o) is a human-generated sample, misclassified as AI-generated with confidence 0.03377231&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600844 [940] (title: Digital Twins o) is a human-generated sample, misclassified as AI-generated with confidence 0.03377231
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i292]" time="0.025"><properties><property name="score" value="0.02990700526437351" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600847 [969] (title: Authentication ) is a human-generated sample, misclassified as AI-generated with confidence 0.02990701&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600847 [969] (title: Authentication ) is a human-generated sample, misclassified as AI-generated with confidence 0.02990701
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i293]" time="0.016"><properties><property name="score" value="0.0036931019029295964" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600849 [1034] (title: Analog Ultra Lo) is a human-generated sample, misclassified as AI-generated with confidence 0.0036931&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600849 [1034] (title: Analog Ultra Lo) is a human-generated sample, misclassified as AI-generated with confidence 0.0036931
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i294]" time="0.019"><properties><property name="score" value="0.01506938478546549" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i295]" time="0.020"><properties><property name="score" value="0.018757791037959937" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600852 [1156] (title: Speaker Recogni) is a human-generated sample, misclassified as AI-generated with confidence 0.01875779&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600852 [1156] (title: Speaker Recogni) is a human-generated sample, misclassified as AI-generated with confidence 0.01875779
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i296]" time="0.022"><properties><property name="score" value="0.038191071464493964" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600854 [753] (title: Smart Contracts) is a human-generated sample, misclassified as AI-generated with confidence 0.03819107&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600854 [753] (title: Smart Contracts) is a human-generated sample, misclassified as AI-generated with confidence 0.03819107
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i297]" time="0.015"><properties><property name="score" value="0.018967577272488434" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i298]" time="0.030"><properties><property name="score" value="0.06445456509339026" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600856 [1501] (title: Generative mode) is a human-generated sample, misclassified as AI-generated with confidence 0.06445457&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600856 [1501] (title: Generative mode) is a human-generated sample, misclassified as AI-generated with confidence 0.06445457
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i299]" time="0.038"><properties><property name="score" value="0.05515036324935188" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i300]" time="0.033"><properties><property name="score" value="0.028986819212243012" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600858 [1733] (title: NPK Soil Nutrie) is a human-generated sample, misclassified as AI-generated with confidence 0.02898682&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600858 [1733] (title: NPK Soil Nutrie) is a human-generated sample, misclassified as AI-generated with confidence 0.02898682
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i301]" time="0.031"><properties><property name="score" value="0.0010754403124187248" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600859 [1755] (title: Personal Authen) is a human-generated sample, misclassified as AI-generated with confidence 0.00107544&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600859 [1755] (title: Personal Authen) is a human-generated sample, misclassified as AI-generated with confidence 0.00107544
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i302]" time="0.030"><properties><property name="score" value="0.01991834439404782" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i303]" time="0.015"><properties><property name="score" value="0.031212215564702106" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600862 [1211] (title: Large Data Down) is a human-generated sample, misclassified as AI-generated with confidence 0.03121222&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600862 [1211] (title: Large Data Down) is a human-generated sample, misclassified as AI-generated with confidence 0.03121222
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i304]" time="0.027"><properties><property name="score" value="0.06126749566368646" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600863 [1578] (title: A new method fo) is a human-generated sample, misclassified as AI-generated with confidence 0.0612675&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600863 [1578] (title: A new method fo) is a human-generated sample, misclassified as AI-generated with confidence 0.0612675
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i305]" time="0.016"><properties><property name="score" value="0.0037618979939813" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i306]" time="0.017"><properties><property name="score" value="0.008062396827412943" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600867 [1018] (title: Applying Intell) is a human-generated sample, misclassified as AI-generated with confidence 0.0080624&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600867 [1018] (title: Applying Intell) is a human-generated sample, misclassified as AI-generated with confidence 0.0080624
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i307]" time="0.016"><properties><property name="score" value="0.05709257756999486" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600869 [1273] (title: Flood Modelling) is a human-generated sample, misclassified as AI-generated with confidence 0.05709258&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600869 [1273] (title: Flood Modelling) is a human-generated sample, misclassified as AI-generated with confidence 0.05709258
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i308]" time="0.016"><properties><property name="score" value="0.0019591979371791357" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600870 [888] (title: Sentiment Analy) is a human-generated sample, misclassified as AI-generated with confidence 0.0019592&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600870 [888] (title: Sentiment Analy) is a human-generated sample, misclassified as AI-generated with confidence 0.0019592
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i309]" time="0.014"><properties><property name="score" value="0.000920674001708699" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i310]" time="0.014"><properties><property name="score" value="0.007580440387294773" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600875 [845] (title: An OEE Improvem) is a human-generated sample, misclassified as AI-generated with confidence 0.00758044&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600875 [845] (title: An OEE Improvem) is a human-generated sample, misclassified as AI-generated with confidence 0.00758044
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i311]" time="0.015"><properties><property name="score" value="0.008362872799733134" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i312]" time="0.015"><properties><property name="score" value="0.004604506494676475" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600879 [1032] (title: A Configurable ) is a human-generated sample, misclassified as AI-generated with confidence 0.00460451&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600879 [1032] (title: A Configurable ) is a human-generated sample, misclassified as AI-generated with confidence 0.00460451
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i313]" time="0.016"><properties><property name="score" value="0.023647619620831284" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i314]" time="0.017"><properties><property name="score" value="0.011361308380852808" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600885 [633] (title: Recognition of ) is a human-generated sample, misclassified as AI-generated with confidence 0.01136131&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600885 [633] (title: Recognition of ) is a human-generated sample, misclassified as AI-generated with confidence 0.01136131
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i315]" time="0.015"><properties><property name="score" value="0.027009353125667657" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i316]" time="0.016"><properties><property name="score" value="0.0031183701898640503" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600890 [1328] (title: Opportunistic W) is a human-generated sample, misclassified as AI-generated with confidence 0.00311837&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600890 [1328] (title: Opportunistic W) is a human-generated sample, misclassified as AI-generated with confidence 0.00311837
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i317]" time="0.015"><properties><property name="score" value="0.0015668042350647227" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600891 [1292] (title: Smart ECG Holte) is a human-generated sample, misclassified as AI-generated with confidence 0.0015668&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600891 [1292] (title: Smart ECG Holte) is a human-generated sample, misclassified as AI-generated with confidence 0.0015668
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i318]" time="0.024"><properties><property name="score" value="0.007011508868720595" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i319]" time="0.015"><properties><property name="score" value="0.023157655157479695" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i320]" time="0.018"><properties><property name="score" value="0.051555397697716776" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i321]" time="0.027"><properties><property name="score" value="0.03201816534554314" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600896 [439] (title: Application of ) is a human-generated sample, misclassified as AI-generated with confidence 0.03201817&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600896 [439] (title: Application of ) is a human-generated sample, misclassified as AI-generated with confidence 0.03201817
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i322]" time="0.018"><properties><property name="score" value="0.005424362344985667" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600898 [547] (title: KrishiMitr (Far) is a human-generated sample, misclassified as AI-generated with confidence 0.00542436&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600898 [547] (title: KrishiMitr (Far) is a human-generated sample, misclassified as AI-generated with confidence 0.00542436
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i323]" time="0.023"><properties><property name="score" value="0.030323024403533516" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600900 [795] (title: Range-Based Col) is a human-generated sample, misclassified as AI-generated with confidence 0.03032302&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600900 [795] (title: Range-Based Col) is a human-generated sample, misclassified as AI-generated with confidence 0.03032302
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i324]" time="0.021"><properties><property name="score" value="0.006538433238884034" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i325]" time="0.016"><properties><property name="score" value="0.014552400035139357" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i326]" time="0.017"><properties><property name="score" value="0.005044326640923913" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i327]" time="0.016"><properties><property name="score" value="0.004491106259346678" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i328]" time="0.014"><properties><property name="score" value="0.011500555605026275" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600912 [1064] (title: Automatic mass ) is a human-generated sample, misclassified as AI-generated with confidence 0.01150056&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600912 [1064] (title: Automatic mass ) is a human-generated sample, misclassified as AI-generated with confidence 0.01150056
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i329]" time="0.015"><properties><property name="score" value="0.00965707332570842" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i330]" time="0.015"><properties><property name="score" value="0.05314921457946764" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600916 [757] (title: Modelling of a ) is a human-generated sample, misclassified as AI-generated with confidence 0.05314921&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600916 [757] (title: Modelling of a ) is a human-generated sample, misclassified as AI-generated with confidence 0.05314921
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i331]" time="0.019"><properties><property name="score" value="0.03092398522679396" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i332]" time="0.018"><properties><property name="score" value="0.004717348545801858" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i333]" time="0.015"><properties><property name="score" value="0.01939271296789169" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600922 [1022] (title: Anonymity and C) is a human-generated sample, misclassified as AI-generated with confidence 0.01939271&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600922 [1022] (title: Anonymity and C) is a human-generated sample, misclassified as AI-generated with confidence 0.01939271
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i334]" time="0.014"><properties><property name="score" value="0.020425342079954945" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600924 [1049] (title: SEECSSim - A Pa) is a human-generated sample, misclassified as AI-generated with confidence 0.02042534&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600924 [1049] (title: SEECSSim - A Pa) is a human-generated sample, misclassified as AI-generated with confidence 0.02042534
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i335]" time="0.016"><properties><property name="score" value="0.027694146692125443" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600925 [562] (title: Data Assimilati) is a human-generated sample, misclassified as AI-generated with confidence 0.02769415&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600925 [562] (title: Data Assimilati) is a human-generated sample, misclassified as AI-generated with confidence 0.02769415
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i336]" time="0.014"><properties><property name="score" value="0.03994949490186417" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600926 [1171] (title: Towards a Reali) is a human-generated sample, misclassified as AI-generated with confidence 0.03994949&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600926 [1171] (title: Towards a Reali) is a human-generated sample, misclassified as AI-generated with confidence 0.03994949
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i337]" time="0.016"><properties><property name="score" value="0.08509833435812464" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600928 [1310] (title: Keynote Speaker) is a human-generated sample, misclassified as AI-generated with confidence 0.08509833&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600928 [1310] (title: Keynote Speaker) is a human-generated sample, misclassified as AI-generated with confidence 0.08509833
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i338]" time="0.013"><properties><property name="score" value="0.013092037151318792" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600930 [546] (title: Outage Probabil) is a human-generated sample, misclassified as AI-generated with confidence 0.01309204&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600930 [546] (title: Outage Probabil) is a human-generated sample, misclassified as AI-generated with confidence 0.01309204
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i339]" time="0.014"><properties><property name="score" value="0.02618778816409484" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600932 [702] (title: An Experimentat) is a human-generated sample, misclassified as AI-generated with confidence 0.02618779&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600932 [702] (title: An Experimentat) is a human-generated sample, misclassified as AI-generated with confidence 0.02618779
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i340]" time="0.014"><properties><property name="score" value="0.04252351895972506" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600933 [1400] (title: A Cloud Based S) is a human-generated sample, misclassified as AI-generated with confidence 0.04252352&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600933 [1400] (title: A Cloud Based S) is a human-generated sample, misclassified as AI-generated with confidence 0.04252352
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i341]" time="0.017"><properties><property name="score" value="0.008782470696094835" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600934 [736] (title: Time Synchroniz) is a human-generated sample, misclassified as AI-generated with confidence 0.00878247&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600934 [736] (title: Time Synchroniz) is a human-generated sample, misclassified as AI-generated with confidence 0.00878247
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i342]" time="0.013"><properties><property name="score" value="0.005778753310281548" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i343]" time="0.022"><properties><property name="score" value="0.020912940513906353" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600937 [645] (title: Delay-based Dis) is a human-generated sample, misclassified as AI-generated with confidence 0.02091294&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600937 [645] (title: Delay-based Dis) is a human-generated sample, misclassified as AI-generated with confidence 0.02091294
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i344]" time="0.018"><properties><property name="score" value="0.004320012114394922" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600958 [826] (title: Harvesting Elec) is a human-generated sample, misclassified as AI-generated with confidence 0.00432001&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600958 [826] (title: Harvesting Elec) is a human-generated sample, misclassified as AI-generated with confidence 0.00432001
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i345]" time="0.015"><properties><property name="score" value="0.0197530020883796" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600959 [598] (title: Fall detection ) is a human-generated sample, misclassified as AI-generated with confidence 0.019753&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600959 [598] (title: Fall detection ) is a human-generated sample, misclassified as AI-generated with confidence 0.019753
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i346]" time="0.014"><properties><property name="score" value="0.02022622619980985" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i347]" time="0.013"><properties><property name="score" value="0.029129077438938555" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i348]" time="0.014"><properties><property name="score" value="0.011398395687262086" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600984 [963] (title: Photoplethysmog) is a human-generated sample, misclassified as AI-generated with confidence 0.0113984&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600984 [963] (title: Photoplethysmog) is a human-generated sample, misclassified as AI-generated with confidence 0.0113984
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i349]" time="0.018"><properties><property name="score" value="0.0004677600403735216" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8600993 [393] (title: Performance of ) is a human-generated sample, misclassified as AI-generated with confidence 0.00046776&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8600993 [393] (title: Performance of ) is a human-generated sample, misclassified as AI-generated with confidence 0.00046776
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i350]" time="0.026"><properties><property name="score" value="0.009700581847430759" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i351]" time="0.013"><properties><property name="score" value="0.041015412768669046" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601004 [1027] (title: Towards a Hybri) is a human-generated sample, misclassified as AI-generated with confidence 0.04101541&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601004 [1027] (title: Towards a Hybri) is a human-generated sample, misclassified as AI-generated with confidence 0.04101541
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i352]" time="0.014"><properties><property name="score" value="0.01972491564358192" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601010 [840] (title: Emerging Simula) is a human-generated sample, misclassified as AI-generated with confidence 0.01972492&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601010 [840] (title: Emerging Simula) is a human-generated sample, misclassified as AI-generated with confidence 0.01972492
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i353]" time="0.014"><properties><property name="score" value="0.016766356213138156" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601011 [711] (title: Machine Learnin) is a human-generated sample, misclassified as AI-generated with confidence 0.01676636&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601011 [711] (title: Machine Learnin) is a human-generated sample, misclassified as AI-generated with confidence 0.01676636
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i354]" time="0.016"><properties><property name="score" value="0.037184920469796046" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601013 [899] (title: Simulation of V) is a human-generated sample, misclassified as AI-generated with confidence 0.03718492&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601013 [899] (title: Simulation of V) is a human-generated sample, misclassified as AI-generated with confidence 0.03718492
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i355]" time="0.014"><properties><property name="score" value="0.030402281737479386" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601014 [1153] (title: A Comprehensive) is a human-generated sample, misclassified as AI-generated with confidence 0.03040228&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601014 [1153] (title: A Comprehensive) is a human-generated sample, misclassified as AI-generated with confidence 0.03040228
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i356]" time="0.015"><properties><property name="score" value="0.03369266469656074" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601016 [1428] (title: Exploring Execu) is a human-generated sample, misclassified as AI-generated with confidence 0.03369266&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601016 [1428] (title: Exploring Execu) is a human-generated sample, misclassified as AI-generated with confidence 0.03369266
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i357]" time="0.014"><properties><property name="score" value="0.004259660968797263" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601020 [1131] (title: A Platform for ) is a human-generated sample, misclassified as AI-generated with confidence 0.00425966&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601020 [1131] (title: A Platform for ) is a human-generated sample, misclassified as AI-generated with confidence 0.00425966
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i358]" time="0.014"><properties><property name="score" value="0.02010919397683608" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i359]" time="0.014"><properties><property name="score" value="0.005625057249153409" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601027 [1131] (title: A Comprehensive) is a human-generated sample, misclassified as AI-generated with confidence 0.00562506&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601027 [1131] (title: A Comprehensive) is a human-generated sample, misclassified as AI-generated with confidence 0.00562506
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i360]" time="0.014"><properties><property name="score" value="0.025875718062245978" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601032 [808] (title: Iris Recognitio) is a human-generated sample, misclassified as AI-generated with confidence 0.02587572&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601032 [808] (title: Iris Recognitio) is a human-generated sample, misclassified as AI-generated with confidence 0.02587572
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i361]" time="0.013"><properties><property name="score" value="0.019792067998530838" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i362]" time="0.014"><properties><property name="score" value="0.020121438986009377" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601087 [991] (title: The Design of B) is a human-generated sample, misclassified as AI-generated with confidence 0.02012144&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601087 [991] (title: The Design of B) is a human-generated sample, misclassified as AI-generated with confidence 0.02012144
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i363]" time="0.015"><properties><property name="score" value="0.039088176527279916" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601090 [1096] (title: Artificial Neur) is a human-generated sample, misclassified as AI-generated with confidence 0.03908818&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601090 [1096] (title: Artificial Neur) is a human-generated sample, misclassified as AI-generated with confidence 0.03908818
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i364]" time="0.013"><properties><property name="score" value="0.008442227444094286" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601092 [1315] (title: An Ant Colony H) is a human-generated sample, misclassified as AI-generated with confidence 0.00844223&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601092 [1315] (title: An Ant Colony H) is a human-generated sample, misclassified as AI-generated with confidence 0.00844223
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i365]" time="0.014"><properties><property name="score" value="0.02721954391234349" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i366]" time="0.013"><properties><property name="score" value="0.056728740196915206" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601097 [930] (title: Call Data Recor) is a human-generated sample, misclassified as AI-generated with confidence 0.05672874&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601097 [930] (title: Call Data Recor) is a human-generated sample, misclassified as AI-generated with confidence 0.05672874
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i367]" time="0.013"><properties><property name="score" value="0.006038414137982118" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601139 [732] (title: IoT Sensor Data) is a human-generated sample, misclassified as AI-generated with confidence 0.00603841&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601139 [732] (title: IoT Sensor Data) is a human-generated sample, misclassified as AI-generated with confidence 0.00603841
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i368]" time="0.012"><properties><property name="score" value="0.0002708376014959324" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i369]" time="0.012"><properties><property name="score" value="0.02147302891202796" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i370]" time="0.023"><properties><property name="score" value="0.012377132165009035" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i371]" time="0.024"><properties><property name="score" value="0.02380698976600648" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601157 [759] (title: Stream machine ) is a human-generated sample, misclassified as AI-generated with confidence 0.02380699&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601157 [759] (title: Stream machine ) is a human-generated sample, misclassified as AI-generated with confidence 0.02380699
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i372]" time="0.012"><properties><property name="score" value="0.01246498974634136" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i373]" time="0.014"><properties><property name="score" value="0.00772085689828006" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601183 [1347] (title: Detecting Seizu) is a human-generated sample, misclassified as AI-generated with confidence 0.00772086&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601183 [1347] (title: Detecting Seizu) is a human-generated sample, misclassified as AI-generated with confidence 0.00772086
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i374]" time="0.023"><properties><property name="score" value="0.04331482489474434" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601184 [1673] (title: A Cooperation A) is a human-generated sample, misclassified as AI-generated with confidence 0.04331482&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601184 [1673] (title: A Cooperation A) is a human-generated sample, misclassified as AI-generated with confidence 0.04331482
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i375]" time="0.023"><properties><property name="score" value="0.09070921123517715" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601185 [1719] (title: Hierarchical Fa) is a human-generated sample, misclassified as AI-generated with confidence 0.09070921&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601185 [1719] (title: Hierarchical Fa) is a human-generated sample, misclassified as AI-generated with confidence 0.09070921
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i376]" time="0.014"><properties><property name="score" value="0.022364219669490648" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601186 [1242] (title: A Survey of Ste) is a human-generated sample, misclassified as AI-generated with confidence 0.02236422&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601186 [1242] (title: A Survey of Ste) is a human-generated sample, misclassified as AI-generated with confidence 0.02236422
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i377]" time="0.023"><properties><property name="score" value="0.09539930473609634" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601191 [1796] (title: An Adaptive Col) is a human-generated sample, misclassified as AI-generated with confidence 0.0953993&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601191 [1796] (title: An Adaptive Col) is a human-generated sample, misclassified as AI-generated with confidence 0.0953993
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i378]" time="0.013"><properties><property name="score" value="0.05567557011321722" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601192 [1252] (title: Study on Planet) is a human-generated sample, misclassified as AI-generated with confidence 0.05567557&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601192 [1252] (title: Study on Planet) is a human-generated sample, misclassified as AI-generated with confidence 0.05567557
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i379]" time="0.013"><properties><property name="score" value="0.04043083604010067" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601196 [1429] (title: An Active Learn) is a human-generated sample, misclassified as AI-generated with confidence 0.04043084&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601196 [1429] (title: An Active Learn) is a human-generated sample, misclassified as AI-generated with confidence 0.04043084
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i380]" time="0.014"><properties><property name="score" value="0.030580811244096395" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601199 [925] (title: Augmented-Reali) is a human-generated sample, misclassified as AI-generated with confidence 0.03058081&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601199 [925] (title: Augmented-Reali) is a human-generated sample, misclassified as AI-generated with confidence 0.03058081
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i381]" time="0.025"><properties><property name="score" value="0.04308716096532638" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i382]" time="0.024"><properties><property name="score" value="0.016565531794321638" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i383]" time="0.016"><properties><property name="score" value="0.008600497197608359" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i384]" time="0.012"><properties><property name="score" value="0.0063353268683674635" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i385]" time="0.013"><properties><property name="score" value="0.022036632651378785" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601220 [934] (title: Cloud Robotics ) is a human-generated sample, misclassified as AI-generated with confidence 0.02203663&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601220 [934] (title: Cloud Robotics ) is a human-generated sample, misclassified as AI-generated with confidence 0.02203663
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i386]" time="0.012"><properties><property name="score" value="0.00015454130415926137" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i387]" time="0.013"><properties><property name="score" value="0.018274744980345536" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601223 [1225] (title: Determining Key) is a human-generated sample, misclassified as AI-generated with confidence 0.01827474&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601223 [1225] (title: Determining Key) is a human-generated sample, misclassified as AI-generated with confidence 0.01827474
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i388]" time="0.014"><properties><property name="score" value="0.04276583923750188" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601226 [846] (title: Wireless Commun) is a human-generated sample, misclassified as AI-generated with confidence 0.04276584&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601226 [846] (title: Wireless Commun) is a human-generated sample, misclassified as AI-generated with confidence 0.04276584
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i389]" time="0.023"><properties><property name="score" value="0.03282373205349115" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i390]" time="0.012"><properties><property name="score" value="0.019027027007689767" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601234 [955] (title: Composing Scala) is a human-generated sample, misclassified as AI-generated with confidence 0.01902703&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601234 [955] (title: Composing Scala) is a human-generated sample, misclassified as AI-generated with confidence 0.01902703
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i391]" time="0.013"><properties><property name="score" value="0.04567895745650197" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601235 [971] (title: Artificial Inte) is a human-generated sample, misclassified as AI-generated with confidence 0.04567896&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601235 [971] (title: Artificial Inte) is a human-generated sample, misclassified as AI-generated with confidence 0.04567896
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i392]" time="0.023"><properties><property name="score" value="0.023618165291005255" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601236 [1537] (title: Business Analyt) is a human-generated sample, misclassified as AI-generated with confidence 0.02361817&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601236 [1537] (title: Business Analyt) is a human-generated sample, misclassified as AI-generated with confidence 0.02361817
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i393]" time="0.013"><properties><property name="score" value="0.005888626437083877" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601238 [1426] (title: Discrimination ) is a human-generated sample, misclassified as AI-generated with confidence 0.00588863&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601238 [1426] (title: Discrimination ) is a human-generated sample, misclassified as AI-generated with confidence 0.00588863
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i394]" time="0.011"><properties><property name="score" value="0.006136902386771892" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i395]" time="0.011"><properties><property name="score" value="0.014170178287253865" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601243 [652] (title: Preprocessed Fa) is a human-generated sample, misclassified as AI-generated with confidence 0.01417018&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601243 [652] (title: Preprocessed Fa) is a human-generated sample, misclassified as AI-generated with confidence 0.01417018
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i396]" time="0.012"><properties><property name="score" value="0.02608639686300096" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i397]" time="0.022"><properties><property name="score" value="0.04798540378907473" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601249 [1545] (title: The Value Persp) is a human-generated sample, misclassified as AI-generated with confidence 0.0479854&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601249 [1545] (title: The Value Persp) is a human-generated sample, misclassified as AI-generated with confidence 0.0479854
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i398]" time="0.012"><properties><property name="score" value="0.10753290301053764" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601250 [1410] (title: A Framework to ) is a human-generated sample, misclassified as AI-generated with confidence 0.1075329&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601250 [1410] (title: A Framework to ) is a human-generated sample, misclassified as AI-generated with confidence 0.1075329
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i399]" time="0.011"><properties><property name="score" value="0.007087037585978839" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601253 [439] (title: Parallel Image ) is a human-generated sample, misclassified as AI-generated with confidence 0.00708704&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601253 [439] (title: Parallel Image ) is a human-generated sample, misclassified as AI-generated with confidence 0.00708704
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i400]" time="0.013"><properties><property name="score" value="0.005123637729481656" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601254 [731] (title: Emulating Softw) is a human-generated sample, misclassified as AI-generated with confidence 0.00512364&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601254 [731] (title: Emulating Softw) is a human-generated sample, misclassified as AI-generated with confidence 0.00512364
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i401]" time="0.011"><properties><property name="score" value="0.0180489350859403" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601258 [779] (title: Challenges and ) is a human-generated sample, misclassified as AI-generated with confidence 0.01804894&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601258 [779] (title: Challenges and ) is a human-generated sample, misclassified as AI-generated with confidence 0.01804894
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i402]" time="0.012"><properties><property name="score" value="0.016670174340605026" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601259 [612] (title: Proficient Algo) is a human-generated sample, misclassified as AI-generated with confidence 0.01667017&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601259 [612] (title: Proficient Algo) is a human-generated sample, misclassified as AI-generated with confidence 0.01667017
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i403]" time="0.012"><properties><property name="score" value="0.1182704895876685" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601260 [1306] (title: Improving Class) is a human-generated sample, misclassified as AI-generated with confidence 0.11827049&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601260 [1306] (title: Improving Class) is a human-generated sample, misclassified as AI-generated with confidence 0.11827049
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i404]" time="0.011"><properties><property name="score" value="0.01018260830710771" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i405]" time="0.022"><properties><property name="score" value="0.07722990734346472" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601262 [1620] (title: The Effects of ) is a human-generated sample, misclassified as AI-generated with confidence 0.07722991&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601262 [1620] (title: The Effects of ) is a human-generated sample, misclassified as AI-generated with confidence 0.07722991
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i406]" time="0.011"><properties><property name="score" value="0.05569793710542004" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601264 [991] (title: Cloud Robotic A) is a human-generated sample, misclassified as AI-generated with confidence 0.05569794&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601264 [991] (title: Cloud Robotic A) is a human-generated sample, misclassified as AI-generated with confidence 0.05569794
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i407]" time="0.012"><properties><property name="score" value="0.011650449832206755" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601265 [765] (title: Fuzzy Logic Non) is a human-generated sample, misclassified as AI-generated with confidence 0.01165045&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601265 [765] (title: Fuzzy Logic Non) is a human-generated sample, misclassified as AI-generated with confidence 0.01165045
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i408]" time="0.011"><properties><property name="score" value="0.0392453904558554" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601271 [1462] (title: Employment Law ) is a human-generated sample, misclassified as AI-generated with confidence 0.03924539&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601271 [1462] (title: Employment Law ) is a human-generated sample, misclassified as AI-generated with confidence 0.03924539
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i409]" time="0.013"><properties><property name="score" value="0.04551380408066574" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601272 [1113] (title: Design of an Au) is a human-generated sample, misclassified as AI-generated with confidence 0.0455138&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601272 [1113] (title: Design of an Au) is a human-generated sample, misclassified as AI-generated with confidence 0.0455138
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i410]" time="0.012"><properties><property name="score" value="0.023291262653418743" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i411]" time="0.013"><properties><property name="score" value="0.0017435641106344768" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601277 [1005] (title: Asymptotic Anal) is a human-generated sample, misclassified as AI-generated with confidence 0.00174356&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601277 [1005] (title: Asymptotic Anal) is a human-generated sample, misclassified as AI-generated with confidence 0.00174356
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i412]" time="0.015"><properties><property name="score" value="0.05211539371891005" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601278 [908] (title: A Review of Aut) is a human-generated sample, misclassified as AI-generated with confidence 0.05211539&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601278 [908] (title: A Review of Aut) is a human-generated sample, misclassified as AI-generated with confidence 0.05211539
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i413]" time="0.012"><properties><property name="score" value="0.023119704578428157" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601280 [1292] (title: The Experimenta) is a human-generated sample, misclassified as AI-generated with confidence 0.0231197&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601280 [1292] (title: The Experimenta) is a human-generated sample, misclassified as AI-generated with confidence 0.0231197
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i414]" time="0.015"><properties><property name="score" value="0.03616895835059086" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601281 [968] (title: Virtual Private) is a human-generated sample, misclassified as AI-generated with confidence 0.03616896&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601281 [968] (title: Virtual Private) is a human-generated sample, misclassified as AI-generated with confidence 0.03616896
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i415]" time="0.011"><properties><property name="score" value="0.03713887339060751" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601282 [977] (title: A Hybrid Recomm) is a human-generated sample, misclassified as AI-generated with confidence 0.03713887&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601282 [977] (title: A Hybrid Recomm) is a human-generated sample, misclassified as AI-generated with confidence 0.03713887
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i416]" time="0.013"><properties><property name="score" value="0.06396244111980809" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601284 [934] (title: Switch: A Batte) is a human-generated sample, misclassified as AI-generated with confidence 0.06396244&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601284 [934] (title: Switch: A Batte) is a human-generated sample, misclassified as AI-generated with confidence 0.06396244
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i417]" time="0.012"><properties><property name="score" value="0.045799999880274944" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601291 [1168] (title: Improving Assoc) is a human-generated sample, misclassified as AI-generated with confidence 0.0458&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601291 [1168] (title: Improving Assoc) is a human-generated sample, misclassified as AI-generated with confidence 0.0458
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i418]" time="0.011"><properties><property name="score" value="0.0223976167510076" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i419]" time="0.012"><properties><property name="score" value="0.0008873443618162558" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601298 [1489] (title: Services and Ap) is a human-generated sample, misclassified as AI-generated with confidence 0.00088734&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601298 [1489] (title: Services and Ap) is a human-generated sample, misclassified as AI-generated with confidence 0.00088734
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i420]" time="0.011"><properties><property name="score" value="0.022300906682204413" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i421]" time="0.022"><properties><property name="score" value="0.03263057044149539" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i422]" time="0.011"><properties><property name="score" value="0.04274043590459797" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601304 [1280] (title: Implementation ) is a human-generated sample, misclassified as AI-generated with confidence 0.04274044&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601304 [1280] (title: Implementation ) is a human-generated sample, misclassified as AI-generated with confidence 0.04274044
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i423]" time="0.012"><properties><property name="score" value="0.015766444954468106" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i424]" time="0.012"><properties><property name="score" value="0.031019691841016606" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601318 [592] (title: Design of a Con) is a human-generated sample, misclassified as AI-generated with confidence 0.03101969&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601318 [592] (title: Design of a Con) is a human-generated sample, misclassified as AI-generated with confidence 0.03101969
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i425]" time="0.012"><properties><property name="score" value="0.07014609641405323" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601319 [1463] (title: Deep Learning a) is a human-generated sample, misclassified as AI-generated with confidence 0.0701461&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601319 [1463] (title: Deep Learning a) is a human-generated sample, misclassified as AI-generated with confidence 0.0701461
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i426]" time="0.012"><properties><property name="score" value="0.05789884500905229" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601324 [1166] (title: GAN-Based Semi-) is a human-generated sample, misclassified as AI-generated with confidence 0.05789885&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601324 [1166] (title: GAN-Based Semi-) is a human-generated sample, misclassified as AI-generated with confidence 0.05789885
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i427]" time="0.022"><properties><property name="score" value="0.00531263896133353" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i428]" time="0.012"><properties><property name="score" value="0.04249669696832492" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601331 [1053] (title: A Novel Paralle) is a human-generated sample, misclassified as AI-generated with confidence 0.0424967&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601331 [1053] (title: A Novel Paralle) is a human-generated sample, misclassified as AI-generated with confidence 0.0424967
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i429]" time="0.011"><properties><property name="score" value="0.0006647310675844675" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601338 [821] (title: Top-Down Synthe) is a human-generated sample, misclassified as AI-generated with confidence 0.00066473&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601338 [821] (title: Top-Down Synthe) is a human-generated sample, misclassified as AI-generated with confidence 0.00066473
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i430]" time="0.014"><properties><property name="score" value="0.01753063834808999" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601341 [771] (title: Effective Impro) is a human-generated sample, misclassified as AI-generated with confidence 0.01753064&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601341 [771] (title: Effective Impro) is a human-generated sample, misclassified as AI-generated with confidence 0.01753064
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i431]" time="0.012"><properties><property name="score" value="0.021817985813654683" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601343 [1307] (title: Automatically E) is a human-generated sample, misclassified as AI-generated with confidence 0.02181799&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601343 [1307] (title: Automatically E) is a human-generated sample, misclassified as AI-generated with confidence 0.02181799
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i432]" time="0.012"><properties><property name="score" value="0.03672213313483108" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601345 [1234] (title: Active Object D) is a human-generated sample, misclassified as AI-generated with confidence 0.03672213&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601345 [1234] (title: Active Object D) is a human-generated sample, misclassified as AI-generated with confidence 0.03672213
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i433]" time="0.012"><properties><property name="score" value="0.018433311064980362" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601346 [1498] (title: BLUEs and Relia) is a human-generated sample, misclassified as AI-generated with confidence 0.01843331&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601346 [1498] (title: BLUEs and Relia) is a human-generated sample, misclassified as AI-generated with confidence 0.01843331
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i434]" time="0.012"><properties><property name="score" value="0.011721885239324159" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601351 [1422] (title: PolSAR Image Se) is a human-generated sample, misclassified as AI-generated with confidence 0.01172189&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601351 [1422] (title: PolSAR Image Se) is a human-generated sample, misclassified as AI-generated with confidence 0.01172189
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i435]" time="0.012"><properties><property name="score" value="0.001819626470234148" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601354 [943] (title: Compensation of) is a human-generated sample, misclassified as AI-generated with confidence 0.00181963&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601354 [943] (title: Compensation of) is a human-generated sample, misclassified as AI-generated with confidence 0.00181963
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i436]" time="0.012"><properties><property name="score" value="0.11326275698883004" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601356 [1184] (title: Identification ) is a human-generated sample, misclassified as AI-generated with confidence 0.11326276&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601356 [1184] (title: Identification ) is a human-generated sample, misclassified as AI-generated with confidence 0.11326276
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i437]" time="0.022"><properties><property name="score" value="0.00664481368598254" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i438]" time="0.011"><properties><property name="score" value="0.029114734896584782" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601360 [993] (title: Response Predic) is a human-generated sample, misclassified as AI-generated with confidence 0.02911473&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601360 [993] (title: Response Predic) is a human-generated sample, misclassified as AI-generated with confidence 0.02911473
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i439]" time="0.012"><properties><property name="score" value="0.0024149635788617285" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i440]" time="0.012"><properties><property name="score" value="0.09255410788356899" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601366 [1470] (title: A Novel Framewo) is a human-generated sample, misclassified as AI-generated with confidence 0.09255411&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601366 [1470] (title: A Novel Framewo) is a human-generated sample, misclassified as AI-generated with confidence 0.09255411
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i441]" time="0.012"><properties><property name="score" value="0.02893586569179818" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601373 [958] (title: Snake: A Stocha) is a human-generated sample, misclassified as AI-generated with confidence 0.02893587&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601373 [958] (title: Snake: A Stocha) is a human-generated sample, misclassified as AI-generated with confidence 0.02893587
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i442]" time="0.012"><properties><property name="score" value="0.04347587381494655" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601376 [1240] (title: A Novel Key-Fra) is a human-generated sample, misclassified as AI-generated with confidence 0.04347587&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601376 [1240] (title: A Novel Key-Fra) is a human-generated sample, misclassified as AI-generated with confidence 0.04347587
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i443]" time="0.012"><properties><property name="score" value="0.03262721449455164" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601377 [1476] (title: Effective and E) is a human-generated sample, misclassified as AI-generated with confidence 0.03262721&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601377 [1476] (title: Effective and E) is a human-generated sample, misclassified as AI-generated with confidence 0.03262721
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i444]" time="0.017"><properties><property name="score" value="0.015967207404643524" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601379 [1136] (title: Stiffness Contr) is a human-generated sample, misclassified as AI-generated with confidence 0.01596721&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601379 [1136] (title: Stiffness Contr) is a human-generated sample, misclassified as AI-generated with confidence 0.01596721
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i445]" time="0.016"><properties><property name="score" value="0.01451606784385051" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601383 [1200] (title: Online Leader S) is a human-generated sample, misclassified as AI-generated with confidence 0.01451607&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601383 [1200] (title: Online Leader S) is a human-generated sample, misclassified as AI-generated with confidence 0.01451607
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i446]" time="0.013"><properties><property name="score" value="0.02562461165794283" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601385 [936] (title: A Square-Root-B) is a human-generated sample, misclassified as AI-generated with confidence 0.02562461&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601385 [936] (title: A Square-Root-B) is a human-generated sample, misclassified as AI-generated with confidence 0.02562461
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i447]" time="0.011"><properties><property name="score" value="0.053646409204277656" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601402 [1036] (title: Illumination-In) is a human-generated sample, misclassified as AI-generated with confidence 0.05364641&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601402 [1036] (title: Illumination-In) is a human-generated sample, misclassified as AI-generated with confidence 0.05364641
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i448]" time="0.012"><properties><property name="score" value="0.02278074569512789" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601418 [408] (title: The Classificat) is a human-generated sample, misclassified as AI-generated with confidence 0.02278075&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601418 [408] (title: The Classificat) is a human-generated sample, misclassified as AI-generated with confidence 0.02278075
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i449]" time="0.012"><properties><property name="score" value="0.007816286286227436" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601419 [408] (title: Registration of) is a human-generated sample, misclassified as AI-generated with confidence 0.00781629&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601419 [408] (title: Registration of) is a human-generated sample, misclassified as AI-generated with confidence 0.00781629
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i450]" time="0.011"><properties><property name="score" value="0.0173078468458987" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i451]" time="0.024"><properties><property name="score" value="0.043361945775169414" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601430 [1957] (title: Short-Term Fore) is a human-generated sample, misclassified as AI-generated with confidence 0.04336195&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601430 [1957] (title: Short-Term Fore) is a human-generated sample, misclassified as AI-generated with confidence 0.04336195
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i452]" time="0.012"><properties><property name="score" value="0.013192666713096601" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601432 [638] (title: The algorithm o) is a human-generated sample, misclassified as AI-generated with confidence 0.01319267&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601432 [638] (title: The algorithm o) is a human-generated sample, misclassified as AI-generated with confidence 0.01319267
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i453]" time="0.012"><properties><property name="score" value="0.008621150042187331" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i454]" time="0.013"><properties><property name="score" value="0.043846163430971696" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601446 [797] (title: Development and) is a human-generated sample, misclassified as AI-generated with confidence 0.04384616&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601446 [797] (title: Development and) is a human-generated sample, misclassified as AI-generated with confidence 0.04384616
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i455]" time="0.012"><properties><property name="score" value="0.02722252182710827" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601449 [725] (title: Synthesis of Al) is a human-generated sample, misclassified as AI-generated with confidence 0.02722252&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601449 [725] (title: Synthesis of Al) is a human-generated sample, misclassified as AI-generated with confidence 0.02722252
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i456]" time="0.012"><properties><property name="score" value="0.0645689819894546" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601451 [1328] (title: Evaluation of P) is a human-generated sample, misclassified as AI-generated with confidence 0.06456898&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601451 [1328] (title: Evaluation of P) is a human-generated sample, misclassified as AI-generated with confidence 0.06456898
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i457]" time="0.012"><properties><property name="score" value="0.002546252601542953" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i458]" time="0.011"><properties><property name="score" value="0.0017932659153674368" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i459]" time="0.012"><properties><property name="score" value="0.03220909847769238" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601456 [813] (title: Distribution of) is a human-generated sample, misclassified as AI-generated with confidence 0.0322091&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601456 [813] (title: Distribution of) is a human-generated sample, misclassified as AI-generated with confidence 0.0322091
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i460]" time="0.012"><properties><property name="score" value="0.027867318350410786" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601461 [1009] (title: Attribute-Based) is a human-generated sample, misclassified as AI-generated with confidence 0.02786732&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601461 [1009] (title: Attribute-Based) is a human-generated sample, misclassified as AI-generated with confidence 0.02786732
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i461]" time="0.011"><properties><property name="score" value="0.018599637010519032" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601462 [877] (title: Specifics of Da) is a human-generated sample, misclassified as AI-generated with confidence 0.01859964&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601462 [877] (title: Specifics of Da) is a human-generated sample, misclassified as AI-generated with confidence 0.01859964
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i462]" time="0.012"><properties><property name="score" value="0.002351016184531929" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i463]" time="0.012"><properties><property name="score" value="0.059937125107406564" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601472 [1208] (title: Information Dec) is a human-generated sample, misclassified as AI-generated with confidence 0.05993713&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601472 [1208] (title: Information Dec) is a human-generated sample, misclassified as AI-generated with confidence 0.05993713
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i464]" time="0.012"><properties><property name="score" value="0.07496368899051475" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601477 [1439] (title: Estimation of t) is a human-generated sample, misclassified as AI-generated with confidence 0.07496369&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601477 [1439] (title: Estimation of t) is a human-generated sample, misclassified as AI-generated with confidence 0.07496369
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i465]" time="0.012"><properties><property name="score" value="0.03766358765849298" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601478 [648] (title: Experimental St) is a human-generated sample, misclassified as AI-generated with confidence 0.03766359&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601478 [648] (title: Experimental St) is a human-generated sample, misclassified as AI-generated with confidence 0.03766359
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i466]" time="0.011"><properties><property name="score" value="0.047152911039022305" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601482 [649] (title: Analytical Depe) is a human-generated sample, misclassified as AI-generated with confidence 0.04715291&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601482 [649] (title: Analytical Depe) is a human-generated sample, misclassified as AI-generated with confidence 0.04715291
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i467]" time="0.012"><properties><property name="score" value="0.03682754616420869" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601485 [406] (title: Approaches to s) is a human-generated sample, misclassified as AI-generated with confidence 0.03682755&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601485 [406] (title: Approaches to s) is a human-generated sample, misclassified as AI-generated with confidence 0.03682755
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i468]" time="0.012"><properties><property name="score" value="0.012596831001321362" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601492 [1095] (title: Scalable and An) is a human-generated sample, misclassified as AI-generated with confidence 0.01259683&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601492 [1095] (title: Scalable and An) is a human-generated sample, misclassified as AI-generated with confidence 0.01259683
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i469]" time="0.023"><properties><property name="score" value="0.10763743210893782" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601497 [1656] (title: Estimation of P) is a human-generated sample, misclassified as AI-generated with confidence 0.10763743&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601497 [1656] (title: Estimation of P) is a human-generated sample, misclassified as AI-generated with confidence 0.10763743
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i470]" time="0.012"><properties><property name="score" value="0.04497733812879878" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601500 [576] (title: Network Simulat) is a human-generated sample, misclassified as AI-generated with confidence 0.04497734&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601500 [576] (title: Network Simulat) is a human-generated sample, misclassified as AI-generated with confidence 0.04497734
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i471]" time="0.012"><properties><property name="score" value="0.0856944920168079" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601526 [1248] (title: The Use of Non-) is a human-generated sample, misclassified as AI-generated with confidence 0.08569449&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601526 [1248] (title: The Use of Non-) is a human-generated sample, misclassified as AI-generated with confidence 0.08569449
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i472]" time="0.012"><properties><property name="score" value="0.005866140076676851" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601534 [623] (title: An Overview of ) is a human-generated sample, misclassified as AI-generated with confidence 0.00586614&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601534 [623] (title: An Overview of ) is a human-generated sample, misclassified as AI-generated with confidence 0.00586614
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i473]" time="0.023"><properties><property name="score" value="0.04486558257484293" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601546 [1770] (title: Partial Dischar) is a human-generated sample, misclassified as AI-generated with confidence 0.04486558&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601546 [1770] (title: Partial Dischar) is a human-generated sample, misclassified as AI-generated with confidence 0.04486558
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i474]" time="0.022"><properties><property name="score" value="0.06175376080227108" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601547 [1581] (title: Bridging the Ga) is a human-generated sample, misclassified as AI-generated with confidence 0.06175376&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601547 [1581] (title: Bridging the Ga) is a human-generated sample, misclassified as AI-generated with confidence 0.06175376
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i475]" time="0.012"><properties><property name="score" value="0.0378051498430787" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601550 [1270] (title: A Novel Method ) is a human-generated sample, misclassified as AI-generated with confidence 0.03780515&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601550 [1270] (title: A Novel Method ) is a human-generated sample, misclassified as AI-generated with confidence 0.03780515
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i476]" time="0.022"><properties><property name="score" value="0.053622007851006646" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601560 [1538] (title: Photovoltaic Mo) is a human-generated sample, misclassified as AI-generated with confidence 0.05362201&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601560 [1538] (title: Photovoltaic Mo) is a human-generated sample, misclassified as AI-generated with confidence 0.05362201
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i477]" time="0.012"><properties><property name="score" value="0.05988186711453625" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601563 [1202] (title: BrainSmart: Amb) is a human-generated sample, misclassified as AI-generated with confidence 0.05988187&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601563 [1202] (title: BrainSmart: Amb) is a human-generated sample, misclassified as AI-generated with confidence 0.05988187
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i478]" time="0.011"><properties><property name="score" value="0.04909568483329463" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601567 [849] (title: A new PV genera) is a human-generated sample, misclassified as AI-generated with confidence 0.04909568&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601567 [849] (title: A new PV genera) is a human-generated sample, misclassified as AI-generated with confidence 0.04909568
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i479]" time="0.012"><properties><property name="score" value="0.039183899336991956" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601579 [933] (title: Correlation Min) is a human-generated sample, misclassified as AI-generated with confidence 0.0391839&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601579 [933] (title: Correlation Min) is a human-generated sample, misclassified as AI-generated with confidence 0.0391839
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i480]" time="0.012"><properties><property name="score" value="0.032339267183739695" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601584 [1169] (title: A method of eva) is a human-generated sample, misclassified as AI-generated with confidence 0.03233927&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601584 [1169] (title: A method of eva) is a human-generated sample, misclassified as AI-generated with confidence 0.03233927
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i481]" time="0.014"><properties><property name="score" value="0.024121379584174374" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601588 [1229] (title: An Equivalent M) is a human-generated sample, misclassified as AI-generated with confidence 0.02412138&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601588 [1229] (title: An Equivalent M) is a human-generated sample, misclassified as AI-generated with confidence 0.02412138
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i482]" time="0.012"><properties><property name="score" value="0.0016297245656216042" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i483]" time="0.012"><properties><property name="score" value="0.022695420116175535" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i484]" time="0.013"><properties><property name="score" value="0.010030305159525188" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601601 [964] (title: Data-driven Res) is a human-generated sample, misclassified as AI-generated with confidence 0.01003031&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601601 [964] (title: Data-driven Res) is a human-generated sample, misclassified as AI-generated with confidence 0.01003031
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i485]" time="0.011"><properties><property name="score" value="0.04162125494923119" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601612 [801] (title: Key technologie) is a human-generated sample, misclassified as AI-generated with confidence 0.04162125&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601612 [801] (title: Key technologie) is a human-generated sample, misclassified as AI-generated with confidence 0.04162125
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i486]" time="0.012"><properties><property name="score" value="0.015194512425997786" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601615 [918] (title: Deep learning b) is a human-generated sample, misclassified as AI-generated with confidence 0.01519451&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601615 [918] (title: Deep learning b) is a human-generated sample, misclassified as AI-generated with confidence 0.01519451
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i487]" time="0.012"><properties><property name="score" value="0.01905601889870545" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601631 [878] (title: Optimal Configu) is a human-generated sample, misclassified as AI-generated with confidence 0.01905602&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601631 [878] (title: Optimal Configu) is a human-generated sample, misclassified as AI-generated with confidence 0.01905602
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i488]" time="0.011"><properties><property name="score" value="0.0003129455980022122" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i489]" time="0.013"><properties><property name="score" value="0.012343520265462038" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601654 [1295] (title: Equivalent Mode) is a human-generated sample, misclassified as AI-generated with confidence 0.01234352&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601654 [1295] (title: Equivalent Mode) is a human-generated sample, misclassified as AI-generated with confidence 0.01234352
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i490]" time="0.012"><properties><property name="score" value="0.0005778761281216571" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601661 [959] (title: Research on Pow) is a human-generated sample, misclassified as AI-generated with confidence 0.00057788&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601661 [959] (title: Research on Pow) is a human-generated sample, misclassified as AI-generated with confidence 0.00057788
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i491]" time="0.015"><properties><property name="score" value="0.01853851467498724" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601662 [867] (title: Temperature Ano) is a human-generated sample, misclassified as AI-generated with confidence 0.01853851&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601662 [867] (title: Temperature Ano) is a human-generated sample, misclassified as AI-generated with confidence 0.01853851
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i492]" time="0.013"><properties><property name="score" value="0.04744388381056619" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601666 [965] (title: A Fast Searchin) is a human-generated sample, misclassified as AI-generated with confidence 0.04744388&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601666 [965] (title: A Fast Searchin) is a human-generated sample, misclassified as AI-generated with confidence 0.04744388
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i493]" time="0.012"><properties><property name="score" value="0.04364271474806558" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601668 [1183] (title: Decision Tree-b) is a human-generated sample, misclassified as AI-generated with confidence 0.04364271&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601668 [1183] (title: Decision Tree-b) is a human-generated sample, misclassified as AI-generated with confidence 0.04364271
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i494]" time="0.013"><properties><property name="score" value="0.037675570694412075" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601669 [735] (title: New Dimensional) is a human-generated sample, misclassified as AI-generated with confidence 0.03767557&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601669 [735] (title: New Dimensional) is a human-generated sample, misclassified as AI-generated with confidence 0.03767557
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i495]" time="0.010"><properties><property name="score" value="0.002349002856943949" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i496]" time="0.072"><properties><property name="score" value="0.027259949727405332" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i497]" time="0.022"><properties><property name="score" value="0.002611530593907007" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i498]" time="0.026"><properties><property name="score" value="0.009267143792974709" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_human_jsonl[i499]" time="0.012"><properties><property name="score" value="0.043079330257739246" /></properties><failure message="AssertionError: samples/ieee-init.jsonl:8601703 [1249] (title: Optimal allocat) is a human-generated sample, misclassified as AI-generated with confidence 0.04307933&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/ieee-init.jsonl:8601703 [1249] (title: Optimal allocat) is a human-generated sample, misclassified as AI-generated with confidence 0.04307933
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i0]" time="0.013"><properties><property name="score" value="0.04075764038899421" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i1]" time="0.012"><properties><property name="score" value="0.06893990768391811" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i2]" time="0.015"><properties><property name="score" value="0.0426567030180125" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i3]" time="0.012"><properties><property name="score" value="0.09450831708044949" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i4]" time="0.011"><properties><property name="score" value="0.031017654675202788" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i5]" time="0.013"><properties><property name="score" value="0.06262104223129783" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i6]" time="0.012"><properties><property name="score" value="0.05546667827620588" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i7]" time="0.011"><properties><property name="score" value="0.0445378205513991" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i8]" time="0.011"><properties><property name="score" value="0.020151449039168545" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i9]" time="0.011"><properties><property name="score" value="0.005212228037271283" /></properties><failure message="AssertionError: samples/ieee-chatgpt-generation.jsonl:8600034 (title: Secrecy Anti-jamming Game Learning in D2D Underlay) is an LLM-generated sample, misclassified as human-generated with confidence 0.00521223&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-generation.jsonl:8600034 (title: Secrecy Anti-jamming Game Learning in D2D Underlay) is an LLM-generated sample, misclassified as human-generated with confidence 0.00521223
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i10]" time="0.011"><properties><property name="score" value="0.08715807171289502" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i11]" time="0.011"><properties><property name="score" value="0.05166272233347313" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i12]" time="0.011"><properties><property name="score" value="0.0018760230369063713" /></properties><failure message="AssertionError: samples/ieee-chatgpt-generation.jsonl:8600039 (title: Unsupervised Ego-Motion and Dense Depth Estimation) is an LLM-generated sample, misclassified as human-generated with confidence 0.00187602&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-generation.jsonl:8600039 (title: Unsupervised Ego-Motion and Dense Depth Estimation) is an LLM-generated sample, misclassified as human-generated with confidence 0.00187602
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i13]" time="0.012"><properties><property name="score" value="0.025634887846492926" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i14]" time="0.012"><properties><property name="score" value="0.04943800487837713" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i15]" time="0.014"><properties><property name="score" value="0.05691389337300623" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i16]" time="0.012"><properties><property name="score" value="0.02604505179167882" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i17]" time="0.012"><properties><property name="score" value="0.059374164195047374" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i18]" time="0.013"><properties><property name="score" value="0.023640049386675077" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i19]" time="0.011"><properties><property name="score" value="0.1198461651720828" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i20]" time="0.013"><properties><property name="score" value="0.02604505179167882" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i21]" time="0.011"><properties><property name="score" value="0.06436311018648833" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i22]" time="0.012"><properties><property name="score" value="0.10006755333676343" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i23]" time="0.011"><properties><property name="score" value="0.030065322265515615" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i24]" time="0.011"><properties><property name="score" value="0.08061643249492878" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i25]" time="0.012"><properties><property name="score" value="0.07331190818917754" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i26]" time="0.011"><properties><property name="score" value="0.07748561929388775" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i27]" time="0.011"><properties><property name="score" value="0.05822733795834423" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i28]" time="0.012"><properties><property name="score" value="0.016731745364861883" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i29]" time="0.011"><properties><property name="score" value="0.07178726314656658" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i30]" time="0.011"><properties><property name="score" value="0.0542255718018414" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i31]" time="0.013"><properties><property name="score" value="0.05597762439292775" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i32]" time="0.011"><properties><property name="score" value="0.03384736945770683" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i33]" time="0.011"><properties><property name="score" value="0.028436247976837414" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i34]" time="0.012"><properties><property name="score" value="0.061353264725888534" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i35]" time="0.012"><properties><property name="score" value="0.07213120128862216" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i36]" time="0.013"><properties><property name="score" value="0.09309503565433919" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i37]" time="0.011"><properties><property name="score" value="0.01852460851692328" /></properties><failure message="AssertionError: samples/ieee-chatgpt-generation.jsonl:8600094 (title: NLOS Detection and Mitigation for UWB/IMU Fusion S) is an LLM-generated sample, misclassified as human-generated with confidence 0.01852461&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-generation.jsonl:8600094 (title: NLOS Detection and Mitigation for UWB/IMU Fusion S) is an LLM-generated sample, misclassified as human-generated with confidence 0.01852461
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i38]" time="0.011"><properties><property name="score" value="0.0818235684117985" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i39]" time="0.012"><properties><property name="score" value="0.08960733031419532" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i40]" time="0.011"><properties><property name="score" value="0.0548003726174251" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i41]" time="0.011"><properties><property name="score" value="0.04607140330026849" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i42]" time="0.012"><properties><property name="score" value="0.057872686521659444" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i43]" time="0.011"><properties><property name="score" value="0.16406653821285921" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i44]" time="0.012"><properties><property name="score" value="0.05890651932229374" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i45]" time="0.012"><properties><property name="score" value="0.04642075323825767" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i46]" time="0.011"><properties><property name="score" value="0.045395604288039904" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i47]" time="0.014"><properties><property name="score" value="0.039338338531808015" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i48]" time="0.011"><properties><property name="score" value="0.07186586481654489" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i49]" time="0.014"><properties><property name="score" value="0.04909568483329463" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i50]" time="0.012"><properties><property name="score" value="0.04875354306000945" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i51]" time="0.011"><properties><property name="score" value="0.07622147892691888" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i52]" time="0.015"><properties><property name="score" value="0.0374068260173388" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i53]" time="0.011"><properties><property name="score" value="0.10576017214095068" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i54]" time="0.011"><properties><property name="score" value="0.07955463522248962" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i55]" time="0.011"><properties><property name="score" value="0.09016453736613861" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i56]" time="0.011"><properties><property name="score" value="0.036670228606855426" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i57]" time="0.012"><properties><property name="score" value="0.09374985120109702" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i58]" time="0.011"><properties><property name="score" value="0.038775645528454916" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i59]" time="0.011"><properties><property name="score" value="0.17915863265891008" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i60]" time="0.011"><properties><property name="score" value="0.059929897088129014" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i61]" time="0.011"><properties><property name="score" value="0.01096375006599537" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i62]" time="0.012"><properties><property name="score" value="0.048582261863816534" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i63]" time="0.011"><properties><property name="score" value="0.06640602818286379" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i64]" time="0.011"><properties><property name="score" value="0.06259586364815162" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i65]" time="0.012"><properties><property name="score" value="0.07381982162565359" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i66]" time="0.011"><properties><property name="score" value="0.03664410803940754" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i67]" time="0.011"><properties><property name="score" value="0.11340676424071994" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i68]" time="0.012"><properties><property name="score" value="0.025769588590995873" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i69]" time="0.011"><properties><property name="score" value="0.03465557096518079" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i70]" time="0.011"><properties><property name="score" value="0.010466551151355752" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i71]" time="0.013"><properties><property name="score" value="0.06935138987376743" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i72]" time="0.011"><properties><property name="score" value="0.03746538459478699" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i73]" time="0.011"><properties><property name="score" value="0.06436745585119419" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i74]" time="0.011"><properties><property name="score" value="0.126400322686826" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i75]" time="0.011"><properties><property name="score" value="0.04675947200299402" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i76]" time="0.011"><properties><property name="score" value="0.03916322517094062" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i77]" time="0.011"><properties><property name="score" value="0.021889131460639533" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i78]" time="0.011"><properties><property name="score" value="0.052631823236665154" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i79]" time="0.012"><properties><property name="score" value="0.03431823908035536" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i80]" time="0.011"><properties><property name="score" value="0.05461266113475083" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i81]" time="0.021"><properties><property name="score" value="0.09103719684357749" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i82]" time="0.011"><properties><property name="score" value="0.049094580863623394" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i83]" time="0.011"><properties><property name="score" value="0.06173995133132437" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i84]" time="0.011"><properties><property name="score" value="0.055921717390006176" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i85]" time="0.011"><properties><property name="score" value="0.07710635545729039" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i86]" time="0.011"><properties><property name="score" value="0.07213120128862216" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i87]" time="0.012"><properties><property name="score" value="0.07633997794362357" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i88]" time="0.011"><properties><property name="score" value="0.02130815780753692" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i89]" time="0.012"><properties><property name="score" value="0.01135810137053439" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i90]" time="0.011"><properties><property name="score" value="0.07486615483859316" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i91]" time="0.011"><properties><property name="score" value="0.057741761241042466" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i92]" time="0.014"><properties><property name="score" value="0.057928430083525884" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i93]" time="0.011"><properties><property name="score" value="0.14441109923963857" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i94]" time="0.013"><properties><property name="score" value="0.02218314299665125" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i95]" time="0.021"><properties><property name="score" value="0.14206904758113661" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i96]" time="0.011"><properties><property name="score" value="0.028061664793238572" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i97]" time="0.011"><properties><property name="score" value="0.04118919598899429" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i98]" time="0.011"><properties><property name="score" value="0.06924316826634747" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i99]" time="0.012"><properties><property name="score" value="0.09677753462004612" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i100]" time="0.011"><properties><property name="score" value="0.08858644291350748" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i101]" time="0.011"><properties><property name="score" value="0.06897235618150699" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i102]" time="0.012"><properties><property name="score" value="0.09511337871865866" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i103]" time="0.011"><properties><property name="score" value="0.022678544917925647" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i104]" time="0.013"><properties><property name="score" value="0.05819674405484476" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i105]" time="0.012"><properties><property name="score" value="0.06102362285387297" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i106]" time="0.011"><properties><property name="score" value="0.09521394773525649" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i107]" time="0.012"><properties><property name="score" value="0.040600579785687874" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i108]" time="0.011"><properties><property name="score" value="0.009347600439157944" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i109]" time="0.011"><properties><property name="score" value="0.10041694235805587" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i110]" time="0.011"><properties><property name="score" value="0.11829783915748293" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i111]" time="0.012"><properties><property name="score" value="0.05421379307973817" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i112]" time="0.011"><properties><property name="score" value="0.00569055545544539" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i113]" time="0.011"><properties><property name="score" value="0.057872686521659444" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i114]" time="0.011"><properties><property name="score" value="0.01753063834808999" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i115]" time="0.012"><properties><property name="score" value="0.03079911652439371" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i116]" time="0.011"><properties><property name="score" value="0.04165192501314752" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i117]" time="0.011"><properties><property name="score" value="0.07066226194767378" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i118]" time="0.014"><properties><property name="score" value="0.08117618413155947" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i119]" time="0.012"><properties><property name="score" value="0.049039316909160746" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i120]" time="0.011"><properties><property name="score" value="0.049947308195247286" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i121]" time="0.011"><properties><property name="score" value="0.07817319638233622" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i122]" time="0.011"><properties><property name="score" value="0.07977362770948404" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i123]" time="0.014"><properties><property name="score" value="0.051656455577753224" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i124]" time="0.011"><properties><property name="score" value="0.07523622893855775" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i125]" time="0.011"><properties><property name="score" value="0.020668658208039536" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i126]" time="0.011"><properties><property name="score" value="0.055985408481257215" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i127]" time="0.011"><properties><property name="score" value="0.12319024290807223" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i128]" time="0.011"><properties><property name="score" value="0.028398379669197782" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i129]" time="0.011"><properties><property name="score" value="0.07223883105191287" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i130]" time="0.011"><properties><property name="score" value="0.12056423984516118" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i131]" time="0.011"><properties><property name="score" value="0.05978315914266785" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i132]" time="0.014"><properties><property name="score" value="0.08054552096971035" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i133]" time="0.012"><properties><property name="score" value="0.06904841732157663" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i134]" time="0.015"><properties><property name="score" value="0.09607278299471611" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i135]" time="0.013"><properties><property name="score" value="0.04499362092457715" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i136]" time="0.011"><properties><property name="score" value="0.029409577392980335" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i137]" time="0.010"><properties><property name="score" value="0.05925148947055758" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i138]" time="0.011"><properties><property name="score" value="0.07643992120562326" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i139]" time="0.011"><properties><property name="score" value="0.03646205030942706" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i140]" time="0.011"><properties><property name="score" value="0.03779277149019711" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i141]" time="0.011"><properties><property name="score" value="0.02773965733557371" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i142]" time="0.012"><properties><property name="score" value="0.05410725695697005" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i143]" time="0.011"><properties><property name="score" value="0.0709616414617642" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i144]" time="0.011"><properties><property name="score" value="0.05924016805374244" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i145]" time="0.012"><properties><property name="score" value="0.03225058544734094" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i146]" time="0.011"><properties><property name="score" value="0.021553282714659172" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i147]" time="0.012"><properties><property name="score" value="0.07725121964443638" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i148]" time="0.011"><properties><property name="score" value="0.04903870010811717" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i149]" time="0.013"><properties><property name="score" value="0.08476107563104018" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i150]" time="0.011"><properties><property name="score" value="0.020973043801081337" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i151]" time="0.011"><properties><property name="score" value="0.06737419273218914" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i152]" time="0.010"><properties><property name="score" value="0.08261298022381736" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i153]" time="0.011"><properties><property name="score" value="0.058237580026043734" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i154]" time="0.010"><properties><property name="score" value="0.05433804924648844" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i155]" time="0.014"><properties><property name="score" value="0.06033048840028976" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i156]" time="0.012"><properties><property name="score" value="0.02347106631325202" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i157]" time="0.010"><properties><property name="score" value="0.019079128105708998" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i158]" time="0.012"><properties><property name="score" value="0.029114734896584782" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i159]" time="0.011"><properties><property name="score" value="0.05713757626762406" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i160]" time="0.011"><properties><property name="score" value="0.01541501836084258" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i161]" time="0.011"><properties><property name="score" value="0.08910918425288905" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i162]" time="0.011"><properties><property name="score" value="0.03779338493843376" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i163]" time="0.012"><properties><property name="score" value="0.07638048125352959" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i164]" time="0.013"><properties><property name="score" value="0.06911295492323388" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i165]" time="0.011"><properties><property name="score" value="0.029352999930143397" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i166]" time="0.012"><properties><property name="score" value="0.014479243796777785" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i167]" time="0.011"><properties><property name="score" value="0.07893526407014573" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i168]" time="0.011"><properties><property name="score" value="0.08093153854144552" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i169]" time="0.012"><properties><property name="score" value="0.03115550447246318" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i170]" time="0.011"><properties><property name="score" value="0.028278759054006786" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i171]" time="0.011"><properties><property name="score" value="0.07454302573323135" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i172]" time="0.012"><properties><property name="score" value="0.08383745540604393" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i173]" time="0.011"><properties><property name="score" value="0.09940329236933221" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i174]" time="0.012"><properties><property name="score" value="0.08229473227992501" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i175]" time="0.011"><properties><property name="score" value="0.07299059853086431" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i176]" time="0.012"><properties><property name="score" value="0.07533869703708063" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i177]" time="0.012"><properties><property name="score" value="0.03302087528692654" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i178]" time="0.011"><properties><property name="score" value="0.04584468859256674" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i179]" time="0.011"><properties><property name="score" value="0.0670615339419478" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i180]" time="0.012"><properties><property name="score" value="0.031333480014750226" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i181]" time="0.011"><properties><property name="score" value="0.06953597899779229" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i182]" time="0.014"><properties><property name="score" value="0.02796441332573174" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i183]" time="0.011"><properties><property name="score" value="0.09218231581724634" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i184]" time="0.011"><properties><property name="score" value="0.029290139395865555" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i185]" time="0.011"><properties><property name="score" value="0.08889545753894001" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i186]" time="0.010"><properties><property name="score" value="0.0541246225112757" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i187]" time="0.010"><properties><property name="score" value="0.0497805033346177" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i188]" time="0.012"><properties><property name="score" value="0.055074463898591164" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i189]" time="0.011"><properties><property name="score" value="0.05694331020189569" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i190]" time="0.011"><properties><property name="score" value="0.004493940841049371" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i191]" time="0.011"><properties><property name="score" value="0.053186541275729304" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i192]" time="0.011"><properties><property name="score" value="0.03803854949256946" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i193]" time="0.013"><properties><property name="score" value="0.059684184722830214" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i194]" time="0.011"><properties><property name="score" value="0.06465763005536473" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i195]" time="0.010"><properties><property name="score" value="0.04139694748713785" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i196]" time="0.012"><properties><property name="score" value="0.010175253497590742" /></properties><failure message="AssertionError: samples/ieee-chatgpt-generation.jsonl:8600517 (title: A Pre-Compensation Method for Digital-to-Analog Co) is an LLM-generated sample, misclassified as human-generated with confidence 0.01017525&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-generation.jsonl:8600517 (title: A Pre-Compensation Method for Digital-to-Analog Co) is an LLM-generated sample, misclassified as human-generated with confidence 0.01017525
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i197]" time="0.011"><properties><property name="score" value="0.05120112705269686" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i198]" time="0.010"><properties><property name="score" value="0.04681461943920051" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i199]" time="0.011"><properties><property name="score" value="0.04118104636799402" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i200]" time="0.010"><properties><property name="score" value="0.06412704222215071" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i201]" time="0.010"><properties><property name="score" value="0.11903448816004136" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i202]" time="0.010"><properties><property name="score" value="0.04100707251005198" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i203]" time="0.011"><properties><property name="score" value="0.0453308518149953" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i204]" time="0.010"><properties><property name="score" value="0.04423528321345627" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i205]" time="0.011"><properties><property name="score" value="0.05444433069916044" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i206]" time="0.011"><properties><property name="score" value="0.08066898591024829" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i207]" time="0.012"><properties><property name="score" value="0.05597762439292775" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i208]" time="0.011"><properties><property name="score" value="0.10483628878680262" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i209]" time="0.010"><properties><property name="score" value="0.05724016964587797" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i210]" time="0.011"><properties><property name="score" value="0.07855372120706416" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i211]" time="0.011"><properties><property name="score" value="0.031866535054031475" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i212]" time="0.010"><properties><property name="score" value="0.022464345924105844" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i213]" time="0.010"><properties><property name="score" value="0.05842943896615105" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i214]" time="0.011"><properties><property name="score" value="0.031077095045300984" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i215]" time="0.010"><properties><property name="score" value="0.023046998687942244" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i216]" time="0.011"><properties><property name="score" value="0.10950558853530246" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i217]" time="0.011"><properties><property name="score" value="0.07098553560802578" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i218]" time="0.010"><properties><property name="score" value="0.08530694967950891" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i219]" time="0.012"><properties><property name="score" value="0.043505156012002555" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i220]" time="0.010"><properties><property name="score" value="0.05291864424771808" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i221]" time="0.010"><properties><property name="score" value="0.022622118864512597" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i222]" time="0.012"><properties><property name="score" value="0.05018005379193724" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i223]" time="0.011"><properties><property name="score" value="0.08868946328387528" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i224]" time="0.010"><properties><property name="score" value="0.08515101369211964" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i225]" time="0.010"><properties><property name="score" value="0.05827282909075526" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i226]" time="0.010"><properties><property name="score" value="0.0541246225112757" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i227]" time="0.010"><properties><property name="score" value="0.08033484117392353" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i228]" time="0.010"><properties><property name="score" value="0.03836661130400265" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i229]" time="0.010"><properties><property name="score" value="0.07802484287798839" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i230]" time="0.010"><properties><property name="score" value="0.03848355798999892" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i231]" time="0.010"><properties><property name="score" value="0.05604921958795028" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i232]" time="0.010"><properties><property name="score" value="0.04475061921005996" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i233]" time="0.013"><properties><property name="score" value="0.05177924784941501" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i234]" time="0.013"><properties><property name="score" value="0.07261507329026085" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i235]" time="0.011"><properties><property name="score" value="0.06685876342701347" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i236]" time="0.013"><properties><property name="score" value="0.04201254835768942" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i237]" time="0.011"><properties><property name="score" value="0.06379770536942275" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i238]" time="0.022"><properties><property name="score" value="0.12971398962078884" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i239]" time="0.010"><properties><property name="score" value="0.06031800876103999" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i240]" time="0.010"><properties><property name="score" value="0.06607605039607445" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i241]" time="0.011"><properties><property name="score" value="0.029927288910125416" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i242]" time="0.011"><properties><property name="score" value="0.100422707734249" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i243]" time="0.011"><properties><property name="score" value="0.060441357544177565" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i244]" time="0.012"><properties><property name="score" value="0.03695053021268713" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i245]" time="0.010"><properties><property name="score" value="0.06077360231719342" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i246]" time="0.011"><properties><property name="score" value="0.04367289490631032" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i247]" time="0.011"><properties><property name="score" value="0.0665508750859245" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i248]" time="0.010"><properties><property name="score" value="0.05819674405484476" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i249]" time="0.011"><properties><property name="score" value="0.10173084163317969" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i250]" time="0.011"><properties><property name="score" value="0.05285771726168398" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i251]" time="0.011"><properties><property name="score" value="0.06891814947637132" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i252]" time="0.012"><properties><property name="score" value="0.04539921676081082" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i253]" time="0.011"><properties><property name="score" value="0.04198376817066696" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i254]" time="0.012"><properties><property name="score" value="0.07368629081531686" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i255]" time="0.015"><properties><property name="score" value="0.060028699150449505" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i256]" time="0.011"><properties><property name="score" value="0.06980633036369754" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i257]" time="0.012"><properties><property name="score" value="0.05097226368596108" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i258]" time="0.011"><properties><property name="score" value="0.014948174476742127" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i259]" time="0.011"><properties><property name="score" value="0.09561837002627915" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i260]" time="0.015"><properties><property name="score" value="0.05233364332838941" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i261]" time="0.011"><properties><property name="score" value="0.0663489662000416" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i262]" time="0.010"><properties><property name="score" value="0.013863400763153289" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i263]" time="0.011"><properties><property name="score" value="0.0157122124856246" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i264]" time="0.010"><properties><property name="score" value="0.019693555166433452" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i265]" time="0.010"><properties><property name="score" value="0.06560223680952548" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i266]" time="0.011"><properties><property name="score" value="0.02306965252577564" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i267]" time="0.010"><properties><property name="score" value="0.07261507329026085" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i268]" time="0.010"><properties><property name="score" value="0.06807424150894192" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i269]" time="0.011"><properties><property name="score" value="0.07561306335864826" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i270]" time="0.011"><properties><property name="score" value="0.0448828013561664" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i271]" time="0.011"><properties><property name="score" value="0.060028699150449505" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i272]" time="0.011"><properties><property name="score" value="0.024978998544450404" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i273]" time="0.010"><properties><property name="score" value="0.05307870226926026" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i274]" time="0.011"><properties><property name="score" value="0.09071542322553472" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i275]" time="0.011"><properties><property name="score" value="0.10815772142916957" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i276]" time="0.010"><properties><property name="score" value="0.023316903319964277" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i277]" time="0.011"><properties><property name="score" value="0.12028192389096914" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i278]" time="0.011"><properties><property name="score" value="0.025228751817613082" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i279]" time="0.010"><properties><property name="score" value="0.06492819282758466" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i280]" time="0.013"><properties><property name="score" value="0.03040252757998685" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i281]" time="0.010"><properties><property name="score" value="0.028811734286526836" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i282]" time="0.011"><properties><property name="score" value="0.012496045385157473" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i283]" time="0.011"><properties><property name="score" value="0.06866837434529316" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i284]" time="0.010"><properties><property name="score" value="0.08030081128866384" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i285]" time="0.010"><properties><property name="score" value="0.03050211301898198" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i286]" time="0.011"><properties><property name="score" value="0.07593229609825225" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i287]" time="0.010"><properties><property name="score" value="0.008730590890032586" /></properties><failure message="AssertionError: samples/ieee-chatgpt-generation.jsonl:8600838 (title: NavREn-Rl: Learning to fly in real environment via) is an LLM-generated sample, misclassified as human-generated with confidence 0.00873059&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-generation.jsonl:8600838 (title: NavREn-Rl: Learning to fly in real environment via) is an LLM-generated sample, misclassified as human-generated with confidence 0.00873059
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i288]" time="0.010"><properties><property name="score" value="0.008118448492100416" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i289]" time="0.010"><properties><property name="score" value="0.0514244775390027" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i290]" time="0.010"><properties><property name="score" value="0.03174819120847494" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i291]" time="0.010"><properties><property name="score" value="0.07234350721525562" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i292]" time="0.011"><properties><property name="score" value="0.019236450122556148" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i293]" time="0.010"><properties><property name="score" value="0.019419749745402948" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i294]" time="0.012"><properties><property name="score" value="0.059473852329527954" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i295]" time="0.012"><properties><property name="score" value="0.04065869769507291" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i296]" time="0.011"><properties><property name="score" value="0.022367883665891375" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i297]" time="0.011"><properties><property name="score" value="0.03214978239380328" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i298]" time="0.010"><properties><property name="score" value="0.08533203936289535" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i299]" time="0.010"><properties><property name="score" value="0.04527708509195483" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i300]" time="0.011"><properties><property name="score" value="0.05861740747691857" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i301]" time="0.010"><properties><property name="score" value="0.026692786196486074" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i302]" time="0.010"><properties><property name="score" value="0.06167083158457598" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i303]" time="0.012"><properties><property name="score" value="0.01056721712412445" /></properties><failure message="AssertionError: samples/ieee-chatgpt-generation.jsonl:8600862 (title: Large Data Download Method for IoT Machines Using ) is an LLM-generated sample, misclassified as human-generated with confidence 0.01056722&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-generation.jsonl:8600862 (title: Large Data Download Method for IoT Machines Using ) is an LLM-generated sample, misclassified as human-generated with confidence 0.01056722
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i304]" time="0.010"><properties><property name="score" value="0.04585095395052141" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i305]" time="0.011"><properties><property name="score" value="0.048810360216861026" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i306]" time="0.011"><properties><property name="score" value="0.08577399387356133" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i307]" time="0.010"><properties><property name="score" value="0.04201982741711685" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i308]" time="0.010"><properties><property name="score" value="0.05257944073266407" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i309]" time="0.012"><properties><property name="score" value="0.04207780423545193" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i310]" time="0.011"><properties><property name="score" value="0.03391727906523756" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i311]" time="0.011"><properties><property name="score" value="0.039465027668211805" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i312]" time="0.011"><properties><property name="score" value="0.058373831964686484" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i313]" time="0.011"><properties><property name="score" value="0.13529342198093652" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i314]" time="0.014"><properties><property name="score" value="0.05498240635531948" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i315]" time="0.011"><properties><property name="score" value="0.0504605046896478" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i316]" time="0.012"><properties><property name="score" value="0.023713026905952672" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i317]" time="0.011"><properties><property name="score" value="0.01774807093615216" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i318]" time="0.010"><properties><property name="score" value="0.007033672843093086" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i319]" time="0.010"><properties><property name="score" value="0.046585327474335525" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i320]" time="0.010"><properties><property name="score" value="0.049152223882120705" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i321]" time="0.010"><properties><property name="score" value="0.08407000408245935" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i322]" time="0.086"><properties><property name="score" value="0.007222160444864123" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i323]" time="0.011"><properties><property name="score" value="0.008618863846693303" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i324]" time="0.011"><properties><property name="score" value="0.039756084841957584" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i325]" time="0.011"><properties><property name="score" value="0.026067329677598172" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i326]" time="0.011"><properties><property name="score" value="0.06079000445390892" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i327]" time="0.011"><properties><property name="score" value="0.06787790348248834" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i328]" time="0.011"><properties><property name="score" value="0.08508615644293172" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i329]" time="0.011"><properties><property name="score" value="0.06924316826634747" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i330]" time="0.011"><properties><property name="score" value="0.042592240149819105" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i331]" time="0.012"><properties><property name="score" value="0.014035195653344656" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i332]" time="0.010"><properties><property name="score" value="0.02432440297715499" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i333]" time="0.010"><properties><property name="score" value="0.04653036601174887" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i334]" time="0.011"><properties><property name="score" value="0.05696097565967562" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i335]" time="0.010"><properties><property name="score" value="0.0664389774698626" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i336]" time="0.011"><properties><property name="score" value="0.04163662302387183" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i337]" time="0.011"><properties><property name="score" value="0.06930958396653142" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i338]" time="0.010"><properties><property name="score" value="0.135022680290392" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i339]" time="0.011"><properties><property name="score" value="0.09203805351049033" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i340]" time="0.010"><properties><property name="score" value="0.025117618509989947" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i341]" time="0.010"><properties><property name="score" value="0.08191168517854286" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i342]" time="0.011"><properties><property name="score" value="0.036917437275379994" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i343]" time="0.010"><properties><property name="score" value="0.09521449915199098" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i344]" time="0.010"><properties><property name="score" value="0.02867549821666815" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i345]" time="0.011"><properties><property name="score" value="0.03625932484506711" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i346]" time="0.010"><properties><property name="score" value="0.02240372440750482" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i347]" time="0.011"><properties><property name="score" value="0.07132734078082459" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i348]" time="0.011"><properties><property name="score" value="0.05545945922668105" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i349]" time="0.011"><properties><property name="score" value="0.04646811934959638" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i350]" time="0.011"><properties><property name="score" value="0.06033048840028976" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i351]" time="0.011"><properties><property name="score" value="0.03190756058495392" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i352]" time="0.011"><properties><property name="score" value="0.10524542554553595" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i353]" time="0.012"><properties><property name="score" value="0.1369739815343607" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i354]" time="0.011"><properties><property name="score" value="0.10548923910721486" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i355]" time="0.011"><properties><property name="score" value="0.05211539371891005" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i356]" time="0.012"><properties><property name="score" value="0.0473326751711689" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i357]" time="0.011"><properties><property name="score" value="0.0694277353035122" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i358]" time="0.012"><properties><property name="score" value="0.04274043590459797" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i359]" time="0.012"><properties><property name="score" value="0.063011749192321" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i360]" time="0.011"><properties><property name="score" value="0.04909531605011841" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i361]" time="0.011"><properties><property name="score" value="0.017414257256692656" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i362]" time="0.011"><properties><property name="score" value="0.09713244479924721" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i363]" time="0.010"><properties><property name="score" value="0.05531280607262157" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i364]" time="0.012"><properties><property name="score" value="0.06423670156742167" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i365]" time="0.011"><properties><property name="score" value="0.06405029432684506" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i366]" time="0.011"><properties><property name="score" value="0.05889557278498936" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i367]" time="0.014"><properties><property name="score" value="0.03549413094919296" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i368]" time="0.012"><properties><property name="score" value="0.0944576939519215" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i369]" time="0.012"><properties><property name="score" value="0.06349104132984396" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i370]" time="0.011"><properties><property name="score" value="0.024669913793401443" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i371]" time="0.011"><properties><property name="score" value="0.05309272058032777" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i372]" time="0.011"><properties><property name="score" value="0.03415350219700408" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i373]" time="0.012"><properties><property name="score" value="0.01264968104645603" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i374]" time="0.011"><properties><property name="score" value="0.022225155612909386" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i375]" time="0.012"><properties><property name="score" value="0.03711379331463038" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i376]" time="0.011"><properties><property name="score" value="0.01939271296789169" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i377]" time="0.011"><properties><property name="score" value="0.10005689972273357" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i378]" time="0.013"><properties><property name="score" value="0.06778975503620721" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i379]" time="0.011"><properties><property name="score" value="0.08455302837837575" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i380]" time="0.012"><properties><property name="score" value="0.061092014627578406" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i381]" time="0.011"><properties><property name="score" value="0.03265857478501055" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i382]" time="0.011"><properties><property name="score" value="0.12422444769075724" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i383]" time="0.013"><properties><property name="score" value="0.054051024406681636" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i384]" time="0.011"><properties><property name="score" value="0.036110292035945735" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i385]" time="0.011"><properties><property name="score" value="0.07746354131690225" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i386]" time="0.012"><properties><property name="score" value="0.028795025413330677" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i387]" time="0.011"><properties><property name="score" value="0.09834792676651771" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i388]" time="0.011"><properties><property name="score" value="0.05494890997705948" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i389]" time="0.012"><properties><property name="score" value="0.017770265598043844" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i390]" time="0.011"><properties><property name="score" value="0.049781613230936994" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i391]" time="0.015"><properties><property name="score" value="0.05512367714137767" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i392]" time="0.012"><properties><property name="score" value="0.048182676628910626" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i393]" time="0.012"><properties><property name="score" value="0.04185325605654544" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i394]" time="0.011"><properties><property name="score" value="0.0314927239657814" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i395]" time="0.010"><properties><property name="score" value="0.016608586464994524" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i396]" time="0.014"><properties><property name="score" value="0.04155544233029884" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i397]" time="0.012"><properties><property name="score" value="0.007801394256234939" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i398]" time="0.011"><properties><property name="score" value="0.07555980718879085" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i399]" time="0.015"><properties><property name="score" value="0.04652796535139503" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i400]" time="0.011"><properties><property name="score" value="0.03400819773569097" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i401]" time="0.011"><properties><property name="score" value="0.03781640086690796" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i402]" time="0.012"><properties><property name="score" value="0.04254104956460392" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i403]" time="0.011"><properties><property name="score" value="0.08268113582756698" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i404]" time="0.013"><properties><property name="score" value="0.04257890897139682" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i405]" time="0.011"><properties><property name="score" value="0.0636168010861693" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i406]" time="0.011"><properties><property name="score" value="0.02934986654840044" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i407]" time="0.012"><properties><property name="score" value="0.055593324597386085" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i408]" time="0.011"><properties><property name="score" value="0.032505436178253966" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i409]" time="0.011"><properties><property name="score" value="0.04709909860177719" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i410]" time="0.011"><properties><property name="score" value="0.10349011054009427" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i411]" time="0.011"><properties><property name="score" value="0.05233364332838941" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i412]" time="0.011"><properties><property name="score" value="0.09491005208937531" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i413]" time="0.011"><properties><property name="score" value="0.04579363740130904" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i414]" time="0.011"><properties><property name="score" value="0.008087735224515669" /></properties><failure message="AssertionError: samples/ieee-chatgpt-generation.jsonl:8601281 (title: Virtual Private Network’s Impact on Network Perfor) is an LLM-generated sample, misclassified as human-generated with confidence 0.00808774&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-generation.jsonl:8601281 (title: Virtual Private Network’s Impact on Network Perfor) is an LLM-generated sample, misclassified as human-generated with confidence 0.00808774
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i415]" time="0.016"><properties><property name="score" value="0.14866528957726" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i416]" time="0.011"><properties><property name="score" value="0.009801573461937219" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i417]" time="0.014"><properties><property name="score" value="0.059195860809468925" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i418]" time="0.012"><properties><property name="score" value="0.01451643568473604" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i419]" time="0.012"><properties><property name="score" value="0.029251461982671145" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i420]" time="0.012"><properties><property name="score" value="0.05914021695739846" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i421]" time="0.011"><properties><property name="score" value="0.05336507881670283" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i422]" time="0.012"><properties><property name="score" value="0.05750955286674131" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i423]" time="0.012"><properties><property name="score" value="0.006666973038477186" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i424]" time="0.011"><properties><property name="score" value="0.1080787191816257" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i425]" time="0.012"><properties><property name="score" value="0.04972358239653252" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i426]" time="0.011"><properties><property name="score" value="0.06700244961793111" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i427]" time="0.011"><properties><property name="score" value="0.008829909804647373" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i428]" time="0.012"><properties><property name="score" value="0.11192189952285436" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i429]" time="0.011"><properties><property name="score" value="0.04943726692787087" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i430]" time="0.011"><properties><property name="score" value="0.027296231094725565" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i431]" time="0.012"><properties><property name="score" value="0.05206600786507032" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i432]" time="0.014"><properties><property name="score" value="0.046413194102562505" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i433]" time="0.014"><properties><property name="score" value="0.05741683674167497" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i434]" time="0.013"><properties><property name="score" value="0.07067568123320656" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i435]" time="0.013"><properties><property name="score" value="0.06698730934557307" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i436]" time="0.012"><properties><property name="score" value="0.10595640274855267" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i437]" time="0.011"><properties><property name="score" value="0.026789094089557652" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i438]" time="0.015"><properties><property name="score" value="0.0901132829089546" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i439]" time="0.015"><properties><property name="score" value="0.09566886091128546" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i440]" time="0.013"><properties><property name="score" value="0.07224195965920832" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i441]" time="0.011"><properties><property name="score" value="0.014072607665638337" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i442]" time="0.013"><properties><property name="score" value="0.043505156012002555" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i443]" time="0.012"><properties><property name="score" value="0.06529203640326386" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i444]" time="0.012"><properties><property name="score" value="0.023894324067108563" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i445]" time="0.013"><properties><property name="score" value="0.0312553188875242" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i446]" time="0.011"><properties><property name="score" value="0.03312228333593459" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i447]" time="0.013"><properties><property name="score" value="0.05080463840902438" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i448]" time="0.013"><properties><property name="score" value="0.06660548147823797" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i449]" time="0.011"><properties><property name="score" value="0.05940678777771602" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i450]" time="0.012"><properties><property name="score" value="0.0900620144972053" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i451]" time="0.011"><properties><property name="score" value="0.03708034865761989" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i452]" time="0.011"><properties><property name="score" value="0.028376394576523634" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i453]" time="0.012"><properties><property name="score" value="0.07432923828886162" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i454]" time="0.011"><properties><property name="score" value="0.09759129572556868" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i455]" time="0.015"><properties><property name="score" value="0.055664287511386634" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i456]" time="0.012"><properties><property name="score" value="0.03778091847250198" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i457]" time="0.013"><properties><property name="score" value="0.11749145133359562" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i458]" time="0.012"><properties><property name="score" value="0.08037201501489566" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i459]" time="0.012"><properties><property name="score" value="0.11455034978231637" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i460]" time="0.015"><properties><property name="score" value="0.005799813313311919" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i461]" time="0.011"><properties><property name="score" value="0.06416024137887999" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i462]" time="0.015"><properties><property name="score" value="0.042244615123260276" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i463]" time="0.013"><properties><property name="score" value="0.04149732277944351" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i464]" time="0.012"><properties><property name="score" value="0.10138267131745149" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i465]" time="0.014"><properties><property name="score" value="0.0328682690233173" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i466]" time="0.011"><properties><property name="score" value="0.08717872538749893" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i467]" time="0.013"><properties><property name="score" value="0.07414136914201674" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i468]" time="0.012"><properties><property name="score" value="0.02443732632165241" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i469]" time="0.013"><properties><property name="score" value="0.07510826564660622" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i470]" time="0.012"><properties><property name="score" value="0.09976257254064036" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i471]" time="0.012"><properties><property name="score" value="0.008418054443404266" /></properties><failure message="AssertionError: samples/ieee-chatgpt-generation.jsonl:8601526 (title: The Use of Non-intrusive Load Identification Techn) is an LLM-generated sample, misclassified as human-generated with confidence 0.00841805&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-generation.jsonl:8601526 (title: The Use of Non-intrusive Load Identification Techn) is an LLM-generated sample, misclassified as human-generated with confidence 0.00841805
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i472]" time="0.015"><properties><property name="score" value="0.010839413147956112" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i473]" time="0.012"><properties><property name="score" value="0.00027143238847116" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i474]" time="0.012"><properties><property name="score" value="0.05080242949023206" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i475]" time="0.012"><properties><property name="score" value="0.01990781539873976" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i476]" time="0.011"><properties><property name="score" value="0.04784236290801824" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i477]" time="0.013"><properties><property name="score" value="0.003584411321305714" /></properties><failure message="AssertionError: samples/ieee-chatgpt-generation.jsonl:8601563 (title: BrainSmart: Ambient Assisted Living System Smartph) is an LLM-generated sample, misclassified as human-generated with confidence 0.00358441&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-generation.jsonl:8601563 (title: BrainSmart: Ambient Assisted Living System Smartph) is an LLM-generated sample, misclassified as human-generated with confidence 0.00358441
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i478]" time="0.013"><properties><property name="score" value="0.08772725247254698" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i479]" time="0.013"><properties><property name="score" value="0.11082006232580599" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i480]" time="0.011"><properties><property name="score" value="0.07097952660252149" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i481]" time="0.011"><properties><property name="score" value="0.012395014919336367" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i482]" time="0.012"><properties><property name="score" value="0.08049287256239879" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i483]" time="0.011"><properties><property name="score" value="0.02586496134499572" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i484]" time="0.011"><properties><property name="score" value="0.0862459675358862" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i485]" time="0.012"><properties><property name="score" value="0.073605981705549" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i486]" time="0.011"><properties><property name="score" value="0.04789665824760925" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i487]" time="0.012"><properties><property name="score" value="0.027268958249609954" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i488]" time="0.015"><properties><property name="score" value="0.1000515922445655" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i489]" time="0.012"><properties><property name="score" value="0.057193458843335376" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i490]" time="0.013"><properties><property name="score" value="0.1155782781161574" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i491]" time="0.011"><properties><property name="score" value="0.055769015599138605" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i492]" time="0.012"><properties><property name="score" value="0.09879523500312404" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i493]" time="0.012"><properties><property name="score" value="0.059652220311517956" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i494]" time="0.012"><properties><property name="score" value="0.13946623448169437" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i495]" time="0.015"><properties><property name="score" value="0.045560774225311684" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i496]" time="0.011"><properties><property name="score" value="0.04721195486446228" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i497]" time="0.012"><properties><property name="score" value="0.03966908470877106" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i498]" time="0.011"><properties><property name="score" value="0.030303102984169383" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_generation_jsonl[i499]" time="0.011"><properties><property name="score" value="0.05861740747691857" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i0]" time="0.012"><properties><property name="score" value="0.014121300956104754" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i1]" time="0.011"><properties><property name="score" value="0.05899569309214536" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i2]" time="0.011"><properties><property name="score" value="0.006980609726731091" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i3]" time="0.012"><properties><property name="score" value="0.01270453178102704" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i4]" time="0.011"><properties><property name="score" value="0.002302318201413689" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i5]" time="0.013"><properties><property name="score" value="0.009987670064554033" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i6]" time="0.014"><properties><property name="score" value="0.08339714698463174" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i7]" time="0.012"><properties><property name="score" value="0.010107089391769852" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600031 (title: Energy-Efficient Network Coding Scheme for Two-Way) is an LLM-generated sample, misclassified as human-generated with confidence 0.01010709&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600031 (title: Energy-Efficient Network Coding Scheme for Two-Way) is an LLM-generated sample, misclassified as human-generated with confidence 0.01010709
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i8]" time="0.013"><properties><property name="score" value="0.012463422116754108" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i9]" time="0.012"><properties><property name="score" value="0.03569673031315612" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i10]" time="0.016"><properties><property name="score" value="0.001750090395841175" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600036 (title: Pedestrian Detection and Attribute Analysis Progra) is an LLM-generated sample, misclassified as human-generated with confidence 0.00175009&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600036 (title: Pedestrian Detection and Attribute Analysis Progra) is an LLM-generated sample, misclassified as human-generated with confidence 0.00175009
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i11]" time="0.012"><properties><property name="score" value="0.034581552292289874" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i12]" time="0.013"><properties><property name="score" value="0.03996972536187138" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i13]" time="0.013"><properties><property name="score" value="0.011169070394040359" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i14]" time="0.012"><properties><property name="score" value="0.02352555508459142" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i15]" time="0.014"><properties><property name="score" value="0.08108834685703892" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i16]" time="0.012"><properties><property name="score" value="0.031374053791916046" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i17]" time="0.012"><properties><property name="score" value="0.02905916350675053" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i18]" time="0.012"><properties><property name="score" value="0.023614407510968993" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600058 (title: Demonstration of vCDN Scheme Based on Multi-Access) is an LLM-generated sample, misclassified as human-generated with confidence 0.02361441&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600058 (title: Demonstration of vCDN Scheme Based on Multi-Access) is an LLM-generated sample, misclassified as human-generated with confidence 0.02361441
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i19]" time="0.011"><properties><property name="score" value="0.021553282714659172" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i20]" time="0.012"><properties><property name="score" value="0.0019215623570711937" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i21]" time="0.011"><properties><property name="score" value="0.03219121903469957" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i22]" time="0.013"><properties><property name="score" value="0.0264886464526215" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i23]" time="0.015"><properties><property name="score" value="0.014415932697425804" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600068 (title: Person Tracking and Frontal Face Capture with UAV) is an LLM-generated sample, misclassified as human-generated with confidence 0.01441593&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600068 (title: Person Tracking and Frontal Face Capture with UAV) is an LLM-generated sample, misclassified as human-generated with confidence 0.01441593
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i24]" time="0.025"><properties><property name="score" value="0.024744948231046355" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i25]" time="0.012"><properties><property name="score" value="0.0027990143873779516" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600074 (title: Human Activity Classification in Smartphones using) is an LLM-generated sample, misclassified as human-generated with confidence 0.00279901&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600074 (title: Human Activity Classification in Smartphones using) is an LLM-generated sample, misclassified as human-generated with confidence 0.00279901
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i26]" time="0.013"><properties><property name="score" value="0.01795347335588615" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600079 (title: Noise Suppression Threshold Channel Estimation Met) is an LLM-generated sample, misclassified as human-generated with confidence 0.01795347&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600079 (title: Noise Suppression Threshold Channel Estimation Met) is an LLM-generated sample, misclassified as human-generated with confidence 0.01795347
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i27]" time="0.012"><properties><property name="score" value="0.008947018569205323" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600080 (title: Identification of Coronary Artery Diseased Subject) is an LLM-generated sample, misclassified as human-generated with confidence 0.00894702&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600080 (title: Identification of Coronary Artery Diseased Subject) is an LLM-generated sample, misclassified as human-generated with confidence 0.00894702
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i28]" time="0.014"><properties><property name="score" value="0.02486745632073739" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i29]" time="0.013"><properties><property name="score" value="0.031511310173776286" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i30]" time="0.012"><properties><property name="score" value="0.01442603947161003" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600083 (title: Grouping Subarray for Robust Estimation of Directi) is an LLM-generated sample, misclassified as human-generated with confidence 0.01442604&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600083 (title: Grouping Subarray for Robust Estimation of Directi) is an LLM-generated sample, misclassified as human-generated with confidence 0.01442604
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i31]" time="0.015"><properties><property name="score" value="0.0751829926283587" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i32]" time="0.012"><properties><property name="score" value="0.022778041428761497" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600088 (title: Weighted Nuclear Norm and TV Regularization based ) is an LLM-generated sample, misclassified as human-generated with confidence 0.02277804&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600088 (title: Weighted Nuclear Norm and TV Regularization based ) is an LLM-generated sample, misclassified as human-generated with confidence 0.02277804
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i33]" time="0.015"><properties><property name="score" value="0.034404943842009095" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i34]" time="0.012"><properties><property name="score" value="0.0931427854978295" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i35]" time="0.023"><properties><property name="score" value="0.00032564155232402435" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i36]" time="0.011"><properties><property name="score" value="0.013311197258719654" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600092 (title: Training-Based Joint Antenna and Relay Selection i) is an LLM-generated sample, misclassified as human-generated with confidence 0.0133112&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600092 (title: Training-Based Joint Antenna and Relay Selection i) is an LLM-generated sample, misclassified as human-generated with confidence 0.0133112
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i37]" time="0.012"><properties><property name="score" value="0.044019061735517395" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600094 (title: NLOS Detection and Mitigation for UWB/IMU Fusion S) is an LLM-generated sample, misclassified as human-generated with confidence 0.04401906&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600094 (title: NLOS Detection and Mitigation for UWB/IMU Fusion S) is an LLM-generated sample, misclassified as human-generated with confidence 0.04401906
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i38]" time="0.011"><properties><property name="score" value="0.03841128683146344" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i39]" time="0.011"><properties><property name="score" value="0.013664427404141977" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i40]" time="0.013"><properties><property name="score" value="0.10791753932677395" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i41]" time="0.012"><properties><property name="score" value="0.011348997948229522" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600101 (title: Closed-form Approximations for Coverage and Rate i) is an LLM-generated sample, misclassified as human-generated with confidence 0.011349&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600101 (title: Closed-form Approximations for Coverage and Rate i) is an LLM-generated sample, misclassified as human-generated with confidence 0.011349
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i42]" time="0.012"><properties><property name="score" value="0.03519873093084658" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i43]" time="0.012"><properties><property name="score" value="0.028174500242739153" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i44]" time="0.011"><properties><property name="score" value="0.054370043316581684" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i45]" time="0.014"><properties><property name="score" value="0.03357242733679766" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i46]" time="0.011"><properties><property name="score" value="0.10392414240379955" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i47]" time="0.013"><properties><property name="score" value="0.01932935055736218" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i48]" time="0.013"><properties><property name="score" value="0.015785389495664992" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600116 (title: A New Channel Estimation Algorithm for Fast Linear) is an LLM-generated sample, misclassified as human-generated with confidence 0.01578539&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600116 (title: A New Channel Estimation Algorithm for Fast Linear) is an LLM-generated sample, misclassified as human-generated with confidence 0.01578539
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i49]" time="0.012"><properties><property name="score" value="0.1148144246322591" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i50]" time="0.016"><properties><property name="score" value="0.029472035177841027" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i51]" time="0.012"><properties><property name="score" value="0.03945689802929975" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i52]" time="0.013"><properties><property name="score" value="0.04182518905519461" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i53]" time="0.011"><properties><property name="score" value="0.004604506494676475" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i54]" time="0.013"><properties><property name="score" value="0.02569692900554199" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600132 (title: Spatial Image Steganalysis Based on ResNeXt) is an LLM-generated sample, misclassified as human-generated with confidence 0.02569693&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600132 (title: Spatial Image Steganalysis Based on ResNeXt) is an LLM-generated sample, misclassified as human-generated with confidence 0.02569693
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i55]" time="0.012"><properties><property name="score" value="0.00715934842117627" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i56]" time="0.011"><properties><property name="score" value="0.013812179036509775" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i57]" time="0.012"><properties><property name="score" value="0.008973952721497147" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i58]" time="0.012"><properties><property name="score" value="0.019496740806240753" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600141 (title: Angle-Domain Channel Tracking for High Speed Railw) is an LLM-generated sample, misclassified as human-generated with confidence 0.01949674&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600141 (title: Angle-Domain Channel Tracking for High Speed Railw) is an LLM-generated sample, misclassified as human-generated with confidence 0.01949674
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i59]" time="0.012"><properties><property name="score" value="0.04092477190051902" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i60]" time="0.013"><properties><property name="score" value="0.04596253528833527" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i61]" time="0.012"><properties><property name="score" value="0.004241827602147197" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i62]" time="0.013"><properties><property name="score" value="0.0018068693598771723" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i63]" time="0.011"><properties><property name="score" value="0.012333060898994663" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i64]" time="0.012"><properties><property name="score" value="0.026969165908569126" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i65]" time="0.014"><properties><property name="score" value="0.04756322173276595" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i66]" time="0.012"><properties><property name="score" value="0.039328161449869725" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i67]" time="0.012"><properties><property name="score" value="0.02124452056953552" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i68]" time="0.012"><properties><property name="score" value="0.005258302641764079" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600157 (title: Throat Microphone Speech Enhancement via Progressi) is an LLM-generated sample, misclassified as human-generated with confidence 0.0052583&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600157 (title: Throat Microphone Speech Enhancement via Progressi) is an LLM-generated sample, misclassified as human-generated with confidence 0.0052583
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i69]" time="0.012"><properties><property name="score" value="0.03720981707507898" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i70]" time="0.023"><properties><property name="score" value="0.03557054077841239" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i71]" time="0.013"><properties><property name="score" value="0.014151530993511408" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600169 (title: LayerOS: Scheduling Apps between the Local System ) is an LLM-generated sample, misclassified as human-generated with confidence 0.01415153&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600169 (title: LayerOS: Scheduling Apps between the Local System ) is an LLM-generated sample, misclassified as human-generated with confidence 0.01415153
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i72]" time="0.012"><properties><property name="score" value="0.03077038356621986" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600173 (title: Accurate Specified-Pedestrian Tracking from Unmann) is an LLM-generated sample, misclassified as human-generated with confidence 0.03077038&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600173 (title: Accurate Specified-Pedestrian Tracking from Unmann) is an LLM-generated sample, misclassified as human-generated with confidence 0.03077038
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i73]" time="0.012"><properties><property name="score" value="0.03993052888213722" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i74]" time="0.013"><properties><property name="score" value="0.017441745570340172" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i75]" time="0.012"><properties><property name="score" value="0.04334347791855597" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i76]" time="0.012"><properties><property name="score" value="0.05294795697922683" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i77]" time="0.012"><properties><property name="score" value="0.0004511357176772979" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600195 (title: Dictionary learning based fingerprinting for indoo) is an LLM-generated sample, misclassified as human-generated with confidence 0.00045114&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600195 (title: Dictionary learning based fingerprinting for indoo) is an LLM-generated sample, misclassified as human-generated with confidence 0.00045114
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i78]" time="0.011"><properties><property name="score" value="0.02003070817876984" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i79]" time="0.014"><properties><property name="score" value="0.0006423976317782021" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600198 (title: Cell Search Algorithms at Low SNR for WiMAX System) is an LLM-generated sample, misclassified as human-generated with confidence 0.0006424&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600198 (title: Cell Search Algorithms at Low SNR for WiMAX System) is an LLM-generated sample, misclassified as human-generated with confidence 0.0006424
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i80]" time="0.012"><properties><property name="score" value="0.039485016063506295" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i81]" time="0.012"><properties><property name="score" value="0.025399382759266897" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i82]" time="0.011"><properties><property name="score" value="0.039601230530705855" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i83]" time="0.011"><properties><property name="score" value="0.020715811610583224" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600202 (title: Mridangam Artist Identification from Taniavartanam) is an LLM-generated sample, misclassified as human-generated with confidence 0.02071581&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600202 (title: Mridangam Artist Identification from Taniavartanam) is an LLM-generated sample, misclassified as human-generated with confidence 0.02071581
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i84]" time="0.013"><properties><property name="score" value="0.014405357554175824" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i85]" time="0.012"><properties><property name="score" value="0.015066253971973698" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600206 (title: A 3D Placement of Unmanned Aerial Vehicle Base Sta) is an LLM-generated sample, misclassified as human-generated with confidence 0.01506625&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600206 (title: A 3D Placement of Unmanned Aerial Vehicle Base Sta) is an LLM-generated sample, misclassified as human-generated with confidence 0.01506625
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i86]" time="0.012"><properties><property name="score" value="0.08260345549175874" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i87]" time="0.012"><properties><property name="score" value="0.0652742083471991" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i88]" time="0.012"><properties><property name="score" value="0.02647670574171257" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i89]" time="0.013"><properties><property name="score" value="0.001312237420536544" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i90]" time="0.012"><properties><property name="score" value="0.036670228606855426" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i91]" time="0.013"><properties><property name="score" value="0.007904921529350295" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i92]" time="0.012"><properties><property name="score" value="0.04624633805084266" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i93]" time="0.012"><properties><property name="score" value="0.0157122124856246" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i94]" time="0.013"><properties><property name="score" value="0.017312556933485324" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i95]" time="0.012"><properties><property name="score" value="0.04198376817066696" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i96]" time="0.015"><properties><property name="score" value="0.019201801740831304" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i97]" time="0.013"><properties><property name="score" value="0.046537537018837005" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i98]" time="0.014"><properties><property name="score" value="0.04807053018777707" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i99]" time="0.012"><properties><property name="score" value="0.07821539274274758" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i100]" time="0.012"><properties><property name="score" value="0.016124806023370963" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i101]" time="0.014"><properties><property name="score" value="0.029272482623865503" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i102]" time="0.012"><properties><property name="score" value="0.0020877095073035257" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i103]" time="0.013"><properties><property name="score" value="0.043929926405877895" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i104]" time="0.013"><properties><property name="score" value="0.016377942552392444" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600244 (title: Generalized Statistical Spectrum Occupancy Modelli) is an LLM-generated sample, misclassified as human-generated with confidence 0.01637794&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600244 (title: Generalized Statistical Spectrum Occupancy Modelli) is an LLM-generated sample, misclassified as human-generated with confidence 0.01637794
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i105]" time="0.012"><properties><property name="score" value="0.022132326693213855" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i106]" time="0.015"><properties><property name="score" value="0.0598028488873692" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i107]" time="0.012"><properties><property name="score" value="0.023013507127839228" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600254 (title: Opportunistic Relay Selection for Physical Layer S) is an LLM-generated sample, misclassified as human-generated with confidence 0.02301351&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600254 (title: Opportunistic Relay Selection for Physical Layer S) is an LLM-generated sample, misclassified as human-generated with confidence 0.02301351
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i108]" time="0.013"><properties><property name="score" value="0.008763115661603038" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i109]" time="0.012"><properties><property name="score" value="0.08611331754593277" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i110]" time="0.012"><properties><property name="score" value="0.005501737668028861" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i111]" time="0.013"><properties><property name="score" value="0.024215202541966185" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i112]" time="0.012"><properties><property name="score" value="0.02569692900554199" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600270 (title: Power Allocation of Non-Orthogonal Multiple Access) is an LLM-generated sample, misclassified as human-generated with confidence 0.02569693&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600270 (title: Power Allocation of Non-Orthogonal Multiple Access) is an LLM-generated sample, misclassified as human-generated with confidence 0.02569693
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i113]" time="0.013"><properties><property name="score" value="0.03658612045684406" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i114]" time="0.013"><properties><property name="score" value="0.008825612973811792" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i115]" time="0.013"><properties><property name="score" value="0.012820698311183376" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i116]" time="0.012"><properties><property name="score" value="0.02000086383308064" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i117]" time="0.022"><properties><property name="score" value="0.021093367922555516" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i118]" time="0.022"><properties><property name="score" value="0.07284524988709999" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i119]" time="0.013"><properties><property name="score" value="0.005108456021268859" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600307 (title: Robust Hierarchical Learning for Non-Negative Matr) is an LLM-generated sample, misclassified as human-generated with confidence 0.00510846&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600307 (title: Robust Hierarchical Learning for Non-Negative Matr) is an LLM-generated sample, misclassified as human-generated with confidence 0.00510846
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i120]" time="0.021"><properties><property name="score" value="0.04764062582519042" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i121]" time="0.013"><properties><property name="score" value="0.00848013261360303" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i122]" time="0.011"><properties><property name="score" value="0.08037201501489566" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i123]" time="0.013"><properties><property name="score" value="0.035007867871167386" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i124]" time="0.012"><properties><property name="score" value="0.04318236356740668" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i125]" time="0.012"><properties><property name="score" value="0.05996120409368921" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i126]" time="0.016"><properties><property name="score" value="0.015190783889390547" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i127]" time="0.012"><properties><property name="score" value="0.037021807976256316" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i128]" time="0.012"><properties><property name="score" value="0.07256613456307903" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i129]" time="0.013"><properties><property name="score" value="0.03983709634420851" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600342 (title: Towards High-Efficient Transaction Commitment in a) is an LLM-generated sample, misclassified as human-generated with confidence 0.0398371&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600342 (title: Towards High-Efficient Transaction Commitment in a) is an LLM-generated sample, misclassified as human-generated with confidence 0.0398371
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i130]" time="0.012"><properties><property name="score" value="0.05017087183181168" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i131]" time="0.013"><properties><property name="score" value="0.07163500434715364" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i132]" time="0.012"><properties><property name="score" value="0.00881504233691821" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600358 (title: A Low-Cost High-Speed Neuromorphic Hardware Based ) is an LLM-generated sample, misclassified as human-generated with confidence 0.00881504&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600358 (title: A Low-Cost High-Speed Neuromorphic Hardware Based ) is an LLM-generated sample, misclassified as human-generated with confidence 0.00881504
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i133]" time="0.012"><properties><property name="score" value="0.07210943528705993" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i134]" time="0.012"><properties><property name="score" value="0.007642529055901237" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600361 (title: Communication Network Topology Inference via Trans) is an LLM-generated sample, misclassified as human-generated with confidence 0.00764253&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600361 (title: Communication Network Topology Inference via Trans) is an LLM-generated sample, misclassified as human-generated with confidence 0.00764253
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i135]" time="0.013"><properties><property name="score" value="0.010373184228640309" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600363 (title: A Hierarchical Multiscale Super-Pixel-Based Classi) is an LLM-generated sample, misclassified as human-generated with confidence 0.01037318&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600363 (title: A Hierarchical Multiscale Super-Pixel-Based Classi) is an LLM-generated sample, misclassified as human-generated with confidence 0.01037318
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i136]" time="0.013"><properties><property name="score" value="0.06580227173630937" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i137]" time="0.023"><properties><property name="score" value="0.041135596418747665" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i138]" time="0.013"><properties><property name="score" value="0.02826963484380829" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i139]" time="0.015"><properties><property name="score" value="0.0008338162148180395" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600372 (title: Integration of Wireless Sensor Networks and Smart ) is an LLM-generated sample, misclassified as human-generated with confidence 0.00083382&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600372 (title: Integration of Wireless Sensor Networks and Smart ) is an LLM-generated sample, misclassified as human-generated with confidence 0.00083382
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i140]" time="0.015"><properties><property name="score" value="0.026440928320292656" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i141]" time="0.013"><properties><property name="score" value="0.041724263091291736" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600375 (title: Navion: A 2-mW Fully Integrated Real-Time Visual-I) is an LLM-generated sample, misclassified as human-generated with confidence 0.04172426&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600375 (title: Navion: A 2-mW Fully Integrated Real-Time Visual-I) is an LLM-generated sample, misclassified as human-generated with confidence 0.04172426
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i142]" time="0.024"><properties><property name="score" value="0.0023843832771425037" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i143]" time="0.016"><properties><property name="score" value="0.039504918615584916" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i144]" time="0.012"><properties><property name="score" value="0.021306620752592487" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i145]" time="0.013"><properties><property name="score" value="0.010703401715383976" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600389 (title: Outlier Dirichlet Mixture Mechanism: Adversarial S) is an LLM-generated sample, misclassified as human-generated with confidence 0.0107034&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600389 (title: Outlier Dirichlet Mixture Mechanism: Adversarial S) is an LLM-generated sample, misclassified as human-generated with confidence 0.0107034
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i146]" time="0.012"><properties><property name="score" value="0.005122413098740397" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i147]" time="0.012"><properties><property name="score" value="0.01331565149484959" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600393 (title: Prescribed-Time Observers for Linear Systems in Ob) is an LLM-generated sample, misclassified as human-generated with confidence 0.01331565&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600393 (title: Prescribed-Time Observers for Linear Systems in Ob) is an LLM-generated sample, misclassified as human-generated with confidence 0.01331565
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i148]" time="0.013"><properties><property name="score" value="0.019632485109105646" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i149]" time="0.012"><properties><property name="score" value="0.06164307339703856" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i150]" time="0.013"><properties><property name="score" value="0.0063353268683674635" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600401 (title: HeadNet: An End-to-End Adaptive Relational Network) is an LLM-generated sample, misclassified as human-generated with confidence 0.00633533&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600401 (title: HeadNet: An End-to-End Adaptive Relational Network) is an LLM-generated sample, misclassified as human-generated with confidence 0.00633533
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i151]" time="0.013"><properties><property name="score" value="0.0026923117532551277" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600419 (title: On comparing color spaces for fabric defect classi) is an LLM-generated sample, misclassified as human-generated with confidence 0.00269231&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600419 (title: On comparing color spaces for fabric defect classi) is an LLM-generated sample, misclassified as human-generated with confidence 0.00269231
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i152]" time="0.013"><properties><property name="score" value="0.017832551000573904" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600422 (title: Classification and Clustering of Human Sperm Swimm) is an LLM-generated sample, misclassified as human-generated with confidence 0.01783255&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600422 (title: Classification and Clustering of Human Sperm Swimm) is an LLM-generated sample, misclassified as human-generated with confidence 0.01783255
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i153]" time="0.013"><properties><property name="score" value="0.010762700580546625" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600423 (title: Kinship Verification using Color Features and Extr) is an LLM-generated sample, misclassified as human-generated with confidence 0.0107627&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600423 (title: Kinship Verification using Color Features and Extr) is an LLM-generated sample, misclassified as human-generated with confidence 0.0107627
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i154]" time="0.012"><properties><property name="score" value="0.01771541814695743" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600424 (title: Audiovisual Synchrony Detection with Optimized Aud) is an LLM-generated sample, misclassified as human-generated with confidence 0.01771542&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600424 (title: Audiovisual Synchrony Detection with Optimized Aud) is an LLM-generated sample, misclassified as human-generated with confidence 0.01771542
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i155]" time="0.014"><properties><property name="score" value="0.021578486805368113" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i156]" time="0.012"><properties><property name="score" value="0.030935415356525553" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i157]" time="0.013"><properties><property name="score" value="0.03606528530288866" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i158]" time="0.013"><properties><property name="score" value="0.016088933234054137" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i159]" time="0.012"><properties><property name="score" value="0.003644538174907286" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i160]" time="0.013"><properties><property name="score" value="0.01945229244863511" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600438 (title: The Research of Photoplethysmography Morphology: D) is an LLM-generated sample, misclassified as human-generated with confidence 0.01945229&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600438 (title: The Research of Photoplethysmography Morphology: D) is an LLM-generated sample, misclassified as human-generated with confidence 0.01945229
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i161]" time="0.012"><properties><property name="score" value="0.01618864297279199" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600440 (title: Disguised Facial Recognition Using Neural Networks) is an LLM-generated sample, misclassified as human-generated with confidence 0.01618864&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600440 (title: Disguised Facial Recognition Using Neural Networks) is an LLM-generated sample, misclassified as human-generated with confidence 0.01618864
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i162]" time="0.012"><properties><property name="score" value="0.007967715134160613" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i163]" time="0.012"><properties><property name="score" value="0.021911114703254686" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600444 (title: Interferometric SAR Baseline Estimation by Partiti) is an LLM-generated sample, misclassified as human-generated with confidence 0.02191111&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600444 (title: Interferometric SAR Baseline Estimation by Partiti) is an LLM-generated sample, misclassified as human-generated with confidence 0.02191111
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i164]" time="0.011"><properties><property name="score" value="0.000557537434042743" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600446 (title: A Gradient-Based Adaptive-step Algorithm for Magni) is an LLM-generated sample, misclassified as human-generated with confidence 0.00055754&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600446 (title: A Gradient-Based Adaptive-step Algorithm for Magni) is an LLM-generated sample, misclassified as human-generated with confidence 0.00055754
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i165]" time="0.013"><properties><property name="score" value="0.04602280998191599" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i166]" time="0.012"><properties><property name="score" value="0.015131886791569515" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600451 (title: Using Long Short Term Memory (LSTM) For Radar Cros) is an LLM-generated sample, misclassified as human-generated with confidence 0.01513189&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600451 (title: Using Long Short Term Memory (LSTM) For Radar Cros) is an LLM-generated sample, misclassified as human-generated with confidence 0.01513189
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i167]" time="0.013"><properties><property name="score" value="0.03597552533945447" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i168]" time="0.012"><properties><property name="score" value="0.027209032485886464" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i169]" time="0.012"><properties><property name="score" value="0.008479724403759326" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600460 (title: Vehicle Detection, Tracking and Counting) is an LLM-generated sample, misclassified as human-generated with confidence 0.00847972&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600460 (title: Vehicle Detection, Tracking and Counting) is an LLM-generated sample, misclassified as human-generated with confidence 0.00847972
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i170]" time="0.013"><properties><property name="score" value="0.0347019565469342" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600465 (title: A Novel Body Posture Recognition System on Bed) is an LLM-generated sample, misclassified as human-generated with confidence 0.03470196&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600465 (title: A Novel Body Posture Recognition System on Bed) is an LLM-generated sample, misclassified as human-generated with confidence 0.03470196
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i171]" time="0.011"><properties><property name="score" value="0.007437033809243321" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i172]" time="0.012"><properties><property name="score" value="0.041230902476391273" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i173]" time="0.015"><properties><property name="score" value="0.01650992460949241" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i174]" time="0.012"><properties><property name="score" value="0.00018547695322945756" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600475 (title: A Cascade Calibration of Near-Field Source Gain an) is an LLM-generated sample, misclassified as human-generated with confidence 0.00018548&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600475 (title: A Cascade Calibration of Near-Field Source Gain an) is an LLM-generated sample, misclassified as human-generated with confidence 0.00018548
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i175]" time="0.014"><properties><property name="score" value="0.04186069410597182" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i176]" time="0.012"><properties><property name="score" value="0.028876210125361457" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i177]" time="0.013"><properties><property name="score" value="0.002339154257779974" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i178]" time="0.012"><properties><property name="score" value="0.0326640221533947" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600485 (title: Towards 3D Human Action Recognition Using a Distil) is an LLM-generated sample, misclassified as human-generated with confidence 0.03266402&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600485 (title: Towards 3D Human Action Recognition Using a Distil) is an LLM-generated sample, misclassified as human-generated with confidence 0.03266402
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i179]" time="0.012"><properties><property name="score" value="0.050059240456407084" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i180]" time="0.012"><properties><property name="score" value="0.010985163637272644" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i181]" time="0.011"><properties><property name="score" value="0.002188613622305935" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i182]" time="0.012"><properties><property name="score" value="0.0038101157424585974" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i183]" time="0.013"><properties><property name="score" value="0.06883133881463999" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i184]" time="0.012"><properties><property name="score" value="0.023256991052411635" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600499 (title: Direction-of-Arrival Estimation Using a Sparse Rep) is an LLM-generated sample, misclassified as human-generated with confidence 0.02325699&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600499 (title: Direction-of-Arrival Estimation Using a Sparse Rep) is an LLM-generated sample, misclassified as human-generated with confidence 0.02325699
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i185]" time="0.012"><properties><property name="score" value="0.030895420739524004" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i186]" time="0.012"><properties><property name="score" value="0.00697582213717296" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600502 (title: QPSK Carrier Tracking Method with a Large Pull-in ) is an LLM-generated sample, misclassified as human-generated with confidence 0.00697582&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600502 (title: QPSK Carrier Tracking Method with a Large Pull-in ) is an LLM-generated sample, misclassified as human-generated with confidence 0.00697582
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i187]" time="0.013"><properties><property name="score" value="0.0016582264684039671" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600504 (title: An Object-Based Method Based on a Novel Statistica) is an LLM-generated sample, misclassified as human-generated with confidence 0.00165823&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600504 (title: An Object-Based Method Based on a Novel Statistica) is an LLM-generated sample, misclassified as human-generated with confidence 0.00165823
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i188]" time="0.012"><properties><property name="score" value="0.018007878074571826" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600505 (title: The Event-Driven Power Efficient Wireless Sensor N) is an LLM-generated sample, misclassified as human-generated with confidence 0.01800788&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600505 (title: The Event-Driven Power Efficient Wireless Sensor N) is an LLM-generated sample, misclassified as human-generated with confidence 0.01800788
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i189]" time="0.013"><properties><property name="score" value="0.008346938957048966" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600506 (title: Dynamic Multi-mapping Convolutional Network for Im) is an LLM-generated sample, misclassified as human-generated with confidence 0.00834694&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600506 (title: Dynamic Multi-mapping Convolutional Network for Im) is an LLM-generated sample, misclassified as human-generated with confidence 0.00834694
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i190]" time="0.025"><properties><property name="score" value="0.009494865352769765" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600508 (title: A Raspberry Pi Based Event Driven Quasi Real Time ) is an LLM-generated sample, misclassified as human-generated with confidence 0.00949487&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600508 (title: A Raspberry Pi Based Event Driven Quasi Real Time ) is an LLM-generated sample, misclassified as human-generated with confidence 0.00949487
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i191]" time="0.022"><properties><property name="score" value="0.014208072177606567" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i192]" time="0.012"><properties><property name="score" value="0.0030241629148053395" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600510 (title: Fire Detection in Infrared Video Surveillance Base) is an LLM-generated sample, misclassified as human-generated with confidence 0.00302416&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600510 (title: Fire Detection in Infrared Video Surveillance Base) is an LLM-generated sample, misclassified as human-generated with confidence 0.00302416
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i193]" time="0.016"><properties><property name="score" value="0.02006060869509385" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i194]" time="0.014"><properties><property name="score" value="0.024150383442145262" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i195]" time="0.012"><properties><property name="score" value="0.01899181681133677" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i196]" time="0.011"><properties><property name="score" value="0.0037513878965167446" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i197]" time="0.012"><properties><property name="score" value="0.008106169683325115" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i198]" time="0.012"><properties><property name="score" value="0.004639923384147648" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600520 (title: Spacecraft Detection Based on Deep Convolutional N) is an LLM-generated sample, misclassified as human-generated with confidence 0.00463992&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600520 (title: Spacecraft Detection Based on Deep Convolutional N) is an LLM-generated sample, misclassified as human-generated with confidence 0.00463992
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i199]" time="0.011"><properties><property name="score" value="0.029434579787734894" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600523 (title: Analysis of Depression Based on Facial Cues on A C) is an LLM-generated sample, misclassified as human-generated with confidence 0.02943458&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600523 (title: Analysis of Depression Based on Facial Cues on A C) is an LLM-generated sample, misclassified as human-generated with confidence 0.02943458
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i200]" time="0.012"><properties><property name="score" value="0.022125930225497248" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i201]" time="0.011"><properties><property name="score" value="0.07368647518423699" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i202]" time="0.011"><properties><property name="score" value="0.03738228830633128" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i203]" time="0.012"><properties><property name="score" value="0.02334998687612" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i204]" time="0.012"><properties><property name="score" value="0.09456026932650463" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i205]" time="0.012"><properties><property name="score" value="0.03505155781362013" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i206]" time="0.018"><properties><property name="score" value="0.000877157073075896" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600535 (title: Text Recognition from Silent Lip Movement Video) is an LLM-generated sample, misclassified as human-generated with confidence 0.00087716&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600535 (title: Text Recognition from Silent Lip Movement Video) is an LLM-generated sample, misclassified as human-generated with confidence 0.00087716
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i207]" time="0.031"><properties><property name="score" value="0.040046745740168754" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i208]" time="0.032"><properties><property name="score" value="0.03807453767634136" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i209]" time="0.026"><properties><property name="score" value="0.06316398224775847" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i210]" time="0.015"><properties><property name="score" value="0.03574609986222721" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600543 (title: Real-Time Simulation of Electric Vehicle Battery C) is an LLM-generated sample, misclassified as human-generated with confidence 0.0357461&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600543 (title: Real-Time Simulation of Electric Vehicle Battery C) is an LLM-generated sample, misclassified as human-generated with confidence 0.0357461
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i211]" time="0.014"><properties><property name="score" value="0.029055128070765157" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i212]" time="0.013"><properties><property name="score" value="0.039503324652073735" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i213]" time="0.015"><properties><property name="score" value="0.005674995047877385" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600549 (title: Residential Load Forecasting Using Deep Neural Net) is an LLM-generated sample, misclassified as human-generated with confidence 0.005675&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600549 (title: Residential Load Forecasting Using Deep Neural Net) is an LLM-generated sample, misclassified as human-generated with confidence 0.005675
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i214]" time="0.013"><properties><property name="score" value="0.009126250949631487" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600553 (title: Distribution Test System for Nontechnical Loss Det) is an LLM-generated sample, misclassified as human-generated with confidence 0.00912625&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600553 (title: Distribution Test System for Nontechnical Loss Det) is an LLM-generated sample, misclassified as human-generated with confidence 0.00912625
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i215]" time="0.015"><properties><property name="score" value="0.007449043111368292" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600555 (title: SDP Based Unit Commitment Via Moment Relaxation) is an LLM-generated sample, misclassified as human-generated with confidence 0.00744904&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600555 (title: SDP Based Unit Commitment Via Moment Relaxation) is an LLM-generated sample, misclassified as human-generated with confidence 0.00744904
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i216]" time="0.013"><properties><property name="score" value="0.06117391098207081" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i217]" time="0.014"><properties><property name="score" value="0.05691695414132836" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i218]" time="0.013"><properties><property name="score" value="0.0020989563951190338" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i219]" time="0.013"><properties><property name="score" value="0.05328435805309617" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i220]" time="0.013"><properties><property name="score" value="0.000930913418223378" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600580 (title: Bilevel Programming-Based Unit Commitment for Loca) is an LLM-generated sample, misclassified as human-generated with confidence 0.00093091&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600580 (title: Bilevel Programming-Based Unit Commitment for Loca) is an LLM-generated sample, misclassified as human-generated with confidence 0.00093091
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i221]" time="0.012"><properties><property name="score" value="0.01803754061673324" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600581 (title: Short-Term Forecast for Locational Marginal Pricin) is an LLM-generated sample, misclassified as human-generated with confidence 0.01803754&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600581 (title: Short-Term Forecast for Locational Marginal Pricin) is an LLM-generated sample, misclassified as human-generated with confidence 0.01803754
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i222]" time="0.014"><properties><property name="score" value="0.022234302335066847" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600584 (title: Minimization of Ohmic Losses in Power Networks by ) is an LLM-generated sample, misclassified as human-generated with confidence 0.0222343&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600584 (title: Minimization of Ohmic Losses in Power Networks by ) is an LLM-generated sample, misclassified as human-generated with confidence 0.0222343
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i223]" time="0.013"><properties><property name="score" value="0.007105150990538434" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600594 (title: Application of Machine Learning Algorithm to Forec) is an LLM-generated sample, misclassified as human-generated with confidence 0.00710515&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600594 (title: Application of Machine Learning Algorithm to Forec) is an LLM-generated sample, misclassified as human-generated with confidence 0.00710515
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i224]" time="0.015"><properties><property name="score" value="0.09334894910827951" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i225]" time="0.013"><properties><property name="score" value="0.03431492494990196" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i226]" time="0.016"><properties><property name="score" value="0.009123777608943673" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600597 (title: Synchrophasor Applications as a Service for Power ) is an LLM-generated sample, misclassified as human-generated with confidence 0.00912378&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600597 (title: Synchrophasor Applications as a Service for Power ) is an LLM-generated sample, misclassified as human-generated with confidence 0.00912378
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i227]" time="0.012"><properties><property name="score" value="0.05573136083259489" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i228]" time="0.013"><properties><property name="score" value="0.05684046724395708" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i229]" time="0.013"><properties><property name="score" value="0.02502258932524093" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i230]" time="0.012"><properties><property name="score" value="0.004037689524083454" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i231]" time="0.017"><properties><property name="score" value="0.00041942726879407033" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600635 (title: MILP Model for Reliability Optimization In Active ) is an LLM-generated sample, misclassified as human-generated with confidence 0.00041943&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600635 (title: MILP Model for Reliability Optimization In Active ) is an LLM-generated sample, misclassified as human-generated with confidence 0.00041943
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i232]" time="0.012"><properties><property name="score" value="0.01551756202679" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i233]" time="0.014"><properties><property name="score" value="0.010713719800192756" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i234]" time="0.013"><properties><property name="score" value="0.01894210895231896" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600642 (title: A Comparative Analysis of Two Loosy DC Optimal Pow) is an LLM-generated sample, misclassified as human-generated with confidence 0.01894211&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600642 (title: A Comparative Analysis of Two Loosy DC Optimal Pow) is an LLM-generated sample, misclassified as human-generated with confidence 0.01894211
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i235]" time="0.017"><properties><property name="score" value="0.008319339274487358" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i236]" time="0.013"><properties><property name="score" value="0.053294090073896516" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i237]" time="0.014"><properties><property name="score" value="0.0027264643480906514" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600650 (title: Fast Generator Placement Algorithm using High Perf) is an LLM-generated sample, misclassified as human-generated with confidence 0.00272646&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600650 (title: Fast Generator Placement Algorithm using High Perf) is an LLM-generated sample, misclassified as human-generated with confidence 0.00272646
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i238]" time="0.013"><properties><property name="score" value="0.030125572594430494" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i239]" time="0.013"><properties><property name="score" value="0.06629014541401634" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i240]" time="0.013"><properties><property name="score" value="0.014392930139733862" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i241]" time="0.013"><properties><property name="score" value="0.003406259875637918" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i242]" time="0.025"><properties><property name="score" value="0.12860571714509644" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i243]" time="0.014"><properties><property name="score" value="0.053064774242689294" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i244]" time="0.014"><properties><property name="score" value="0.046672383707163645" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600676 (title: Solar Photovoltaic output prediction using Jackkni) is an LLM-generated sample, misclassified as human-generated with confidence 0.04667238&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600676 (title: Solar Photovoltaic output prediction using Jackkni) is an LLM-generated sample, misclassified as human-generated with confidence 0.04667238
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i245]" time="0.014"><properties><property name="score" value="0.07683396538956111" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i246]" time="0.017"><properties><property name="score" value="0.006657599932913261" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600683 (title: Unscented Transformations for Probabilistic Load F) is an LLM-generated sample, misclassified as human-generated with confidence 0.0066576&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600683 (title: Unscented Transformations for Probabilistic Load F) is an LLM-generated sample, misclassified as human-generated with confidence 0.0066576
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i247]" time="0.013"><properties><property name="score" value="0.011314647973975633" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i248]" time="0.013"><properties><property name="score" value="0.04590518778579411" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i249]" time="0.012"><properties><property name="score" value="0.016334583362281885" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i250]" time="0.027"><properties><property name="score" value="0.0237910650830902" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i251]" time="0.015"><properties><property name="score" value="0.04278516014220024" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i252]" time="0.013"><properties><property name="score" value="0.0370264478614879" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i253]" time="0.013"><properties><property name="score" value="0.012066654682085787" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600733 (title: An Outlier-Insensitive Unmixing Algorithm With Spa) is an LLM-generated sample, misclassified as human-generated with confidence 0.01206665&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600733 (title: An Outlier-Insensitive Unmixing Algorithm With Spa) is an LLM-generated sample, misclassified as human-generated with confidence 0.01206665
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i254]" time="0.015"><properties><property name="score" value="0.044613207227728324" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i255]" time="0.024"><properties><property name="score" value="0.028368555493645378" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i256]" time="0.012"><properties><property name="score" value="0.005218715269283969" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i257]" time="0.023"><properties><property name="score" value="0.17915299210779034" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i258]" time="0.012"><properties><property name="score" value="0.02048607307393313" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i259]" time="0.013"><properties><property name="score" value="0.013005561225087892" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i260]" time="0.014"><properties><property name="score" value="0.02267424148420516" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i261]" time="0.022"><properties><property name="score" value="0.123983943210652" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i262]" time="0.012"><properties><property name="score" value="0.010715008370520973" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i263]" time="0.013"><properties><property name="score" value="0.056387596375728766" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600754 (title: A Castle of Glass: Leaky IoT Appliances in Modern ) is an LLM-generated sample, misclassified as human-generated with confidence 0.0563876&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600754 (title: A Castle of Glass: Leaky IoT Appliances in Modern ) is an LLM-generated sample, misclassified as human-generated with confidence 0.0563876
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i264]" time="0.011"><properties><property name="score" value="0.031077095045300984" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i265]" time="0.011"><properties><property name="score" value="0.02254917599707751" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i266]" time="0.012"><properties><property name="score" value="0.04368930620280431" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i267]" time="0.012"><properties><property name="score" value="0.0163355047372582" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i268]" time="0.012"><properties><property name="score" value="0.04574602468967459" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i269]" time="0.022"><properties><property name="score" value="0.12498962485397412" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i270]" time="0.015"><properties><property name="score" value="0.08852663248571924" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i271]" time="0.012"><properties><property name="score" value="0.0375601481857335" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i272]" time="0.024"><properties><property name="score" value="0.06412600227263011" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i273]" time="0.011"><properties><property name="score" value="0.04434083616133844" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i274]" time="0.012"><properties><property name="score" value="0.1352904894812046" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i275]" time="0.022"><properties><property name="score" value="0.05226297355300616" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i276]" time="0.015"><properties><property name="score" value="0.041296949966007634" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i277]" time="0.012"><properties><property name="score" value="0.022420775881171195" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i278]" time="0.012"><properties><property name="score" value="0.015317580688850585" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i279]" time="0.012"><properties><property name="score" value="0.03829479182069773" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i280]" time="0.012"><properties><property name="score" value="0.03355649644681824" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600819 (title: Small Parts Classification with Flexible Machine V) is an LLM-generated sample, misclassified as human-generated with confidence 0.0335565&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600819 (title: Small Parts Classification with Flexible Machine V) is an LLM-generated sample, misclassified as human-generated with confidence 0.0335565
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i281]" time="0.013"><properties><property name="score" value="0.0009688027121312892" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600824 (title: IoT Device Fingerprint using Deep Learning) is an LLM-generated sample, misclassified as human-generated with confidence 0.0009688&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600824 (title: IoT Device Fingerprint using Deep Learning) is an LLM-generated sample, misclassified as human-generated with confidence 0.0009688
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i282]" time="0.011"><properties><property name="score" value="0.08192599002556089" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i283]" time="0.012"><properties><property name="score" value="0.018206315388902272" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600827 (title: Towards Real-Time Capable Simulations with a Conta) is an LLM-generated sample, misclassified as human-generated with confidence 0.01820632&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600827 (title: Towards Real-Time Capable Simulations with a Conta) is an LLM-generated sample, misclassified as human-generated with confidence 0.01820632
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i284]" time="0.022"><properties><property name="score" value="0.05347083892875073" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i285]" time="0.013"><properties><property name="score" value="0.021124496900132295" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i286]" time="0.012"><properties><property name="score" value="0.0621521263585334" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i287]" time="0.012"><properties><property name="score" value="0.030484708649797243" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600838 (title: NavREn-Rl: Learning to fly in real environment via) is an LLM-generated sample, misclassified as human-generated with confidence 0.03048471&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600838 (title: NavREn-Rl: Learning to fly in real environment via) is an LLM-generated sample, misclassified as human-generated with confidence 0.03048471
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i288]" time="0.012"><properties><property name="score" value="0.0011216808809499934" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i289]" time="0.022"><properties><property name="score" value="0.061527662043214616" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i290]" time="0.013"><properties><property name="score" value="0.006034234373325509" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i291]" time="0.012"><properties><property name="score" value="0.0583079168305245" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i292]" time="0.013"><properties><property name="score" value="0.04341266431500124" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i293]" time="0.012"><properties><property name="score" value="0.008124839478407253" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i294]" time="0.014"><properties><property name="score" value="0.015259894640834437" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600851 (title: Ultra-low-latency Video Coding Method for Autonomo) is an LLM-generated sample, misclassified as human-generated with confidence 0.01525989&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600851 (title: Ultra-low-latency Video Coding Method for Autonomo) is an LLM-generated sample, misclassified as human-generated with confidence 0.01525989
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i295]" time="0.012"><properties><property name="score" value="0.0046983195809091605" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i296]" time="0.012"><properties><property name="score" value="0.04682295261305147" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i297]" time="0.015"><properties><property name="score" value="0.015006970180458934" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i298]" time="0.022"><properties><property name="score" value="0.09352544677893505" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i299]" time="0.023"><properties><property name="score" value="0.030984549791697492" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600857 (title: A Vision Aid for the Visually Impaired using Commo) is an LLM-generated sample, misclassified as human-generated with confidence 0.03098455&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600857 (title: A Vision Aid for the Visually Impaired using Commo) is an LLM-generated sample, misclassified as human-generated with confidence 0.03098455
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i300]" time="0.022"><properties><property name="score" value="0.02792598351990428" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i301]" time="0.022"><properties><property name="score" value="0.0011044639037496027" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i302]" time="0.012"><properties><property name="score" value="0.02884837100733395" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600860 (title: The Industrial IoT for Nusantara) is an LLM-generated sample, misclassified as human-generated with confidence 0.02884837&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600860 (title: The Industrial IoT for Nusantara) is an LLM-generated sample, misclassified as human-generated with confidence 0.02884837
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i303]" time="0.013"><properties><property name="score" value="0.018420518605644087" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i304]" time="0.023"><properties><property name="score" value="0.043424780065487424" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i305]" time="0.013"><properties><property name="score" value="0.002123056462410844" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600864 (title: Unsupervised Video Prediction Network with Spatio-) is an LLM-generated sample, misclassified as human-generated with confidence 0.00212306&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600864 (title: Unsupervised Video Prediction Network with Spatio-) is an LLM-generated sample, misclassified as human-generated with confidence 0.00212306
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i306]" time="0.016"><properties><property name="score" value="0.011460420253472936" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i307]" time="0.012"><properties><property name="score" value="0.005138896680983995" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i308]" time="0.014"><properties><property name="score" value="0.0014730121311312683" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600870 (title: Sentiment Analysis on the Level of Customer Satisf) is an LLM-generated sample, misclassified as human-generated with confidence 0.00147301&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600870 (title: Sentiment Analysis on the Level of Customer Satisf) is an LLM-generated sample, misclassified as human-generated with confidence 0.00147301
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i309]" time="0.011"><properties><property name="score" value="0.0025058971936497976" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i310]" time="0.013"><properties><property name="score" value="0.0017574891965022044" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600875 (title: An OEE Improvement Method Based on TOC) is an LLM-generated sample, misclassified as human-generated with confidence 0.00175749&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600875 (title: An OEE Improvement Method Based on TOC) is an LLM-generated sample, misclassified as human-generated with confidence 0.00175749
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i311]" time="0.012"><properties><property name="score" value="0.030283874108993913" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i312]" time="0.012"><properties><property name="score" value="0.10184987880432761" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i313]" time="0.012"><properties><property name="score" value="0.006989267574980529" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600884 (title: Sentiment Analysis on User Satisfaction Level of M) is an LLM-generated sample, misclassified as human-generated with confidence 0.00698927&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600884 (title: Sentiment Analysis on User Satisfaction Level of M) is an LLM-generated sample, misclassified as human-generated with confidence 0.00698927
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i314]" time="0.012"><properties><property name="score" value="0.007635239581765818" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i315]" time="0.013"><properties><property name="score" value="0.005796571248462001" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i316]" time="0.012"><properties><property name="score" value="0.009138547902226568" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600890 (title: Opportunistic Wardriving Through Neighborhood Publ) is an LLM-generated sample, misclassified as human-generated with confidence 0.00913855&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600890 (title: Opportunistic Wardriving Through Neighborhood Publ) is an LLM-generated sample, misclassified as human-generated with confidence 0.00913855
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i317]" time="0.025"><properties><property name="score" value="0.01017990191900223" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i318]" time="0.012"><properties><property name="score" value="0.027099248753426064" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i319]" time="0.012"><properties><property name="score" value="0.02732458897154455" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i320]" time="0.011"><properties><property name="score" value="0.011547344729900821" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600895 (title: Realization of IPv6 connectivity via RS485 fieldbu) is an LLM-generated sample, misclassified as human-generated with confidence 0.01154734&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600895 (title: Realization of IPv6 connectivity via RS485 fieldbu) is an LLM-generated sample, misclassified as human-generated with confidence 0.01154734
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i321]" time="0.011"><properties><property name="score" value="0.023543706984102686" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i322]" time="0.014"><properties><property name="score" value="0.014430001201182208" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i323]" time="0.011"><properties><property name="score" value="0.02899125877983688" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i324]" time="0.021"><properties><property name="score" value="0.047907129350152136" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i325]" time="0.011"><properties><property name="score" value="0.0031195373792147763" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600906 (title: A preliminary visual system for assistant diagnosi) is an LLM-generated sample, misclassified as human-generated with confidence 0.00311954&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600906 (title: A preliminary visual system for assistant diagnosi) is an LLM-generated sample, misclassified as human-generated with confidence 0.00311954
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i326]" time="0.011"><properties><property name="score" value="0.013914306419460232" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i327]" time="0.011"><properties><property name="score" value="0.004219709304820185" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600910 (title: Knowledge Capturing Platform in Multi-Stage Produc) is an LLM-generated sample, misclassified as human-generated with confidence 0.00421971&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600910 (title: Knowledge Capturing Platform in Multi-Stage Produc) is an LLM-generated sample, misclassified as human-generated with confidence 0.00421971
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i328]" time="0.011"><properties><property name="score" value="0.023426236603435457" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i329]" time="0.013"><properties><property name="score" value="0.006206540472492028" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600914 (title: Naive Bayes Multi-Label Classification Approach fo) is an LLM-generated sample, misclassified as human-generated with confidence 0.00620654&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600914 (title: Naive Bayes Multi-Label Classification Approach fo) is an LLM-generated sample, misclassified as human-generated with confidence 0.00620654
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i330]" time="0.011"><properties><property name="score" value="0.05508811548707304" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i331]" time="0.013"><properties><property name="score" value="0.006276066622018517" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600918 (title: Lie Detector With The Analysis Of The Change Of Di) is an LLM-generated sample, misclassified as human-generated with confidence 0.00627607&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600918 (title: Lie Detector With The Analysis Of The Change Of Di) is an LLM-generated sample, misclassified as human-generated with confidence 0.00627607
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i332]" time="0.013"><properties><property name="score" value="0.03366032327762336" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i333]" time="0.025"><properties><property name="score" value="0.15728463040424268" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i334]" time="0.011"><properties><property name="score" value="0.05810088210626563" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i335]" time="0.011"><properties><property name="score" value="0.029008392855217613" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i336]" time="0.012"><properties><property name="score" value="0.08507217456039662" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i337]" time="0.011"><properties><property name="score" value="0.08442413817816297" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i338]" time="0.011"><properties><property name="score" value="0.0037173137884038976" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i339]" time="0.011"><properties><property name="score" value="0.013935878460569207" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i340]" time="0.011"><properties><property name="score" value="0.062111517593171905" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i341]" time="0.011"><properties><property name="score" value="0.01613319529117363" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i342]" time="0.011"><properties><property name="score" value="0.01267445497598052" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i343]" time="0.012"><properties><property name="score" value="0.005733757119941174" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600937 (title: Delay-based Distribution and Optimization of A Sim) is an LLM-generated sample, misclassified as human-generated with confidence 0.00573376&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600937 (title: Delay-based Distribution and Optimization of A Sim) is an LLM-generated sample, misclassified as human-generated with confidence 0.00573376
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i344]" time="0.012"><properties><property name="score" value="0.09229656752277049" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i345]" time="0.011"><properties><property name="score" value="0.0021382161693117574" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i346]" time="0.022"><properties><property name="score" value="0.036321988319822876" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i347]" time="0.012"><properties><property name="score" value="0.016971952779182287" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600980 (title: Minimization of the High-Level Fault Model for Mic) is an LLM-generated sample, misclassified as human-generated with confidence 0.01697195&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600980 (title: Minimization of the High-Level Fault Model for Mic) is an LLM-generated sample, misclassified as human-generated with confidence 0.01697195
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i348]" time="0.013"><properties><property name="score" value="0.08323360792563128" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i349]" time="0.011"><properties><property name="score" value="8.379762994614737e-06" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i350]" time="0.022"><properties><property name="score" value="0.03244342765904248" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8600998 (title: Adaptive Event Driven Framework for Real Time Mult) is an LLM-generated sample, misclassified as human-generated with confidence 0.03244343&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8600998 (title: Adaptive Event Driven Framework for Real Time Mult) is an LLM-generated sample, misclassified as human-generated with confidence 0.03244343
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i351]" time="0.011"><properties><property name="score" value="0.026692786196486074" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i352]" time="0.022"><properties><property name="score" value="0.1523927487368921" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i353]" time="0.011"><properties><property name="score" value="0.06866953841294099" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i354]" time="0.018"><properties><property name="score" value="0.05267562326313757" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i355]" time="0.011"><properties><property name="score" value="0.03674181825032985" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i356]" time="0.011"><properties><property name="score" value="0.007783087263690813" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i357]" time="0.012"><properties><property name="score" value="0.012372298360269207" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i358]" time="0.011"><properties><property name="score" value="0.002899873938488895" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i359]" time="0.012"><properties><property name="score" value="0.010244589566599238" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8601027 (title: A Comprehensive Monitoring and Controlling System ) is an LLM-generated sample, misclassified as human-generated with confidence 0.01024459&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8601027 (title: A Comprehensive Monitoring and Controlling System ) is an LLM-generated sample, misclassified as human-generated with confidence 0.01024459
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i360]" time="0.012"><properties><property name="score" value="0.08119373429806287" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i361]" time="0.011"><properties><property name="score" value="0.013887514492938458" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i362]" time="0.011"><properties><property name="score" value="0.026966716941007052" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i363]" time="0.011"><properties><property name="score" value="0.02347106631325202" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i364]" time="0.021"><properties><property name="score" value="0.024921258078514245" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i365]" time="0.012"><properties><property name="score" value="0.029433322190108857" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i366]" time="0.011"><properties><property name="score" value="0.051257706587576335" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i367]" time="0.013"><properties><property name="score" value="0.017169112075055004" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i368]" time="0.011"><properties><property name="score" value="0.015589236572238763" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i369]" time="0.011"><properties><property name="score" value="0.008282111496132893" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8601146 (title: Automized, Simulation-Based, Technical and Economi) is an LLM-generated sample, misclassified as human-generated with confidence 0.00828211&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8601146 (title: Automized, Simulation-Based, Technical and Economi) is an LLM-generated sample, misclassified as human-generated with confidence 0.00828211
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i370]" time="0.012"><properties><property name="score" value="0.0027772126572511713" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8601150 (title: Device Classification for NILM using FIT-PS compar) is an LLM-generated sample, misclassified as human-generated with confidence 0.00277721&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8601150 (title: Device Classification for NILM using FIT-PS compar) is an LLM-generated sample, misclassified as human-generated with confidence 0.00277721
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i371]" time="0.011"><properties><property name="score" value="0.02934986654840044" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i372]" time="0.011"><properties><property name="score" value="0.03637153423243222" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i373]" time="0.011"><properties><property name="score" value="0.007614484815332556" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i374]" time="0.011"><properties><property name="score" value="0.04060399892513167" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i375]" time="0.022"><properties><property name="score" value="0.05340275471501929" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i376]" time="0.011"><properties><property name="score" value="0.016571170215562425" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i377]" time="0.022"><properties><property name="score" value="0.09065533050310592" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i378]" time="0.011"><properties><property name="score" value="0.08399842985989392" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i379]" time="0.012"><properties><property name="score" value="0.03383662264971887" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i380]" time="0.011"><properties><property name="score" value="0.016923699244431045" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i381]" time="0.015"><properties><property name="score" value="0.03799982019415116" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8601202 (title: Delivery of an Interactive Audio Course on Fisheri) is an LLM-generated sample, misclassified as human-generated with confidence 0.03799982&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8601202 (title: Delivery of an Interactive Audio Course on Fisheri) is an LLM-generated sample, misclassified as human-generated with confidence 0.03799982
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i382]" time="0.012"><properties><property name="score" value="0.0352885571891115" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8601203 (title: Performance Analysis of Machine Learning Classifie) is an LLM-generated sample, misclassified as human-generated with confidence 0.03528856&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8601203 (title: Performance Analysis of Machine Learning Classifie) is an LLM-generated sample, misclassified as human-generated with confidence 0.03528856
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i383]" time="0.012"><properties><property name="score" value="0.015728349636934258" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i384]" time="0.012"><properties><property name="score" value="0.01538773150899786" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8601218 (title: Detecting and Removing the Impulsive Noise in OFDM) is an LLM-generated sample, misclassified as human-generated with confidence 0.01538773&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8601218 (title: Detecting and Removing the Impulsive Noise in OFDM) is an LLM-generated sample, misclassified as human-generated with confidence 0.01538773
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i385]" time="0.011"><properties><property name="score" value="0.030402404277812822" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i386]" time="0.012"><properties><property name="score" value="0.01616835872300104" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i387]" time="0.022"><properties><property name="score" value="0.048659537636638044" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i388]" time="0.012"><properties><property name="score" value="0.036585325159788074" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i389]" time="0.012"><properties><property name="score" value="0.021598625622060164" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8601232 (title: A Training Utility for Estimating the Bowling Spee) is an LLM-generated sample, misclassified as human-generated with confidence 0.02159863&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8601232 (title: A Training Utility for Estimating the Bowling Spee) is an LLM-generated sample, misclassified as human-generated with confidence 0.02159863
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i390]" time="0.013"><properties><property name="score" value="0.015686383288571193" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i391]" time="0.012"><properties><property name="score" value="0.03900902883686719" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i392]" time="0.022"><properties><property name="score" value="0.02621289917249307" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i393]" time="0.012"><properties><property name="score" value="0.02284175797846455" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i394]" time="0.011"><properties><property name="score" value="0.023552711489455147" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i395]" time="0.013"><properties><property name="score" value="0.0205772923245251" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i396]" time="0.012"><properties><property name="score" value="0.011938047866710821" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8601248 (title: Modeling the Influence of Large-Scale Circulation ) is an LLM-generated sample, misclassified as human-generated with confidence 0.01193805&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8601248 (title: Modeling the Influence of Large-Scale Circulation ) is an LLM-generated sample, misclassified as human-generated with confidence 0.01193805
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i397]" time="0.012"><properties><property name="score" value="0.0035581067796275345" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i398]" time="0.014"><properties><property name="score" value="0.10270494547094944" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i399]" time="0.013"><properties><property name="score" value="0.10469355605683406" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i400]" time="0.012"><properties><property name="score" value="0.017236426464622978" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8601254 (title: Emulating Software Defined Network Using Mininet a) is an LLM-generated sample, misclassified as human-generated with confidence 0.01723643&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8601254 (title: Emulating Software Defined Network Using Mininet a) is an LLM-generated sample, misclassified as human-generated with confidence 0.01723643
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i401]" time="0.013"><properties><property name="score" value="0.027822236812469203" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i402]" time="0.015"><properties><property name="score" value="0.03623229016286311" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i403]" time="0.012"><properties><property name="score" value="0.0700223466158778" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i404]" time="0.021"><properties><property name="score" value="0.00909221208422189" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i405]" time="0.021"><properties><property name="score" value="0.06421393571198615" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i406]" time="0.014"><properties><property name="score" value="0.06755175620440368" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i407]" time="0.011"><properties><property name="score" value="0.026548768805217948" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i408]" time="0.025"><properties><property name="score" value="0.05232105426649939" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i409]" time="0.011"><properties><property name="score" value="0.05091566770023537" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i410]" time="0.013"><properties><property name="score" value="0.010361750271192793" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i411]" time="0.011"><properties><property name="score" value="0.0030553977766634155" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8601277 (title: Asymptotic Analysis of A Fuzzy Based Intrusion Det) is an LLM-generated sample, misclassified as human-generated with confidence 0.0030554&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8601277 (title: Asymptotic Analysis of A Fuzzy Based Intrusion Det) is an LLM-generated sample, misclassified as human-generated with confidence 0.0030554
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i412]" time="0.011"><properties><property name="score" value="0.08535700783859346" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i413]" time="0.012"><properties><property name="score" value="0.0013264554623193536" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i414]" time="0.011"><properties><property name="score" value="0.03268304486174212" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i415]" time="0.023"><properties><property name="score" value="0.10847497798258371" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i416]" time="0.012"><properties><property name="score" value="0.05166272233347313" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i417]" time="0.012"><properties><property name="score" value="0.0715768354221269" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i418]" time="0.011"><properties><property name="score" value="0.007169841902593399" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8601295 (title: WeChat®: The Present and the Future) is an LLM-generated sample, misclassified as human-generated with confidence 0.00716984&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8601295 (title: WeChat®: The Present and the Future) is an LLM-generated sample, misclassified as human-generated with confidence 0.00716984
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i419]" time="0.011"><properties><property name="score" value="8.830227921641232e-05" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i420]" time="0.011"><properties><property name="score" value="0.018392371162367915" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8601302 (title: Overview on Intrusion Detection Schemes for Intern) is an LLM-generated sample, misclassified as human-generated with confidence 0.01839237&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8601302 (title: Overview on Intrusion Detection Schemes for Intern) is an LLM-generated sample, misclassified as human-generated with confidence 0.01839237
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i421]" time="0.011"><properties><property name="score" value="0.026139393182139647" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8601303 (title: A Controllable Deflection Routing and Wavelength A) is an LLM-generated sample, misclassified as human-generated with confidence 0.02613939&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8601303 (title: A Controllable Deflection Routing and Wavelength A) is an LLM-generated sample, misclassified as human-generated with confidence 0.02613939
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i422]" time="0.021"><properties><property name="score" value="0.061164467389425894" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i423]" time="0.011"><properties><property name="score" value="0.04139411846875163" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8601309 (title: One Pixel Attack for Fooling Deep Neural Networks) is an LLM-generated sample, misclassified as human-generated with confidence 0.04139412&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8601309 (title: One Pixel Attack for Fooling Deep Neural Networks) is an LLM-generated sample, misclassified as human-generated with confidence 0.04139412
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i424]" time="0.012"><properties><property name="score" value="0.04875354306000945" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i425]" time="0.011"><properties><property name="score" value="0.09315194471755106" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i426]" time="0.011"><properties><property name="score" value="0.036637714028009416" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i427]" time="0.024"><properties><property name="score" value="0.0073487479576969566" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8601329 (title: Efficient CCA2 Secure Flexible and Publicly-Verifi) is an LLM-generated sample, misclassified as human-generated with confidence 0.00734875&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8601329 (title: Efficient CCA2 Secure Flexible and Publicly-Verifi) is an LLM-generated sample, misclassified as human-generated with confidence 0.00734875
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i428]" time="0.012"><properties><property name="score" value="0.0370012507242945" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i429]" time="0.012"><properties><property name="score" value="0.0019721192787869324" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i430]" time="0.011"><properties><property name="score" value="0.017619674709712374" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i431]" time="0.016"><properties><property name="score" value="0.03518380097311535" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i432]" time="0.016"><properties><property name="score" value="0.02588959464380758" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i433]" time="0.013"><properties><property name="score" value="0.005372592620761507" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i434]" time="0.011"><properties><property name="score" value="0.002511645297054166" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i435]" time="0.013"><properties><property name="score" value="0.020027058644889184" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8601354 (title: Compensation of the Fluctuations of Differential D) is an LLM-generated sample, misclassified as human-generated with confidence 0.02002706&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8601354 (title: Compensation of the Fluctuations of Differential D) is an LLM-generated sample, misclassified as human-generated with confidence 0.02002706
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i436]" time="0.012"><properties><property name="score" value="0.09854506238768934" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i437]" time="0.022"><properties><property name="score" value="0.014627550913879338" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i438]" time="0.011"><properties><property name="score" value="0.07517581222372116" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i439]" time="0.011"><properties><property name="score" value="0.017089009512222253" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8601362 (title: Cofactor-Based Efficient Endmember Extraction for ) is an LLM-generated sample, misclassified as human-generated with confidence 0.01708901&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8601362 (title: Cofactor-Based Efficient Endmember Extraction for ) is an LLM-generated sample, misclassified as human-generated with confidence 0.01708901
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i440]" time="0.023"><properties><property name="score" value="0.09030159858193594" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i441]" time="0.011"><properties><property name="score" value="0.015825588796197776" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i442]" time="0.011"><properties><property name="score" value="0.01797005646949623" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i443]" time="0.011"><properties><property name="score" value="0.05363065749823814" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i444]" time="0.011"><properties><property name="score" value="0.01650992460949241" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i445]" time="0.011"><properties><property name="score" value="0.009223118995249635" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i446]" time="0.011"><properties><property name="score" value="0.0020989563951190338" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i447]" time="0.014"><properties><property name="score" value="0.06580227173630937" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i448]" time="0.011"><properties><property name="score" value="0.00925655921937829" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i449]" time="0.013"><properties><property name="score" value="0.029944947531851396" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i450]" time="0.012"><properties><property name="score" value="0.02201637852285998" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8601423 (title: Behavioral Modeling a Wireless Borehole Communicat) is an LLM-generated sample, misclassified as human-generated with confidence 0.02201638&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8601423 (title: Behavioral Modeling a Wireless Borehole Communicat) is an LLM-generated sample, misclassified as human-generated with confidence 0.02201638
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i451]" time="0.024"><properties><property name="score" value="0.033700285818418385" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i452]" time="0.024"><properties><property name="score" value="0.012067933590320656" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i453]" time="0.012"><properties><property name="score" value="0.018835466307665083" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8601440 (title: Perspective Neural Network Algorithms for Dynamic ) is an LLM-generated sample, misclassified as human-generated with confidence 0.01883547&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8601440 (title: Perspective Neural Network Algorithms for Dynamic ) is an LLM-generated sample, misclassified as human-generated with confidence 0.01883547
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i454]" time="0.012"><properties><property name="score" value="0.030998494788830877" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i455]" time="0.012"><properties><property name="score" value="0.013812179036509775" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i456]" time="0.012"><properties><property name="score" value="0.03216211313017969" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i457]" time="0.011"><properties><property name="score" value="0.009464795572372564" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i458]" time="0.012"><properties><property name="score" value="0.019510294923336113" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i459]" time="0.010"><properties><property name="score" value="0.00416062707349063" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i460]" time="0.011"><properties><property name="score" value="0.014948174476742127" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i461]" time="0.011"><properties><property name="score" value="0.027702424783532242" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i462]" time="0.010"><properties><property name="score" value="0.08700340203493528" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i463]" time="0.011"><properties><property name="score" value="0.028799871392559684" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i464]" time="0.012"><properties><property name="score" value="0.051026397857967254" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i465]" time="0.011"><properties><property name="score" value="0.03276690104132007" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i466]" time="0.012"><properties><property name="score" value="0.05104115573744128" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i467]" time="0.011"><properties><property name="score" value="0.007690350477074226" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i468]" time="0.011"><properties><property name="score" value="0.017753714696672196" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i469]" time="0.011"><properties><property name="score" value="0.08935812480935779" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i470]" time="0.011"><properties><property name="score" value="0.05868378312356426" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i471]" time="0.012"><properties><property name="score" value="0.0693169288469686" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i472]" time="0.012"><properties><property name="score" value="0.037501799359385135" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i473]" time="0.023"><properties><property name="score" value="0.038108109711753844" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i474]" time="0.011"><properties><property name="score" value="0.05855396566782911" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i475]" time="0.024"><properties><property name="score" value="0.05701917848155898" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i476]" time="0.020"><properties><property name="score" value="0.04985118043830256" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i477]" time="0.011"><properties><property name="score" value="0.017441745570340172" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i478]" time="0.011"><properties><property name="score" value="0.05137364034535907" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i479]" time="0.010"><properties><property name="score" value="0.01990781539873976" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i480]" time="0.012"><properties><property name="score" value="0.07767880267978144" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i481]" time="0.010"><properties><property name="score" value="0.045974516374402974" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i482]" time="0.011"><properties><property name="score" value="0.000323702586479957" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i483]" time="0.011"><properties><property name="score" value="0.009206593592991918" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i484]" time="0.011"><properties><property name="score" value="0.02594957317887947" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i485]" time="0.010"><properties><property name="score" value="0.04675731991420351" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i486]" time="0.011"><properties><property name="score" value="0.021941477744225057" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i487]" time="0.010"><properties><property name="score" value="0.0016841206377404827" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8601631 (title: Optimal Configuration of PMU Based on False Data I) is an LLM-generated sample, misclassified as human-generated with confidence 0.00168412&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8601631 (title: Optimal Configuration of PMU Based on False Data I) is an LLM-generated sample, misclassified as human-generated with confidence 0.00168412
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i488]" time="0.012"><properties><property name="score" value="0.017821055390532248" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8601641 (title: Analysis on the Simplified Simulation and Modellin) is an LLM-generated sample, misclassified as human-generated with confidence 0.01782106&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8601641 (title: Analysis on the Simplified Simulation and Modellin) is an LLM-generated sample, misclassified as human-generated with confidence 0.01782106
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i489]" time="0.011"><properties><property name="score" value="0.014345471231064222" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8601654 (title: Equivalent Modeling of Distributed Photovoltaic Po) is an LLM-generated sample, misclassified as human-generated with confidence 0.01434547&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8601654 (title: Equivalent Modeling of Distributed Photovoltaic Po) is an LLM-generated sample, misclassified as human-generated with confidence 0.01434547
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i490]" time="0.011"><properties><property name="score" value="0.028329149696404432" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i491]" time="0.012"><properties><property name="score" value="0.005769969292662713" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i492]" time="0.010"><properties><property name="score" value="0.056320885652527686" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i493]" time="0.010"><properties><property name="score" value="0.049264891152550616" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i494]" time="0.012"><properties><property name="score" value="0.0555445769549745" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i495]" time="0.019"><properties><property name="score" value="0.01028889492950962" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i496]" time="0.023"><properties><property name="score" value="0.01310729729945879" /></properties></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i497]" time="0.011"><properties><property name="score" value="0.0051602373050962935" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8601681 (title: mHealth4Afrika Beta v1 Validation in Rural and Dee) is an LLM-generated sample, misclassified as human-generated with confidence 0.00516024&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8601681 (title: mHealth4Afrika Beta v1 Validation in Rural and Dee) is an LLM-generated sample, misclassified as human-generated with confidence 0.00516024
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i498]" time="0.022"><properties><property name="score" value="0.0032719628664281825" /></properties><failure message="AssertionError: samples/ieee-chatgpt-polish.jsonl:8601701 (title: IoT Sensor Network Approach for Smart Farming: An ) is an LLM-generated sample, misclassified as human-generated with confidence 0.00327196&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/ieee-chatgpt-polish.jsonl:8601701 (title: IoT Sensor Network Approach for Smart Farming: An ) is an LLM-generated sample, misclassified as human-generated with confidence 0.00327196
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_lzma_detect" name="test_cheat_polish_jsonl[i499]" time="0.062"><properties><property name="score" value="0.022001918686892497" /></properties></testcase></testsuite></testsuites>
\ No newline at end of file
diff --git a/roberta-report.xml b/roberta-report.xml
index bdf2680..388735e 100644
--- a/roberta-report.xml
+++ b/roberta-report.xml
@@ -1,394 +1,1225 @@
-<?xml version="1.0" encoding="utf-8"?><testsuites><testsuite name="pytest" errors="0" failures="131" skipped="0" tests="472" time="1697.363" timestamp="2023-05-23T20:05:55.507040" hostname="FALCON-PUNCH"><testcase classname="test_roberta_detect" name="test_training_file" time="60.701"><properties><property name="score" value="0.45471779037924376" /></properties><failure message="AssertionError: The training corpus should always be detected as AI-generated... since it is (score: 0.45471779)&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: The training corpus should always be detected as AI-generated... since it is (score: 0.45471779)
+<?xml version="1.0" encoding="utf-8"?><testsuites><testsuite name="pytest" errors="0" failures="1914" skipped="0" tests="4225" time="5166.116" timestamp="2023-05-31T08:25:31.802765" hostname="FALCON-PUNCH"><testcase classname="test_roberta_detect" name="test_training_file" time="46.767"><properties><property name="score" value="0.19114503905742983" /></properties><failure message="AssertionError: The training corpus should always be detected as AI-generated... since it is (score: 0.19114504)&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: The training corpus should always be detected as AI-generated... since it is (score: 0.19114504)
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_human_samples[IFAW1.txt]" time="3.005"><properties><property name="score" value="0.9623879790306091" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[cdc_covid.txt]" time="5.938"><properties><property name="score" value="0.9858141303062439" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[How_soon-Fans.txt]" time="3.092"><properties><property name="score" value="0.9844722300767899" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[hsus4.txt]" time="3.776"><properties><property name="score" value="0.3929385713168553" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[pg_robinson_crusoe.txt]" time="16.330"><properties><property name="score" value="0.7777995908701861" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[How_soon-Lebron-James.txt]" time="2.151"><properties><property name="score" value="0.4997263699769974" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Ohio_Steel.txt]" time="12.049"><properties><property name="score" value="0.9954873353242875" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[chapter-10.txt]" time="21.524"><properties><property name="score" value="0.9410458881604044" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Env_Prot_Agency-nov1.txt]" time="17.246"><properties><property name="score" value="0.8591362622476393" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[ch5.txt]" time="14.856"><properties><property name="score" value="0.6686606429241322" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[lw1.txt]" time="3.084"><properties><property name="score" value="0.9679946780204773" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Madame_White_Snake.txt]" time="6.481"><properties><property name="score" value="0.9970659494400025" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[110CYL070.txt]" time="0.783"><properties><property name="score" value="0.9972526729106903" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[audubon1.txt]" time="4.111"><properties><property name="score" value="0.9997499329703194" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[littleshelter2.txt]" time="0.983"><properties><property name="score" value="0.9984914064407349" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[guidedogs1.txt]" time="2.108"><properties><property name="score" value="0.884106770157814" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[anth_essay_4.txt]" time="11.039"><properties><property name="score" value="0.8249280764943078" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[113CWL018.txt]" time="0.787"><properties><property name="score" value="0.9133711755275726" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[blog-monastery.txt]" time="3.660"><properties><property name="score" value="0.9993360340595245" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[115CVL037.txt]" time="1.575"><properties><property name="score" value="0.8667178153991699" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[115CVL036.txt]" time="1.142"><properties><property name="score" value="0.9998140931129456" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[A_defense_of_Michael_Moore.txt]" time="16.695"><properties><property name="score" value="0.8974886372171599" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[captured_moments.txt]" time="21.971"><properties><property name="score" value="0.9325130462646485" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[115CVL035.txt]" time="1.570"><properties><property name="score" value="0.3331179618835449" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[NWF1.txt]" time="3.182"><properties><property name="score" value="0.9921465516090393" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[WhereToHongKong.txt]" time="41.698"><properties><property name="score" value="0.9925749024440502" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[att2.txt]" time="1.506"><properties><property name="score" value="0.9954832990964254" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[c4_GOD.txt]" time="0.844"><properties><property name="score" value="0.9995656311511993" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[cdc_covid_asthma.txt]" time="2.055"><properties><property name="score" value="0.47734813392162323" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[fcic_final_report_conclusions.txt]" time="27.779"><properties><property name="score" value="0.9933638129123422" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[118CWL050.txt]" time="2.503"><properties><property name="score" value="0.9997527897357941" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[116CUL033.txt]" time="1.110"><properties><property name="score" value="0.9996890425682068" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[112C-L012.txt]" time="1.044"><properties><property name="score" value="0.954500287771225" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[appalachian1.txt]" time="5.520"><properties><property name="score" value="0.9469614492522346" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[united1.txt]" time="1.342"><properties><property name="score" value="0.9993507464726766" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[117CWL009.txt]" time="0.959"><properties><property name="score" value="0.9864437878131866" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[marine1.txt]" time="3.142"><properties><property name="score" value="0.99975745677948" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Ant_Robot.txt]" time="31.463"><properties><property name="score" value="0.8720748307912246" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[110CYL069.txt]" time="1.942"><properties><property name="score" value="0.9011939018964767" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[CUP2.txt]" time="20.688"><properties><property name="score" value="0.9389931393994225" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Tupelo-Honey-Cafe.txt]" time="3.000"><properties><property name="score" value="0.991249144077301" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[119CWL041.txt]" time="0.845"><properties><property name="score" value="0.9188057780265808" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[lzma_readme.txt]" time="0.967"><properties><property name="score" value="0.9958797097206116" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Homosexuality.txt]" time="10.564"><properties><property name="score" value="0.9907277673482895" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[116CUL032.txt]" time="1.344"><properties><property name="score" value="0.9745008945465088" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[112C-L014.txt]" time="1.087"><properties><property name="score" value="0.9768432378768921" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Fermentation_HR5034.txt]" time="9.703"><properties><property name="score" value="0.8626657327016195" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Acephalous-Internet.txt]" time="2.762"><properties><property name="score" value="0.9897918343544007" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Effing-Idiot.txt]" time="18.687"><properties><property name="score" value="0.9893337627251942" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[114CUL059.txt]" time="1.082"><properties><property name="score" value="0.9975789487361908" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[sucker.txt]" time="2.104"><properties><property name="score" value="0.361471692721049" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[c4_denver.txt]" time="0.513"><properties><property name="score" value="0.9998093247413635" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[blog-jet-lag.txt]" time="3.710"><properties><property name="score" value="0.9812272701944623" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[warner1.txt]" time="1.496"><properties><property name="score" value="0.30424970388412476" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Fermentation_Eminent-Domain.txt]" time="10.246"><properties><property name="score" value="0.9964820529733386" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[113CWL017.txt]" time="1.231"><properties><property name="score" value="0.9997550249099731" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[112C-L015.txt]" time="1.067"><properties><property name="score" value="0.9983396828174591" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[blog-new-year's-resolutions.txt]" time="3.574"><properties><property name="score" value="0.9424317138535636" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[audubon2.txt]" time="2.454"><properties><property name="score" value="0.9635114073753357" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[The_Black_Willow.txt]" time="13.351"><properties><property name="score" value="0.9973260326818987" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Postal_Rate_Comm-ReportToCongress2002WEB.txt]" time="21.242"><properties><property name="score" value="0.9018126036971807" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[alumnifund1.txt]" time="1.098"><properties><property name="score" value="0.9982951581478119" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[112C-L016.txt]" time="0.749"><properties><property name="score" value="0.8658565878868103" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[114CUL057.txt]" time="1.275"><properties><property name="score" value="0.14687404036521912" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[110CYL067.txt]" time="2.081"><properties><property name="score" value="0.9997860789299011" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[hs_keyescrow.txt]" time="10.911"><properties><property name="score" value="0.746738661738003" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[lessig_blog-carbon.txt]" time="6.345"><properties><property name="score" value="0.9617442621125115" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[114CUL060.txt]" time="1.792"><properties><property name="score" value="0.9909478823343912" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[vampires.txt]" time="3.185"><properties><property name="score" value="0.9907969951629638" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[chZ.txt]" time="3.529"><properties><property name="score" value="0.994496742884318" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[ts_theme_intro.txt]" time="0.750"><properties><property name="score" value="0.09102800488471985" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[wildelifewatch1.txt]" time="2.128"><properties><property name="score" value="0.9997594505548477" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[aspca1.txt]" time="2.493"><properties><property name="score" value="0.9997801184654236" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[reuters_santos.txt]" time="0.982"><properties><property name="score" value="0.986989438533783" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[110CYL071.txt]" time="1.494"><properties><property name="score" value="0.998834490776062" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Uprooted_Bike.txt]" time="4.829"><properties><property name="score" value="0.7292217271668571" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[110CYL072.txt]" time="0.547"><properties><property name="score" value="0.9998257756233215" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Black_and_white.txt]" time="12.875"><properties><property name="score" value="0.9739026462330538" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[110CYL200.txt]" time="1.650"><properties><property name="score" value="0.37836694717407227" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[A_Wasted_Day.txt]" time="22.459"><properties><property name="score" value="0.9969487827399681" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[HistoryGreek.txt]" time="10.701"><properties><property name="score" value="0.8857396315125858" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[HistoryLasVegas.txt]" time="11.409"><properties><property name="score" value="0.9985819353776819" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[cnn_santos.txt]" time="1.850"><properties><property name="score" value="0.9937650710344315" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[detroit.txt]" time="11.793"><properties><property name="score" value="0.8865748188074898" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[118CWL049.txt]" time="1.690"><properties><property name="score" value="0.4098638693491618" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[hotel-california.txt]" time="27.519"><properties><property name="score" value="0.9366267288432402" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[about_me.txt]" time="0.573"><properties><property name="score" value="0.9998273849487305" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[602CZL285.txt]" time="1.546"><properties><property name="score" value="0.9996192455291748" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[WhatToHongKong.txt]" time="16.268"><properties><property name="score" value="0.974881879308007" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Bartok.txt]" time="18.024"><properties><property name="score" value="0.9174929582155668" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Uprooted_Farming-on-Sand.txt]" time="4.080"><properties><property name="score" value="0.9851189951101939" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[116CUL034.txt]" time="2.361"><properties><property name="score" value="0.9125429868698121" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[114CUL058.txt]" time="1.268"><properties><property name="score" value="0.9181902805964152" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[IntroDublin.txt]" time="4.217"><properties><property name="score" value="0.9997940829821995" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[defenders5.txt]" time="4.221"><properties><property name="score" value="0.9768698215484619" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Acephalous-Cant-believe.txt]" time="2.216"><properties><property name="score" value="0.9983697235584259" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[118CWL048.txt]" time="0.810"><properties><property name="score" value="0.9684148728847504" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[110CYL068.txt]" time="1.777"><properties><property name="score" value="0.3436531027158101" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[HistoryJerusalem.txt]" time="13.001"><properties><property name="score" value="0.999798309057951" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[rybczynski-ch3.txt]" time="30.650"><properties><property name="score" value="0.9246672485388961" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[117CWL008.txt]" time="0.768"><properties><property name="score" value="0.943294495344162" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Italy.txt]" time="3.213"><properties><property name="score" value="0.9892012973626455" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[CUP1.txt]" time="22.035"><properties><property name="score" value="0.9955122502226579" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Seedbombing.txt]" time="4.359"><properties><property name="score" value="0.9969821998051235" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Nathans_Bylichka.txt]" time="40.558"><properties><property name="score" value="0.5669655745679681" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Anti-Terrorist.txt]" time="3.229"><properties><property name="score" value="0.9985781411329905" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Protocol_Regarding_Access.txt]" time="8.159"><properties><property name="score" value="0.7049760477883475" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[IntroHongKong.txt]" time="3.257"><properties><property name="score" value="0.9996477007865906" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[wwf12.txt]" time="3.390"><properties><property name="score" value="0.6688169002532959" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[112C-L013.txt]" time="1.375"><properties><property name="score" value="0.9484502474466959" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[blog-varsity-athletics.txt]" time="3.167"><properties><property name="score" value="0.9997489353020986" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[AMC2.txt]" time="2.578"><properties><property name="score" value="0.5021124184131622" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.00667_generated.txt]" time="5.942"><properties><property name="score" value="0.9938243882996696" /></properties><failure message="AssertionError: 2111.00667_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99382439&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.00667_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99382439
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_human_samples[IFAW1.txt]" time="3.535"><properties><property name="score" value="0.9623879790306091" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[cdc_covid.txt]" time="7.320"><properties><property name="score" value="0.9858141303062439" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[How_soon-Fans.txt]" time="3.254"><properties><property name="score" value="0.9844722300767899" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[hsus4.txt]" time="5.512"><properties><property name="score" value="0.3929385713168553" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[pg_robinson_crusoe.txt]" time="27.789"><properties><property name="score" value="0.7777995908701861" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[How_soon-Lebron-James.txt]" time="3.517"><properties><property name="score" value="0.4997263699769974" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Ohio_Steel.txt]" time="17.426"><properties><property name="score" value="0.9954873353242875" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[chapter-10.txt]" time="42.449"><properties><property name="score" value="0.9410458881604044" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Env_Prot_Agency-nov1.txt]" time="28.796"><properties><property name="score" value="0.8591362622476393" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[ch5.txt]" time="23.830"><properties><property name="score" value="0.6686606429241322" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[lw1.txt]" time="3.970"><properties><property name="score" value="0.9679946780204773" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Madame_White_Snake.txt]" time="9.430"><properties><property name="score" value="0.9970659494400025" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[110CYL070.txt]" time="1.190"><properties><property name="score" value="0.9972526729106903" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[audubon1.txt]" time="8.090"><properties><property name="score" value="0.9997499329703194" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[littleshelter2.txt]" time="3.590"><properties><property name="score" value="0.9984914064407349" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[guidedogs1.txt]" time="4.367"><properties><property name="score" value="0.884106770157814" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[anth_essay_4.txt]" time="17.863"><properties><property name="score" value="0.8249280764943078" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[113CWL018.txt]" time="1.046"><properties><property name="score" value="0.9133711755275726" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[blog-monastery.txt]" time="5.162"><properties><property name="score" value="0.9993360340595245" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[115CVL037.txt]" time="2.433"><properties><property name="score" value="0.8667178153991699" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[115CVL036.txt]" time="1.934"><properties><property name="score" value="0.9998140931129456" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[A_defense_of_Michael_Moore.txt]" time="26.633"><properties><property name="score" value="0.8974886372171599" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[captured_moments.txt]" time="34.880"><properties><property name="score" value="0.9325130462646485" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[115CVL035.txt]" time="1.993"><properties><property name="score" value="0.3331179618835449" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[NWF1.txt]" time="4.233"><properties><property name="score" value="0.9921465516090393" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[WhereToHongKong.txt]" time="60.874"><properties><property name="score" value="0.9925749024440502" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[att2.txt]" time="2.359"><properties><property name="score" value="0.9954832990964254" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[c4_GOD.txt]" time="1.271"><properties><property name="score" value="0.9995656311511993" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[cdc_covid_asthma.txt]" time="2.848"><properties><property name="score" value="0.47734813392162323" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[fcic_final_report_conclusions.txt]" time="40.703"><properties><property name="score" value="0.9933638129123422" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[118CWL050.txt]" time="3.401"><properties><property name="score" value="0.9997527897357941" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[116CUL033.txt]" time="1.731"><properties><property name="score" value="0.9996890425682068" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[112C-L012.txt]" time="1.512"><properties><property name="score" value="0.954500287771225" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[appalachian1.txt]" time="7.810"><properties><property name="score" value="0.9469614492522346" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[united1.txt]" time="2.268"><properties><property name="score" value="0.9993507464726766" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[117CWL009.txt]" time="1.452"><properties><property name="score" value="0.9864437878131866" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[marine1.txt]" time="4.197"><properties><property name="score" value="0.99975745677948" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Ant_Robot.txt]" time="47.608"><properties><property name="score" value="0.8720748307912246" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[110CYL069.txt]" time="2.952"><properties><property name="score" value="0.9011939018964767" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[CUP2.txt]" time="30.323"><properties><property name="score" value="0.9389931393994225" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Tupelo-Honey-Cafe.txt]" time="4.162"><properties><property name="score" value="0.991249144077301" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[119CWL041.txt]" time="1.357"><properties><property name="score" value="0.9188057780265808" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[lzma_readme.txt]" time="1.575"><properties><property name="score" value="0.9958797097206116" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Homosexuality.txt]" time="18.861"><properties><property name="score" value="0.9907277673482895" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[116CUL032.txt]" time="1.978"><properties><property name="score" value="0.9745008945465088" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[112C-L014.txt]" time="1.263"><properties><property name="score" value="0.9768432378768921" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Fermentation_HR5034.txt]" time="12.775"><properties><property name="score" value="0.8626657327016195" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Acephalous-Internet.txt]" time="4.484"><properties><property name="score" value="0.9897918343544007" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Effing-Idiot.txt]" time="31.096"><properties><property name="score" value="0.9893337627251942" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[114CUL059.txt]" time="1.979"><properties><property name="score" value="0.9975789487361908" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[sucker.txt]" time="3.915"><properties><property name="score" value="0.361471692721049" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[c4_denver.txt]" time="0.780"><properties><property name="score" value="0.9998093247413635" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[blog-jet-lag.txt]" time="5.294"><properties><property name="score" value="0.9812272701944623" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[warner1.txt]" time="2.861"><properties><property name="score" value="0.30424970388412476" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Fermentation_Eminent-Domain.txt]" time="13.763"><properties><property name="score" value="0.9964820529733386" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[113CWL017.txt]" time="1.544"><properties><property name="score" value="0.9997550249099731" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[112C-L015.txt]" time="2.474"><properties><property name="score" value="0.9983396828174591" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[blog-new-year's-resolutions.txt]" time="6.421"><properties><property name="score" value="0.9424317138535636" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[audubon2.txt]" time="3.891"><properties><property name="score" value="0.9635114073753357" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[The_Black_Willow.txt]" time="23.699"><properties><property name="score" value="0.9973260326818987" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Postal_Rate_Comm-ReportToCongress2002WEB.txt]" time="28.905"><properties><property name="score" value="0.9018126036971807" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[alumnifund1.txt]" time="2.199"><properties><property name="score" value="0.9982951581478119" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[112C-L016.txt]" time="1.862"><properties><property name="score" value="0.8658565878868103" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[114CUL057.txt]" time="2.081"><properties><property name="score" value="0.14687404036521912" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[110CYL067.txt]" time="3.058"><properties><property name="score" value="0.9997860789299011" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[hs_keyescrow.txt]" time="18.733"><properties><property name="score" value="0.746738661738003" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[lessig_blog-carbon.txt]" time="9.146"><properties><property name="score" value="0.9617442621125115" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[114CUL060.txt]" time="2.295"><properties><property name="score" value="0.9909478823343912" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[vampires.txt]" time="4.712"><properties><property name="score" value="0.9907969951629638" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[chZ.txt]" time="5.504"><properties><property name="score" value="0.994496742884318" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[ts_theme_intro.txt]" time="1.284"><properties><property name="score" value="0.09102800488471985" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[wildelifewatch1.txt]" time="3.468"><properties><property name="score" value="0.9997594505548477" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[aspca1.txt]" time="3.651"><properties><property name="score" value="0.9997801184654236" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[reuters_santos.txt]" time="1.326"><properties><property name="score" value="0.986989438533783" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[110CYL071.txt]" time="2.211"><properties><property name="score" value="0.998834490776062" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Uprooted_Bike.txt]" time="6.970"><properties><property name="score" value="0.7292217271668571" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[110CYL072.txt]" time="0.868"><properties><property name="score" value="0.9998257756233215" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Black_and_white.txt]" time="17.183"><properties><property name="score" value="0.9739026462330538" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[110CYL200.txt]" time="2.285"><properties><property name="score" value="0.37836694717407227" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[A_Wasted_Day.txt]" time="33.177"><properties><property name="score" value="0.9969487827399681" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[HistoryGreek.txt]" time="22.494"><properties><property name="score" value="0.8857396315125858" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[HistoryLasVegas.txt]" time="20.553"><properties><property name="score" value="0.9985819353776819" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[cnn_santos.txt]" time="3.884"><properties><property name="score" value="0.9937650710344315" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[detroit.txt]" time="18.334"><properties><property name="score" value="0.8865748188074898" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[118CWL049.txt]" time="2.435"><properties><property name="score" value="0.4098638693491618" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[hotel-california.txt]" time="40.351"><properties><property name="score" value="0.9366267288432402" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[about_me.txt]" time="0.845"><properties><property name="score" value="0.9998273849487305" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[602CZL285.txt]" time="1.770"><properties><property name="score" value="0.9996192455291748" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[WhatToHongKong.txt]" time="23.302"><properties><property name="score" value="0.974881879308007" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Bartok.txt]" time="25.092"><properties><property name="score" value="0.9174929582155668" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Uprooted_Farming-on-Sand.txt]" time="6.216"><properties><property name="score" value="0.9851189951101939" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[116CUL034.txt]" time="3.628"><properties><property name="score" value="0.9125429868698121" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[114CUL058.txt]" time="2.074"><properties><property name="score" value="0.9181902805964152" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[IntroDublin.txt]" time="6.701"><properties><property name="score" value="0.9997940829821995" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[defenders5.txt]" time="8.230"><properties><property name="score" value="0.9768698215484619" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Acephalous-Cant-believe.txt]" time="4.936"><properties><property name="score" value="0.9983697235584259" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[118CWL048.txt]" time="1.308"><properties><property name="score" value="0.9684148728847504" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[110CYL068.txt]" time="2.455"><properties><property name="score" value="0.3436531027158101" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[HistoryJerusalem.txt]" time="15.797"><properties><property name="score" value="0.999798309057951" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[rybczynski-ch3.txt]" time="47.052"><properties><property name="score" value="0.9246672485388961" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[117CWL008.txt]" time="1.406"><properties><property name="score" value="0.943294495344162" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Italy.txt]" time="5.381"><properties><property name="score" value="0.9892012973626455" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[CUP1.txt]" time="44.265"><properties><property name="score" value="0.9955122502226579" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Seedbombing.txt]" time="6.518"><properties><property name="score" value="0.9969821998051235" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Nathans_Bylichka.txt]" time="67.208"><properties><property name="score" value="0.5669655745679681" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Anti-Terrorist.txt]" time="6.182"><properties><property name="score" value="0.9985781411329905" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[Protocol_Regarding_Access.txt]" time="12.866"><properties><property name="score" value="0.7049760477883475" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[IntroHongKong.txt]" time="4.979"><properties><property name="score" value="0.9996477007865906" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[wwf12.txt]" time="3.770"><properties><property name="score" value="0.6688169002532959" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[112C-L013.txt]" time="2.009"><properties><property name="score" value="0.9484502474466959" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[blog-varsity-athletics.txt]" time="5.211"><properties><property name="score" value="0.9997489353020986" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_samples[AMC2.txt]" time="5.505"><properties><property name="score" value="0.5021124184131622" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.00667_generated.txt]" time="10.555"><properties><property name="score" value="0.9938243882996696" /></properties><failure message="AssertionError: 2111.00667_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99382439&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.00667_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99382439
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.13472_generated.txt]" time="4.865"><properties><property name="score" value="0.9998128414154053" /></properties><failure message="AssertionError: 2110.13472_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99981284&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.13472_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99981284
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.13472_generated.txt]" time="7.963"><properties><property name="score" value="0.9998128414154053" /></properties><failure message="AssertionError: 2110.13472_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99981284&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.13472_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99981284
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.03612_generated.txt]" time="8.227"><properties><property name="score" value="0.9892661479803232" /></properties><failure message="AssertionError: 2111.03612_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.98926615&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.03612_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.98926615
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.03612_generated.txt]" time="11.656"><properties><property name="score" value="0.9892661479803232" /></properties><failure message="AssertionError: 2111.03612_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.98926615&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.03612_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.98926615
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.00310_generated.txt]" time="4.166"><properties><property name="score" value="0.9975122043064663" /></properties><failure message="AssertionError: 2111.00310_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9975122&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.00310_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9975122
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.00310_generated.txt]" time="5.745"><properties><property name="score" value="0.9975122043064663" /></properties><failure message="AssertionError: 2111.00310_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9975122&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.00310_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9975122
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.06230_generated.txt]" time="7.091"><properties><property name="score" value="0.8617523816915659" /></properties><failure message="AssertionError: 2111.06230_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.86175238&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.06230_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.86175238
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.06230_generated.txt]" time="13.873"><properties><property name="score" value="0.8617523816915659" /></properties><failure message="AssertionError: 2111.06230_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.86175238&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.06230_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.86175238
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.06012_generated.txt]" time="6.777"><properties><property name="score" value="0.9996202046220953" /></properties><failure message="AssertionError: 2111.06012_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9996202&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.06012_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9996202
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.06012_generated.txt]" time="19.817"><properties><property name="score" value="0.9996202046220953" /></properties><failure message="AssertionError: 2111.06012_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9996202&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.06012_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9996202
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.01676_generated.txt]" time="3.896"><properties><property name="score" value="0.985426528113229" /></properties><failure message="AssertionError: 2111.01676_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.98542653&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.01676_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.98542653
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.01676_generated.txt]" time="8.558"><properties><property name="score" value="0.985426528113229" /></properties><failure message="AssertionError: 2111.01676_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.98542653&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.01676_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.98542653
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2112.00405_generated.txt]" time="4.929"><properties><property name="score" value="0.99945011309215" /></properties><failure message="AssertionError: 2112.00405_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99945011&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2112.00405_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99945011
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2112.00405_generated.txt]" time="13.863"><properties><property name="score" value="0.99945011309215" /></properties><failure message="AssertionError: 2112.00405_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99945011&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2112.00405_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99945011
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[gpt_patient_summary.txt]" time="0.333"><properties><property name="score" value="0.9997474551200867" /></properties><failure message="AssertionError: gpt_patient_summary.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99974746&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: gpt_patient_summary.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99974746
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[gpt_patient_summary.txt]" time="0.687"><properties><property name="score" value="0.9997474551200867" /></properties><failure message="AssertionError: gpt_patient_summary.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99974746&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: gpt_patient_summary.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99974746
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[gpt_exercise.txt]" time="0.938"><properties><property name="score" value="0.9049538373947144" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.05241_generated.txt]" time="4.137"><properties><property name="score" value="0.9992387550217765" /></properties><failure message="AssertionError: 2111.05241_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99923876&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.05241_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99923876
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[gpt_exercise.txt]" time="1.877"><properties><property name="score" value="0.9049538373947144" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.05241_generated.txt]" time="8.222"><properties><property name="score" value="0.9992387550217765" /></properties><failure message="AssertionError: 2111.05241_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99923876&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.05241_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99923876
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.12645_generated.txt]" time="3.437"><properties><property name="score" value="0.999110996723175" /></properties><failure message="AssertionError: 2110.12645_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.999111&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.12645_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.999111
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.12645_generated.txt]" time="5.459"><properties><property name="score" value="0.999110996723175" /></properties><failure message="AssertionError: 2110.12645_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.999111&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.12645_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.999111
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.07793_generated.txt]" time="9.251"><properties><property name="score" value="0.9762440621852875" /></properties><failure message="AssertionError: 2111.07793_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.97624406&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.07793_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.97624406
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.07793_generated.txt]" time="13.137"><properties><property name="score" value="0.9762440621852875" /></properties><failure message="AssertionError: 2111.07793_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.97624406&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.07793_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.97624406
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[cleaned_2111.00086_generated.txt]" time="4.944"><properties><property name="score" value="0.9996407702565193" /></properties><failure message="AssertionError: cleaned_2111.00086_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99964077&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: cleaned_2111.00086_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99964077
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[cleaned_2111.00086_generated.txt]" time="7.697"><properties><property name="score" value="0.9996407702565193" /></properties><failure message="AssertionError: cleaned_2111.00086_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99964077&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: cleaned_2111.00086_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99964077
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.10340_generated.txt]" time="4.205"><properties><property name="score" value="0.06167797909842597" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.11115_generated.txt]" time="3.133"><properties><property name="score" value="0.3836793005466461" /></properties><failure message="AssertionError: 2110.11115_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.3836793&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.11115_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.3836793
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.10340_generated.txt]" time="6.923"><properties><property name="score" value="0.06167797909842597" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.11115_generated.txt]" time="4.805"><properties><property name="score" value="0.3836793005466461" /></properties><failure message="AssertionError: 2110.11115_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.3836793&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.11115_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.3836793
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.15093_generated.txt]" time="4.628"><properties><property name="score" value="0.9997433573007584" /></properties><failure message="AssertionError: 2111.15093_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99974336&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.15093_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99974336
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.15093_generated.txt]" time="5.843"><properties><property name="score" value="0.9997433573007584" /></properties><failure message="AssertionError: 2111.15093_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99974336&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.15093_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99974336
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.01515_generated.txt]" time="10.679"><properties><property name="score" value="0.8974505828486549" /></properties><failure message="AssertionError: 2111.01515_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.89745058&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.01515_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.89745058
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.01515_generated.txt]" time="14.225"><properties><property name="score" value="0.8974505828486549" /></properties><failure message="AssertionError: 2111.01515_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.89745058&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.01515_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.89745058
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.12501_generated.txt]" time="4.353"><properties><property name="score" value="0.8262916141086154" /></properties><failure message="AssertionError: 2110.12501_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.82629161&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.12501_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.82629161
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.12501_generated.txt]" time="8.391"><properties><property name="score" value="0.8262916141086154" /></properties><failure message="AssertionError: 2110.12501_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.82629161&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.12501_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.82629161
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.02041_generated.txt]" time="8.233"><properties><property name="score" value="0.5906160632769267" /></properties><failure message="AssertionError: 2111.02041_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.59061606&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.02041_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.59061606
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.02041_generated.txt]" time="17.898"><properties><property name="score" value="0.5906160632769267" /></properties><failure message="AssertionError: 2111.02041_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.59061606&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.02041_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.59061606
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.01023_generated.txt]" time="3.235"><properties><property name="score" value="0.9254761834939321" /></properties><failure message="AssertionError: 2111.01023_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.92547618&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.01023_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.92547618
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.01023_generated.txt]" time="4.678"><properties><property name="score" value="0.9254761834939321" /></properties><failure message="AssertionError: 2111.01023_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.92547618&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.01023_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.92547618
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.03715_generated.txt]" time="3.436"><properties><property name="score" value="0.9991788864135742" /></properties><failure message="AssertionError: 2111.03715_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99917889&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.03715_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99917889
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.03715_generated.txt]" time="6.013"><properties><property name="score" value="0.9991788864135742" /></properties><failure message="AssertionError: 2111.03715_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99917889&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.03715_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99917889
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.02760_generated.txt]" time="5.008"><properties><property name="score" value="0.3858499593204922" /></properties><failure message="AssertionError: 2111.02760_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.38584996&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.02760_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.38584996
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[cgpt_analysis_explosion.txt]" time="3.303"><properties><property name="score" value="0.6463437557220459" /></properties><failure message="AssertionError: cgpt_analysis_explosion.txt is an LLM-generated file, misclassified as human-generated with confidence 0.64634376&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: cgpt_analysis_explosion.txt is an LLM-generated file, misclassified as human-generated with confidence 0.64634376
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.02574_generated.txt]" time="4.811"><properties><property name="score" value="0.9997794255614281" /></properties><failure message="AssertionError: 2111.02574_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99977943&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.02574_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99977943
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.02760_generated.txt]" time="7.630"><properties><property name="score" value="0.3858499593204922" /></properties><failure message="AssertionError: 2111.02760_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.38584996&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.02760_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.38584996
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.15023_generated.txt]" time="6.500"><properties><property name="score" value="0.9958954321013557" /></properties><failure message="AssertionError: 2110.15023_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99589543&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.15023_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99589543
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.02574_generated.txt]" time="8.127"><properties><property name="score" value="0.9997794255614281" /></properties><failure message="AssertionError: 2111.02574_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99977943&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.02574_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99977943
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.03320_generated.txt]" time="9.298"><properties><property name="score" value="0.8900772578576032" /></properties><failure message="AssertionError: 2111.03320_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.89007726&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.03320_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.89007726
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.15023_generated.txt]" time="7.991"><properties><property name="score" value="0.9958954321013557" /></properties><failure message="AssertionError: 2110.15023_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99589543&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.15023_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99589543
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[gpt_patient_narrative.txt]" time="0.769"><properties><property name="score" value="0.8774709403514862" /></properties><failure message="AssertionError: gpt_patient_narrative.txt is an LLM-generated file, misclassified as human-generated with confidence 0.87747094&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: gpt_patient_narrative.txt is an LLM-generated file, misclassified as human-generated with confidence 0.87747094
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.03320_generated.txt]" time="14.274"><properties><property name="score" value="0.8900772578576032" /></properties><failure message="AssertionError: 2111.03320_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.89007726&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.03320_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.89007726
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.15317_generated.txt]" time="4.265"><properties><property name="score" value="0.9997736641338894" /></properties><failure message="AssertionError: 2110.15317_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99977366&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.15317_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99977366
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[gpt_patient_narrative.txt]" time="1.172"><properties><property name="score" value="0.8774709403514862" /></properties><failure message="AssertionError: gpt_patient_narrative.txt is an LLM-generated file, misclassified as human-generated with confidence 0.87747094&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: gpt_patient_narrative.txt is an LLM-generated file, misclassified as human-generated with confidence 0.87747094
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.06181_generated.txt]" time="5.321"><properties><property name="score" value="0.9908169865608215" /></properties><failure message="AssertionError: 2111.06181_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99081699&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.06181_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99081699
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.15317_generated.txt]" time="6.298"><properties><property name="score" value="0.9997736641338894" /></properties><failure message="AssertionError: 2110.15317_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99977366&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.15317_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99977366
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.03294_generated.txt]" time="3.438"><properties><property name="score" value="0.7371706366539001" /></properties><failure message="AssertionError: 2111.03294_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.73717064&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.03294_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.73717064
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.06181_generated.txt]" time="8.047"><properties><property name="score" value="0.9908169865608215" /></properties><failure message="AssertionError: 2111.06181_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99081699&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.06181_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99081699
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.15436_generated.txt]" time="7.586"><properties><property name="score" value="0.9997595995664597" /></properties><failure message="AssertionError: 2111.15436_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9997596&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.15436_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9997596
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.03294_generated.txt]" time="4.798"><properties><property name="score" value="0.7371706366539001" /></properties><failure message="AssertionError: 2111.03294_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.73717064&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.03294_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.73717064
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.00867_generated.txt]" time="5.042"><properties><property name="score" value="0.9912795946002007" /></properties><failure message="AssertionError: 2111.00867_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99127959&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.00867_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99127959
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.15436_generated.txt]" time="12.015"><properties><property name="score" value="0.9997595995664597" /></properties><failure message="AssertionError: 2111.15436_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9997596&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.15436_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9997596
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.05754_generated.txt]" time="4.674"><properties><property name="score" value="0.9741079012552897" /></properties><failure message="AssertionError: 2111.05754_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9741079&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.05754_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9741079
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.00867_generated.txt]" time="6.809"><properties><property name="score" value="0.9912795946002007" /></properties><failure message="AssertionError: 2111.00867_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99127959&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.00867_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99127959
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.12341_generated.txt]" time="3.419"><properties><property name="score" value="0.9998130400975546" /></properties><failure message="AssertionError: 2110.12341_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99981304&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.12341_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99981304
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.05754_generated.txt]" time="8.284"><properties><property name="score" value="0.9741079012552897" /></properties><failure message="AssertionError: 2111.05754_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9741079&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.05754_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9741079
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.00086_generated.txt]" time="5.717"><properties><property name="score" value="0.9976274743676186" /></properties><failure message="AssertionError: 2111.00086_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99762747&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.00086_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99762747
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[dv_t1_aliens.txt]" time="3.022"><properties><property name="score" value="0.20863925218582152" /></properties><failure message="AssertionError: dv_t1_aliens.txt is an LLM-generated file, misclassified as human-generated with confidence 0.20863925&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: dv_t1_aliens.txt is an LLM-generated file, misclassified as human-generated with confidence 0.20863925
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.02326_generated.txt]" time="6.592"><properties><property name="score" value="0.9994895512407477" /></properties><failure message="AssertionError: 2111.02326_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99948955&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.02326_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99948955
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.12341_generated.txt]" time="5.679"><properties><property name="score" value="0.9998130400975546" /></properties><failure message="AssertionError: 2110.12341_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99981304&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.12341_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99981304
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.00514_generated.txt]" time="3.214"><properties><property name="score" value="0.9932700395584106" /></properties><failure message="AssertionError: 2111.00514_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99327004&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.00514_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99327004
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.00086_generated.txt]" time="7.672"><properties><property name="score" value="0.9976274743676186" /></properties><failure message="AssertionError: 2111.00086_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99762747&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.00086_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99762747
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.14532_generated.txt]" time="4.240"><properties><property name="score" value="0.7764513625038995" /></properties><failure message="AssertionError: 2110.14532_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.77645136&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.14532_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.77645136
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.02326_generated.txt]" time="10.664"><properties><property name="score" value="0.9994895512407477" /></properties><failure message="AssertionError: 2111.02326_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99948955&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.02326_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99948955
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.07408_generated.txt]" time="3.155"><properties><property name="score" value="0.9980983833471934" /></properties><failure message="AssertionError: 2111.07408_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99809838&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.07408_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99809838
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.00514_generated.txt]" time="4.909"><properties><property name="score" value="0.9932700395584106" /></properties><failure message="AssertionError: 2111.00514_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99327004&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.00514_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99327004
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.00035_generated.txt]" time="6.054"><properties><property name="score" value="0.9911740964109247" /></properties><failure message="AssertionError: 2111.00035_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9911741&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.00035_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9911741
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.14532_generated.txt]" time="6.520"><properties><property name="score" value="0.7764513625038995" /></properties><failure message="AssertionError: 2110.14532_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.77645136&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.14532_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.77645136
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.02687_generated.txt]" time="3.233"><properties><property name="score" value="0.6927005478313991" /></properties><failure message="AssertionError: 2111.02687_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.69270055&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.02687_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.69270055
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.07408_generated.txt]" time="4.823"><properties><property name="score" value="0.9980983833471934" /></properties><failure message="AssertionError: 2111.07408_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99809838&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.07408_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99809838
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.05204_generated.txt]" time="4.524"><properties><property name="score" value="0.8179490433798896" /></properties><failure message="AssertionError: 2111.05204_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.81794904&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.05204_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.81794904
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.00035_generated.txt]" time="10.402"><properties><property name="score" value="0.9911740964109247" /></properties><failure message="AssertionError: 2111.00035_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9911741&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.00035_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9911741
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.10478_generated.txt]" time="2.928"><properties><property name="score" value="0.9730845212936401" /></properties><failure message="AssertionError: 2110.10478_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.97308452&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.10478_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.97308452
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.02687_generated.txt]" time="5.185"><properties><property name="score" value="0.6927005478313991" /></properties><failure message="AssertionError: 2111.02687_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.69270055&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.02687_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.69270055
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.02188_generated.txt]" time="8.648"><properties><property name="score" value="0.8440222859382629" /></properties><failure message="AssertionError: 2111.02188_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.84402229&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.02188_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.84402229
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.05204_generated.txt]" time="4.352"><properties><property name="score" value="0.8179490433798896" /></properties><failure message="AssertionError: 2111.05204_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.81794904&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.05204_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.81794904
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.10575_generated.txt]" time="3.801"><properties><property name="score" value="0.9796056747436523" /></properties><failure message="AssertionError: 2110.10575_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.97960567&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.10575_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.97960567
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.10478_generated.txt]" time="3.424"><properties><property name="score" value="0.9730845212936401" /></properties><failure message="AssertionError: 2110.10478_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.97308452&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.10478_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.97308452
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.07525_generated.txt]" time="3.933"><properties><property name="score" value="0.7460605725646019" /></properties><failure message="AssertionError: 2111.07525_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.74606057&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.07525_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.74606057
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.02188_generated.txt]" time="8.698"><properties><property name="score" value="0.8440222859382629" /></properties><failure message="AssertionError: 2111.02188_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.84402229&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.02188_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.84402229
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.10577_generated.txt]" time="3.014"><properties><property name="score" value="0.9996822973092397" /></properties><failure message="AssertionError: 2110.10577_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9996823&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.10577_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9996823
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.10575_generated.txt]" time="4.129"><properties><property name="score" value="0.9796056747436523" /></properties><failure message="AssertionError: 2110.10575_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.97960567&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.10575_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.97960567
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.15473_generated.txt]" time="3.109"><properties><property name="score" value="0.993756890296936" /></properties><failure message="AssertionError: 2111.15473_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99375689&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.15473_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99375689
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.07525_generated.txt]" time="3.999"><properties><property name="score" value="0.7460605725646019" /></properties><failure message="AssertionError: 2111.07525_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.74606057&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.07525_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.74606057
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[bing_c_abstract.txt]" time="0.720"><properties><property name="score" value="0.5164026021957397" /></properties><failure message="AssertionError: bing_c_abstract.txt is an LLM-generated file, misclassified as human-generated with confidence 0.5164026&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: bing_c_abstract.txt is an LLM-generated file, misclassified as human-generated with confidence 0.5164026
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.10577_generated.txt]" time="2.941"><properties><property name="score" value="0.9996822973092397" /></properties><failure message="AssertionError: 2110.10577_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9996823&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.10577_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9996823
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.10778_generated.txt]" time="3.382"><properties><property name="score" value="0.9753133739743914" /></properties><failure message="AssertionError: 2110.10778_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.97531337&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.10778_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.97531337
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.15473_generated.txt]" time="3.521"><properties><property name="score" value="0.993756890296936" /></properties><failure message="AssertionError: 2111.15473_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99375689&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.15473_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99375689
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.13900_generated.txt]" time="5.926"><properties><property name="score" value="0.9968726813793183" /></properties><failure message="AssertionError: 2110.13900_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99687268&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.13900_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99687268
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[bing_c_abstract.txt]" time="0.483"><properties><property name="score" value="0.5164026021957397" /></properties><failure message="AssertionError: bing_c_abstract.txt is an LLM-generated file, misclassified as human-generated with confidence 0.5164026&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: bing_c_abstract.txt is an LLM-generated file, misclassified as human-generated with confidence 0.5164026
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.04416_generated.txt]" time="5.713"><properties><property name="score" value="0.8292752265930176" /></properties><failure message="AssertionError: 2111.04416_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.82927523&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.04416_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.82927523
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.10778_generated.txt]" time="4.007"><properties><property name="score" value="0.9753133739743914" /></properties><failure message="AssertionError: 2110.10778_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.97531337&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.10778_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.97531337
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.03913_generated.txt]" time="5.448"><properties><property name="score" value="0.9997698962688446" /></properties><failure message="AssertionError: 2111.03913_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9997699&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.03913_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9997699
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.13900_generated.txt]" time="7.880"><properties><property name="score" value="0.9968726813793183" /></properties><failure message="AssertionError: 2110.13900_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99687268&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.13900_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99687268
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.01706_generated.txt]" time="3.770"><properties><property name="score" value="0.9997688361576625" /></properties><failure message="AssertionError: 2111.01706_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99976884&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.01706_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99976884
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.04416_generated.txt]" time="5.719"><properties><property name="score" value="0.8292752265930176" /></properties><failure message="AssertionError: 2111.04416_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.82927523&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.04416_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.82927523
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.01243_generated.txt]" time="3.272"><properties><property name="score" value="0.9960611561934153" /></properties><failure message="AssertionError: 2111.01243_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99606116&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.01243_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99606116
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.03913_generated.txt]" time="4.633"><properties><property name="score" value="0.9997698962688446" /></properties><failure message="AssertionError: 2111.03913_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9997699&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.03913_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9997699
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.11879_generated.txt]" time="3.323"><properties><property name="score" value="0.1788279414176941" /></properties><failure message="AssertionError: 2110.11879_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.17882794&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.11879_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.17882794
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.01706_generated.txt]" time="3.521"><properties><property name="score" value="0.9997688361576625" /></properties><failure message="AssertionError: 2111.01706_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99976884&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.01706_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99976884
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.10329_generated.txt]" time="4.820"><properties><property name="score" value="0.9983042137963432" /></properties><failure message="AssertionError: 2110.10329_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99830421&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.10329_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99830421
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.01243_generated.txt]" time="3.453"><properties><property name="score" value="0.9960611561934153" /></properties><failure message="AssertionError: 2111.01243_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99606116&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.01243_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99606116
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.04507_generated.txt]" time="4.988"><properties><property name="score" value="0.9944851199785868" /></properties><failure message="AssertionError: 2111.04507_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99448512&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.04507_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99448512
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.11879_generated.txt]" time="3.198"><properties><property name="score" value="0.1788279414176941" /></properties><failure message="AssertionError: 2110.11879_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.17882794&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.11879_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.17882794
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.03837_generated.txt]" time="5.261"><properties><property name="score" value="0.8360573974522677" /></properties><failure message="AssertionError: 2111.03837_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.8360574&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.03837_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.8360574
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.10329_generated.txt]" time="4.507"><properties><property name="score" value="0.9983042137963432" /></properties><failure message="AssertionError: 2110.10329_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99830421&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.10329_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99830421
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.00572_generated.txt]" time="3.462"><properties><property name="score" value="0.9997765251568386" /></properties><failure message="AssertionError: 2111.00572_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99977653&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.00572_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99977653
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.04507_generated.txt]" time="4.719"><properties><property name="score" value="0.9944851199785868" /></properties><failure message="AssertionError: 2111.04507_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99448512&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.04507_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99448512
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.11984_generated.txt]" time="3.994"><properties><property name="score" value="0.9923993178776332" /></properties><failure message="AssertionError: 2110.11984_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99239932&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.11984_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99239932
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.03837_generated.txt]" time="5.021"><properties><property name="score" value="0.8360573974522677" /></properties><failure message="AssertionError: 2111.03837_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.8360574&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.03837_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.8360574
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.06464_generated.txt]" time="4.367"><properties><property name="score" value="0.960891492664814" /></properties><failure message="AssertionError: 2111.06464_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.96089149&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.06464_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.96089149
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.00572_generated.txt]" time="3.331"><properties><property name="score" value="0.9997765251568386" /></properties><failure message="AssertionError: 2111.00572_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99977653&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.00572_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99977653
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[cgpt_hs_essay2.txt]" time="2.363"><properties><property name="score" value="0.9998067617416382" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.02110_generated.txt]" time="3.963"><properties><property name="score" value="0.7906729951500893" /></properties><failure message="AssertionError: 2111.02110_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.790673&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.02110_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.790673
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.11984_generated.txt]" time="3.359"><properties><property name="score" value="0.9923993178776332" /></properties><failure message="AssertionError: 2110.11984_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99239932&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.11984_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99239932
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[cgpt_hs_essay.txt]" time="2.366"><properties><property name="score" value="0.9992981255054474" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.15802_generated.txt]" time="4.151"><properties><property name="score" value="0.9988721534609795" /></properties><failure message="AssertionError: 2110.15802_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99887215&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.15802_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99887215
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.06464_generated.txt]" time="4.087"><properties><property name="score" value="0.960891492664814" /></properties><failure message="AssertionError: 2111.06464_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.96089149&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.06464_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.96089149
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.11589_generated.txt]" time="3.968"><properties><property name="score" value="0.9998159408569336" /></properties><failure message="AssertionError: 2110.11589_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99981594&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.11589_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99981594
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[cgpt_hs_essay2.txt]" time="1.941"><properties><property name="score" value="0.9998067617416382" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.02110_generated.txt]" time="4.978"><properties><property name="score" value="0.7906729951500893" /></properties><failure message="AssertionError: 2111.02110_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.790673&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.02110_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.790673
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.15724_generated.txt]" time="4.763"><properties><property name="score" value="0.9987454017003378" /></properties><failure message="AssertionError: 2110.15724_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9987454&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.15724_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9987454
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[cgpt_hs_essay.txt]" time="2.474"><properties><property name="score" value="0.9992981255054474" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.15802_generated.txt]" time="4.158"><properties><property name="score" value="0.9988721534609795" /></properties><failure message="AssertionError: 2110.15802_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99887215&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.15802_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99887215
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.15130_generated.txt]" time="5.247"><properties><property name="score" value="0.9994887471199035" /></properties><failure message="AssertionError: 2110.15130_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99948875&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.15130_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99948875
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.11589_generated.txt]" time="4.027"><properties><property name="score" value="0.9998159408569336" /></properties><failure message="AssertionError: 2110.11589_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99981594&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.11589_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99981594
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.06644_generated.txt]" time="4.579"><properties><property name="score" value="0.7927331659528944" /></properties><failure message="AssertionError: 2111.06644_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.79273317&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.06644_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.79273317
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.15724_generated.txt]" time="4.555"><properties><property name="score" value="0.9987454017003378" /></properties><failure message="AssertionError: 2110.15724_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9987454&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.15724_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9987454
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[bard_news.txt]" time="1.562"><properties><property name="score" value="0.8677253127098083" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.01322_generated.txt]" time="4.787"><properties><property name="score" value="0.9837439954280853" /></properties><failure message="AssertionError: 2111.01322_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.983744&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.01322_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.983744
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.15130_generated.txt]" time="5.301"><properties><property name="score" value="0.9994887471199035" /></properties><failure message="AssertionError: 2110.15130_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99948875&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.15130_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99948875
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[about_me_short.txt]" time="0.459"><properties><property name="score" value="0.6147785782814026" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[bing_b_song.txt]" time="0.509"><properties><property name="score" value="0.5042322278022766" /></properties><failure message="AssertionError: bing_b_song.txt is an LLM-generated file, misclassified as human-generated with confidence 0.50423223&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: bing_b_song.txt is an LLM-generated file, misclassified as human-generated with confidence 0.50423223
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.06644_generated.txt]" time="4.652"><properties><property name="score" value="0.7927331659528944" /></properties><failure message="AssertionError: 2111.06644_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.79273317&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.06644_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.79273317
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.00526_generated.txt]" time="4.679"><properties><property name="score" value="0.7780005931854248" /></properties><failure message="AssertionError: 2111.00526_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.77800059&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.00526_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.77800059
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[bard_news.txt]" time="1.125"><properties><property name="score" value="0.8677253127098083" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.01322_generated.txt]" time="4.312"><properties><property name="score" value="0.9837439954280853" /></properties><failure message="AssertionError: 2111.01322_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.983744&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.01322_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.983744
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.12010_generated.txt]" time="3.059"><properties><property name="score" value="0.9997117122014364" /></properties><failure message="AssertionError: 2110.12010_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99971171&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.12010_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99971171
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[about_me_short.txt]" time="0.512"><properties><property name="score" value="0.6147785782814026" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[bing_b_song.txt]" time="0.521"><properties><property name="score" value="0.5042322278022766" /></properties><failure message="AssertionError: bing_b_song.txt is an LLM-generated file, misclassified as human-generated with confidence 0.50423223&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: bing_b_song.txt is an LLM-generated file, misclassified as human-generated with confidence 0.50423223
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.13317_generated.txt]" time="6.564"><properties><property name="score" value="0.9980548849472632" /></properties><failure message="AssertionError: 2110.13317_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99805488&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.13317_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99805488
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.00526_generated.txt]" time="5.388"><properties><property name="score" value="0.7780005931854248" /></properties><failure message="AssertionError: 2111.00526_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.77800059&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.00526_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.77800059
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[gpt_santos_long.txt]" time="1.057"><properties><property name="score" value="0.9126072923342387" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[gpt_santos_short.txt]" time="0.480"><properties><property name="score" value="0.9998300075531006" /></properties><failure message="AssertionError: gpt_santos_short.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99983001&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: gpt_santos_short.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99983001
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.12010_generated.txt]" time="3.240"><properties><property name="score" value="0.9997117122014364" /></properties><failure message="AssertionError: 2110.12010_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99971171&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.12010_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99971171
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.02643_generated.txt]" time="5.068"><properties><property name="score" value="0.999600750207901" /></properties><failure message="AssertionError: 2111.02643_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99960075&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.02643_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99960075
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[cgpt_planet_abs.txt]" time="2.159"><properties><property name="score" value="0.8793291449546814" /></properties><failure message="AssertionError: cgpt_planet_abs.txt is an LLM-generated file, misclassified as human-generated with confidence 0.87932914&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: cgpt_planet_abs.txt is an LLM-generated file, misclassified as human-generated with confidence 0.87932914
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[about_me.txt]" time="1.011"><properties><property name="score" value="0.030726104974746704" /></properties><failure message="AssertionError: about_me.txt is an LLM-generated file, misclassified as human-generated with confidence 0.0307261&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: about_me.txt is an LLM-generated file, misclassified as human-generated with confidence 0.0307261
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.13317_generated.txt]" time="8.053"><properties><property name="score" value="0.9980548849472632" /></properties><failure message="AssertionError: 2110.13317_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99805488&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.13317_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99805488
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.15534_generated.txt]" time="3.164"><properties><property name="score" value="0.9991713265577952" /></properties><failure message="AssertionError: 2110.15534_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99917133&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.15534_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99917133
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[gpt_santos_long.txt]" time="1.065"><properties><property name="score" value="0.9126072923342387" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[gpt_santos_short.txt]" time="0.289"><properties><property name="score" value="0.9998300075531006" /></properties><failure message="AssertionError: gpt_santos_short.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99983001&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: gpt_santos_short.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99983001
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.07611_generated.txt]" time="3.045"><properties><property name="score" value="0.9987724125385284" /></properties><failure message="AssertionError: 2111.07611_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99877241&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.07611_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99877241
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.02643_generated.txt]" time="5.713"><properties><property name="score" value="0.999600750207901" /></properties><failure message="AssertionError: 2111.02643_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99960075&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.02643_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99960075
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.13658_generated.txt]" time="6.960"><properties><property name="score" value="0.9993734142997048" /></properties><failure message="AssertionError: 2110.13658_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99937341&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.13658_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99937341
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[about_me.txt]" time="1.007"><properties><property name="score" value="0.030726104974746704" /></properties><failure message="AssertionError: about_me.txt is an LLM-generated file, misclassified as human-generated with confidence 0.0307261&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: about_me.txt is an LLM-generated file, misclassified as human-generated with confidence 0.0307261
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.15725_generated.txt]" time="5.212"><properties><property name="score" value="0.3305565317471822" /></properties><failure message="AssertionError: 2110.15725_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.33055653&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.15725_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.33055653
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.15534_generated.txt]" time="2.914"><properties><property name="score" value="0.9991713265577952" /></properties><failure message="AssertionError: 2110.15534_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99917133&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.15534_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99917133
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.01231_generated.txt]" time="4.990"><properties><property name="score" value="0.9993047383096483" /></properties><failure message="AssertionError: 2111.01231_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99930474&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.01231_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99930474
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.07611_generated.txt]" time="2.818"><properties><property name="score" value="0.9987724125385284" /></properties><failure message="AssertionError: 2111.07611_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99877241&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.07611_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99877241
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.10319_generated.txt]" time="2.727"><properties><property name="score" value="0.7068025370438894" /></properties><failure message="AssertionError: 2110.10319_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.70680254&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.10319_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.70680254
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.13658_generated.txt]" time="5.674"><properties><property name="score" value="0.9993734142997048" /></properties><failure message="AssertionError: 2110.13658_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99937341&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.13658_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99937341
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.01340_generated.txt]" time="4.177"><properties><property name="score" value="0.9997531026601791" /></properties><failure message="AssertionError: 2111.01340_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9997531&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.01340_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9997531
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.15725_generated.txt]" time="4.780"><properties><property name="score" value="0.3305565317471822" /></properties><failure message="AssertionError: 2110.15725_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.33055653&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.15725_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.33055653
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.07699_generated.txt]" time="6.287"><properties><property name="score" value="0.9587874611218771" /></properties><failure message="AssertionError: 2111.07699_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.95878746&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.07699_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.95878746
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.01231_generated.txt]" time="4.562"><properties><property name="score" value="0.9993047383096483" /></properties><failure message="AssertionError: 2111.01231_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99930474&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.01231_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99930474
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.15705_generated.txt]" time="6.402"><properties><property name="score" value="0.9943132617256858" /></properties><failure message="AssertionError: 2110.15705_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99431326&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.15705_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99431326
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.10319_generated.txt]" time="2.525"><properties><property name="score" value="0.7068025370438894" /></properties><failure message="AssertionError: 2110.10319_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.70680254&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.10319_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.70680254
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.02844_generated.txt]" time="3.750"><properties><property name="score" value="0.21748248487710953" /></properties><failure message="AssertionError: 2111.02844_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.21748248&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.02844_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.21748248
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.01340_generated.txt]" time="3.872"><properties><property name="score" value="0.9997531026601791" /></properties><failure message="AssertionError: 2111.01340_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9997531&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.01340_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9997531
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.04130_generated.txt]" time="6.056"><properties><property name="score" value="0.9952034890651703" /></properties><failure message="AssertionError: 2111.04130_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99520349&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.04130_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99520349
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.07699_generated.txt]" time="6.219"><properties><property name="score" value="0.9587874611218771" /></properties><failure message="AssertionError: 2111.07699_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.95878746&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.07699_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.95878746
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.02259_generated.txt]" time="3.746"><properties><property name="score" value="0.9978996600423541" /></properties><failure message="AssertionError: 2111.02259_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99789966&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.02259_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99789966
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.15705_generated.txt]" time="6.468"><properties><property name="score" value="0.9943132617256858" /></properties><failure message="AssertionError: 2110.15705_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99431326&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.15705_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99431326
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.00180_generated.txt]" time="5.064"><properties><property name="score" value="0.7624386012554168" /></properties><failure message="AssertionError: 2111.00180_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.7624386&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.00180_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.7624386
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.02844_generated.txt]" time="3.790"><properties><property name="score" value="0.21748248487710953" /></properties><failure message="AssertionError: 2111.02844_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.21748248&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.02844_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.21748248
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.04574_generated.txt]" time="4.123"><properties><property name="score" value="0.9996647068432399" /></properties><failure message="AssertionError: 2111.04574_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99966471&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.04574_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99966471
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.04130_generated.txt]" time="5.777"><properties><property name="score" value="0.9952034890651703" /></properties><failure message="AssertionError: 2111.04130_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99520349&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.04130_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99520349
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.00808_generated.txt]" time="3.078"><properties><property name="score" value="0.999459832906723" /></properties><failure message="AssertionError: 2111.00808_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99945983&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.00808_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99945983
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.02259_generated.txt]" time="3.402"><properties><property name="score" value="0.9978996600423541" /></properties><failure message="AssertionError: 2111.02259_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99789966&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.02259_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99789966
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.12765_generated.txt]" time="3.646"><properties><property name="score" value="0.9997301399707794" /></properties><failure message="AssertionError: 2110.12765_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99973014&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.12765_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99973014
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.00180_generated.txt]" time="5.161"><properties><property name="score" value="0.7624386012554168" /></properties><failure message="AssertionError: 2111.00180_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.7624386&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.00180_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.7624386
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.13229_generated.txt]" time="2.835"><properties><property name="score" value="0.9883219798405966" /></properties><failure message="AssertionError: 2110.13229_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.98832198&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.13229_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.98832198
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.04574_generated.txt]" time="3.444"><properties><property name="score" value="0.9996647068432399" /></properties><failure message="AssertionError: 2111.04574_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99966471&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.04574_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99966471
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.00554_generated.txt]" time="3.860"><properties><property name="score" value="0.7493788376450539" /></properties><failure message="AssertionError: 2111.00554_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.74937884&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.00554_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.74937884
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.00808_generated.txt]" time="3.033"><properties><property name="score" value="0.999459832906723" /></properties><failure message="AssertionError: 2111.00808_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99945983&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.00808_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99945983
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.11207_generated.txt]" time="2.923"><properties><property name="score" value="0.995137631893158" /></properties><failure message="AssertionError: 2110.11207_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99513763&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.11207_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99513763
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.12765_generated.txt]" time="3.321"><properties><property name="score" value="0.9997301399707794" /></properties><failure message="AssertionError: 2110.12765_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99973014&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.12765_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99973014
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.00607_generated.txt]" time="3.626"><properties><property name="score" value="0.9975566182817731" /></properties><failure message="AssertionError: 2111.00607_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99755662&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.00607_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99755662
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.13229_generated.txt]" time="2.823"><properties><property name="score" value="0.9883219798405966" /></properties><failure message="AssertionError: 2110.13229_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.98832198&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.13229_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.98832198
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.07267_generated.txt]" time="3.883"><properties><property name="score" value="0.9632482975721359" /></properties><failure message="AssertionError: 2111.07267_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9632483&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.07267_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9632483
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.00554_generated.txt]" time="3.501"><properties><property name="score" value="0.7493788376450539" /></properties><failure message="AssertionError: 2111.00554_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.74937884&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.00554_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.74937884
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.11205_generated.txt]" time="5.105"><properties><property name="score" value="0.9983960568904877" /></properties><failure message="AssertionError: 2110.11205_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99839606&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.11205_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99839606
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.11207_generated.txt]" time="2.877"><properties><property name="score" value="0.995137631893158" /></properties><failure message="AssertionError: 2110.11207_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99513763&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.11207_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99513763
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[bard_internet_abstract.txt]" time="1.710"><properties><property name="score" value="0.9997635881106058" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.03945_generated.txt]" time="4.427"><properties><property name="score" value="0.9988931342959404" /></properties><failure message="AssertionError: 2111.03945_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99889313&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.03945_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99889313
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.00607_generated.txt]" time="3.355"><properties><property name="score" value="0.9975566182817731" /></properties><failure message="AssertionError: 2111.00607_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99755662&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.00607_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99755662
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.12383_generated.txt]" time="2.929"><properties><property name="score" value="0.999808649222056" /></properties><failure message="AssertionError: 2110.12383_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99980865&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.12383_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99980865
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.07267_generated.txt]" time="3.591"><properties><property name="score" value="0.9632482975721359" /></properties><failure message="AssertionError: 2111.07267_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9632483&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.07267_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.9632483
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.02362_generated.txt]" time="6.064"><properties><property name="score" value="0.5322309624065052" /></properties><failure message="AssertionError: 2111.02362_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.53223096&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.02362_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.53223096
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.11205_generated.txt]" time="4.543"><properties><property name="score" value="0.9983960568904877" /></properties><failure message="AssertionError: 2110.11205_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99839606&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.11205_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99839606
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.15799_generated.txt]" time="5.190"><properties><property name="score" value="0.9998103777567545" /></properties><failure message="AssertionError: 2110.15799_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99981038&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.15799_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99981038
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[bard_internet_abstract.txt]" time="1.235"><properties><property name="score" value="0.9997635881106058" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.03945_generated.txt]" time="4.111"><properties><property name="score" value="0.9988931342959404" /></properties><failure message="AssertionError: 2111.03945_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99889313&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.03945_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99889313
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[bard_abstract.txt]" time="0.571"><properties><property name="score" value="0.998390257358551" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i0]" time="0.345"><properties><property name="score" value="0.5571487545967102" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i1]" time="1.857"><properties><property name="score" value="0.5741835981607437" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i2]" time="0.425"><properties><property name="score" value="0.9997851252555847" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i3]" time="1.414"><properties><property name="score" value="0.3033061623573303" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255003 (len: 490) is a human-generated sample, misclassified as AI-generated with confidence 0.30330616&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255003 (len: 490) is a human-generated sample, misclassified as AI-generated with confidence 0.30330616
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.12383_generated.txt]" time="3.031"><properties><property name="score" value="0.999808649222056" /></properties><failure message="AssertionError: 2110.12383_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99980865&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.12383_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99980865
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2111.02362_generated.txt]" time="5.329"><properties><property name="score" value="0.5322309624065052" /></properties><failure message="AssertionError: 2111.02362_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.53223096&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2111.02362_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.53223096
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[2110.15799_generated.txt]" time="4.632"><properties><property name="score" value="0.9998103777567545" /></properties><failure message="AssertionError: 2110.15799_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99981038&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: 2110.15799_generated.txt is an LLM-generated file, misclassified as human-generated with confidence 0.99981038
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_sample[bard_abstract.txt]" time="0.492"><properties><property name="score" value="0.998390257358551" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i0]" time="0.328"><properties><property name="score" value="0.5571487545967102" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i1]" time="1.799"><properties><property name="score" value="0.5741835981607437" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i2]" time="0.524"><properties><property name="score" value="0.9997851252555847" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i3]" time="1.197"><properties><property name="score" value="0.3033061623573303" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255003 (len: 490) is a human-generated sample, misclassified as AI-generated with confidence 0.30330616&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255003 (len: 490) is a human-generated sample, misclassified as AI-generated with confidence 0.30330616
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i4]" time="0.548"><properties><property name="score" value="0.9997872710227966" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i5]" time="0.564"><properties><property name="score" value="0.9581142067909241" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i6]" time="1.010"><properties><property name="score" value="0.00045305490493774414" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i7]" time="0.473"><properties><property name="score" value="0.9992738366127014" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i8]" time="0.598"><properties><property name="score" value="0.9997966885566711" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i9]" time="0.207"><properties><property name="score" value="0.7315866351127625" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255009 (len: 55) is a human-generated sample, misclassified as AI-generated with confidence 0.73158664&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255009 (len: 55) is a human-generated sample, misclassified as AI-generated with confidence 0.73158664
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i4]" time="0.475"><properties><property name="score" value="0.9997872710227966" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i5]" time="0.420"><properties><property name="score" value="0.9581142067909241" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i6]" time="0.885"><properties><property name="score" value="0.00045305490493774414" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i7]" time="0.464"><properties><property name="score" value="0.9992738366127014" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i8]" time="0.630"><properties><property name="score" value="0.9997966885566711" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i9]" time="0.192"><properties><property name="score" value="0.7315866351127625" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255009 (len: 55) is a human-generated sample, misclassified as AI-generated with confidence 0.73158664&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255009 (len: 55) is a human-generated sample, misclassified as AI-generated with confidence 0.73158664
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i10]" time="0.348"><properties><property name="score" value="0.8875606060028076" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i11]" time="2.178"><properties><property name="score" value="0.99895378947258" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i12]" time="1.008"><properties><property name="score" value="0.71474888920784" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255012 (len: 359) is a human-generated sample, misclassified as AI-generated with confidence 0.71474889&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255012 (len: 359) is a human-generated sample, misclassified as AI-generated with confidence 0.71474889
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i10]" time="0.271"><properties><property name="score" value="0.8875606060028076" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i11]" time="2.013"><properties><property name="score" value="0.99895378947258" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i12]" time="0.805"><properties><property name="score" value="0.71474888920784" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255012 (len: 359) is a human-generated sample, misclassified as AI-generated with confidence 0.71474889&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255012 (len: 359) is a human-generated sample, misclassified as AI-generated with confidence 0.71474889
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i13]" time="2.415"><properties><property name="score" value="0.8826677054166794" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i14]" time="0.242"><properties><property name="score" value="0.9525750279426575" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255014 (len: 83) is a human-generated sample, misclassified as AI-generated with confidence 0.95257503&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255014 (len: 83) is a human-generated sample, misclassified as AI-generated with confidence 0.95257503
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i13]" time="2.185"><properties><property name="score" value="0.8826677054166794" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i14]" time="0.230"><properties><property name="score" value="0.9525750279426575" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255014 (len: 83) is a human-generated sample, misclassified as AI-generated with confidence 0.95257503&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255014 (len: 83) is a human-generated sample, misclassified as AI-generated with confidence 0.95257503
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i15]" time="2.177"><properties><property name="score" value="0.05446718633174896" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i16]" time="0.302"><properties><property name="score" value="0.997799813747406" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i17]" time="0.275"><properties><property name="score" value="0.8389936685562134" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i18]" time="1.081"><properties><property name="score" value="0.1324148178100586" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i19]" time="1.504"><properties><property name="score" value="0.48974796136220294" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i20]" time="1.827"><properties><property name="score" value="0.9978846907615662" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i21]" time="2.862"><properties><property name="score" value="0.6673129439353943" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i22]" time="2.618"><properties><property name="score" value="0.9785245895385742" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i23]" time="0.846"><properties><property name="score" value="0.9664595425128937" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i24]" time="0.250"><properties><property name="score" value="0.9607297778129578" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i25]" time="2.307"><properties><property name="score" value="0.8946043372154235" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i26]" time="2.708"><properties><property name="score" value="0.4704836706320445" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i27]" time="1.772"><properties><property name="score" value="0.9968965649604797" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i28]" time="1.366"><properties><property name="score" value="0.9996445576349894" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i29]" time="1.026"><properties><property name="score" value="0.999719113111496" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i30]" time="1.843"><properties><property name="score" value="0.8905469328165054" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i31]" time="0.208"><properties><property name="score" value="0.8489285707473755" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255031 (len: 60) is a human-generated sample, misclassified as AI-generated with confidence 0.84892857&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255031 (len: 60) is a human-generated sample, misclassified as AI-generated with confidence 0.84892857
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i15]" time="2.097"><properties><property name="score" value="0.05446718633174896" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i16]" time="0.321"><properties><property name="score" value="0.997799813747406" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i17]" time="0.206"><properties><property name="score" value="0.8389936685562134" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i18]" time="0.910"><properties><property name="score" value="0.1324148178100586" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i19]" time="1.413"><properties><property name="score" value="0.48974796136220294" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i20]" time="1.693"><properties><property name="score" value="0.9978846907615662" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i21]" time="2.474"><properties><property name="score" value="0.6673129439353943" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i22]" time="2.313"><properties><property name="score" value="0.9785245895385742" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i23]" time="0.576"><properties><property name="score" value="0.9664595425128937" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i24]" time="0.225"><properties><property name="score" value="0.9607297778129578" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i25]" time="2.228"><properties><property name="score" value="0.8946043372154235" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i26]" time="2.436"><properties><property name="score" value="0.4704836706320445" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i27]" time="2.423"><properties><property name="score" value="0.9968965649604797" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i28]" time="1.598"><properties><property name="score" value="0.9996445576349894" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i29]" time="1.089"><properties><property name="score" value="0.999719113111496" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i30]" time="1.924"><properties><property name="score" value="0.8905469328165054" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i31]" time="0.187"><properties><property name="score" value="0.8489285707473755" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255031 (len: 60) is a human-generated sample, misclassified as AI-generated with confidence 0.84892857&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255031 (len: 60) is a human-generated sample, misclassified as AI-generated with confidence 0.84892857
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i32]" time="0.281"><properties><property name="score" value="0.9778472185134888" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i33]" time="2.483"><properties><property name="score" value="0.5959689855575562" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i34]" time="0.878"><properties><property name="score" value="0.9874178767204285" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i35]" time="2.781"><properties><property name="score" value="0.9883281707763671" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i36]" time="1.491"><properties><property name="score" value="0.9998198449611664" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i37]" time="1.140"><properties><property name="score" value="0.048420339822769165" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255037 (len: 505) is a human-generated sample, misclassified as AI-generated with confidence 0.04842034&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255037 (len: 505) is a human-generated sample, misclassified as AI-generated with confidence 0.04842034
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i32]" time="0.284"><properties><property name="score" value="0.9778472185134888" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i33]" time="2.466"><properties><property name="score" value="0.5959689855575562" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i34]" time="0.995"><properties><property name="score" value="0.9874178767204285" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i35]" time="2.452"><properties><property name="score" value="0.9883281707763671" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i36]" time="1.114"><properties><property name="score" value="0.9998198449611664" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i37]" time="1.289"><properties><property name="score" value="0.048420339822769165" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255037 (len: 505) is a human-generated sample, misclassified as AI-generated with confidence 0.04842034&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255037 (len: 505) is a human-generated sample, misclassified as AI-generated with confidence 0.04842034
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i38]" time="0.322"><properties><property name="score" value="0.7256137132644653" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255038 (len: 102) is a human-generated sample, misclassified as AI-generated with confidence 0.72561371&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255038 (len: 102) is a human-generated sample, misclassified as AI-generated with confidence 0.72561371
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i38]" time="0.241"><properties><property name="score" value="0.7256137132644653" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255038 (len: 102) is a human-generated sample, misclassified as AI-generated with confidence 0.72561371&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255038 (len: 102) is a human-generated sample, misclassified as AI-generated with confidence 0.72561371
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i39]" time="0.474"><properties><property name="score" value="0.9022691249847412" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i40]" time="2.656"><properties><property name="score" value="0.27648659944534304" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i41]" time="1.345"><properties><property name="score" value="0.9824651678403219" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i42]" time="1.284"><properties><property name="score" value="0.361291766166687" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i43]" time="0.265"><properties><property name="score" value="0.9992792010307312" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i44]" time="0.811"><properties><property name="score" value="0.9960125684738159" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i45]" time="3.170"><properties><property name="score" value="0.982548189163208" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i46]" time="2.603"><properties><property name="score" value="0.6579800963401794" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i47]" time="2.654"><properties><property name="score" value="0.9225528120994568" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i48]" time="2.762"><properties><property name="score" value="0.7329963246981303" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i49]" time="1.127"><properties><property name="score" value="0.0896996259689331" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255049 (len: 424) is a human-generated sample, misclassified as AI-generated with confidence 0.08969963&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255049 (len: 424) is a human-generated sample, misclassified as AI-generated with confidence 0.08969963
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i39]" time="0.473"><properties><property name="score" value="0.9022691249847412" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i40]" time="2.568"><properties><property name="score" value="0.27648659944534304" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i41]" time="1.146"><properties><property name="score" value="0.9824651678403219" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i42]" time="1.325"><properties><property name="score" value="0.361291766166687" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i43]" time="0.245"><properties><property name="score" value="0.9992792010307312" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i44]" time="0.612"><properties><property name="score" value="0.9960125684738159" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i45]" time="2.486"><properties><property name="score" value="0.982548189163208" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i46]" time="3.431"><properties><property name="score" value="0.6579800963401794" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i47]" time="2.713"><properties><property name="score" value="0.9225528120994568" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i48]" time="2.854"><properties><property name="score" value="0.7329963246981303" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i49]" time="1.040"><properties><property name="score" value="0.0896996259689331" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255049 (len: 424) is a human-generated sample, misclassified as AI-generated with confidence 0.08969963&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255049 (len: 424) is a human-generated sample, misclassified as AI-generated with confidence 0.08969963
     assert 'AI' == 'Human'
       - Human
-      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i0]" time="0.463"><properties><property name="score" value="0.9967435598373413" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255001 (text: What is this?  This is a neat little plugin to giv) is an LLM-generated sample, misclassified as human-generated with confidence 0.99674356&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255001 (text: What is this?  This is a neat little plugin to giv) is an LLM-generated sample, misclassified as human-generated with confidence 0.99674356
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i50]" time="2.393"><properties><property name="score" value="0.6270938158035279" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i51]" time="2.176"><properties><property name="score" value="0.27273480892181395" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i52]" time="0.304"><properties><property name="score" value="0.9998220801353455" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i53]" time="0.701"><properties><property name="score" value="0.92865851521492" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i54]" time="1.253"><properties><property name="score" value="0.99956014752388" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i55]" time="2.321"><properties><property name="score" value="0.5540543317794799" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i56]" time="2.141"><properties><property name="score" value="0.9659956336021424" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i57]" time="2.483"><properties><property name="score" value="0.9990162491798401" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i58]" time="2.413"><properties><property name="score" value="0.9985167384147644" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i59]" time="0.599"><properties><property name="score" value="0.8944965898990631" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i60]" time="2.358"><properties><property name="score" value="0.9997556209564209" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i61]" time="2.412"><properties><property name="score" value="0.6299816727638244" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i62]" time="0.174"><properties><property name="score" value="0.7945960760116577" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i63]" time="0.254"><properties><property name="score" value="0.9948675632476807" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i64]" time="0.204"><properties><property name="score" value="0.9937396049499512" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i65]" time="2.501"><properties><property name="score" value="0.988049104809761" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i66]" time="0.229"><properties><property name="score" value="0.9783594012260437" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i67]" time="1.949"><properties><property name="score" value="0.9922092258930206" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i68]" time="0.837"><properties><property name="score" value="0.9994023144245148" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i69]" time="0.656"><properties><property name="score" value="0.9806088805198669" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i70]" time="0.305"><properties><property name="score" value="0.9997535347938538" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i71]" time="2.534"><properties><property name="score" value="0.9956087946891785" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i72]" time="0.940"><properties><property name="score" value="0.9990681707859039" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i73]" time="1.278"><properties><property name="score" value="0.9997933506965637" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i74]" time="0.782"><properties><property name="score" value="0.00035774707794189453" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i75]" time="1.009"><properties><property name="score" value="0.9998205602169037" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i76]" time="0.557"><properties><property name="score" value="0.6158799529075623" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i77]" time="0.686"><properties><property name="score" value="0.010744303464889526" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i78]" time="2.391"><properties><property name="score" value="0.5589596271514893" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i79]" time="0.498"><properties><property name="score" value="0.9740335941314697" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i80]" time="2.376"><properties><property name="score" value="0.9860053777694702" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i81]" time="2.320"><properties><property name="score" value="0.9953763484954834" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i82]" time="0.203"><properties><property name="score" value="0.9818570017814636" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i83]" time="1.690"><properties><property name="score" value="0.996516764163971" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i84]" time="2.082"><properties><property name="score" value="0.6087557792663574" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i85]" time="0.707"><properties><property name="score" value="0.00816836953163147" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i86]" time="1.811"><properties><property name="score" value="0.9997578561306" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i87]" time="1.172"><properties><property name="score" value="0.9866769909858704" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i88]" time="0.460"><properties><property name="score" value="0.9997960925102234" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i89]" time="2.534"><properties><property name="score" value="0.08046053647994995" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255089 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.08046054&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255089 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.08046054
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i90]" time="0.198"><properties><property name="score" value="0.6618620157241821" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i91]" time="0.282"><properties><property name="score" value="0.9997391104698181" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i92]" time="2.409"><properties><property name="score" value="0.9091811180114746" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i93]" time="1.536"><properties><property name="score" value="0.9996700882911682" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i94]" time="0.923"><properties><property name="score" value="0.999814361333847" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i95]" time="2.411"><properties><property name="score" value="0.9933454155921936" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i96]" time="0.693"><failure message="RuntimeError: The expanded size of the tensor (539) must match the existing size (514) at non-singleton dimension 1.  Target sizes: [1, 539].  Tensor sizes: [1, 514]">E   RuntimeError: The expanded size of the tensor (539) must match the existing size (514) at non-singleton dimension 1.  Target sizes: [1, 539].  Tensor sizes: [1, 514]</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i97]" time="0.182"><properties><property name="score" value="0.9996429681777954" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i98]" time="0.963"><properties><property name="score" value="0.99540776014328" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i99]" time="0.220"><properties><property name="score" value="0.6158614158630371" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i100]" time="1.259"><properties><property name="score" value="0.9705980817476908" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i101]" time="2.367"><properties><property name="score" value="0.6226354837417603" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i102]" time="0.850"><properties><property name="score" value="0.0003025233745574951" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i103]" time="2.323"><properties><property name="score" value="0.19912527799606322" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i104]" time="0.706"><properties><property name="score" value="0.12030091881752014" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i105]" time="2.474"><properties><property name="score" value="0.5980462074279785" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i106]" time="1.744"><properties><property name="score" value="0.41522733370463055" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i107]" time="0.419"><properties><property name="score" value="0.9998192191123962" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i108]" time="0.811"><properties><property name="score" value="0.14455854892730713" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i109]" time="1.467"><properties><property name="score" value="0.22687806685765585" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i110]" time="0.301"><properties><property name="score" value="0.9861132502555847" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i111]" time="1.345"><properties><property name="score" value="0.9929361144701639" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i112]" time="1.272"><properties><property name="score" value="0.9818121194839478" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i113]" time="1.278"><properties><property name="score" value="0.9264129400253296" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i114]" time="0.485"><properties><property name="score" value="0.9998237490653992" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i115]" time="2.592"><properties><property name="score" value="0.6989482045173645" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i116]" time="2.474"><properties><property name="score" value="0.9965414047241211" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i117]" time="2.320"><properties><property name="score" value="0.49254806339740753" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i118]" time="1.398"><properties><property name="score" value="0.9910686612129211" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i119]" time="1.985"><properties><property name="score" value="0.9994007796049118" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i120]" time="0.607"><properties><property name="score" value="0.8510706424713135" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i121]" time="1.310"><properties><property name="score" value="0.4513130982716878" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i122]" time="1.203"><properties><property name="score" value="0.011024892330169678" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i123]" time="1.755"><properties><property name="score" value="0.5841767638921738" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i124]" time="0.952"><properties><property name="score" value="0.09872302412986755" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i125]" time="0.781"><properties><property name="score" value="0.01979959011077881" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i126]" time="0.783"><properties><property name="score" value="0.23591172695159912" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i127]" time="1.422"><properties><property name="score" value="0.38568369547526044" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i128]" time="2.875"><properties><property name="score" value="0.9939494371414185" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i129]" time="1.795"><properties><property name="score" value="0.9091822803020477" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i130]" time="1.845"><properties><property name="score" value="0.9535469561815262" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i131]" time="2.406"><properties><property name="score" value="0.9821225762367248" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i132]" time="0.230"><properties><property name="score" value="0.7774047255516052" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i133]" time="2.733"><properties><property name="score" value="0.6506359179814657" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i134]" time="0.500"><properties><property name="score" value="0.6261691451072693" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255134 (len: 227) is a human-generated sample, misclassified as AI-generated with confidence 0.62616915&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255134 (len: 227) is a human-generated sample, misclassified as AI-generated with confidence 0.62616915
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i135]" time="1.402"><properties><property name="score" value="0.9968183636665344" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i136]" time="0.216"><properties><property name="score" value="0.9135029315948486" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i137]" time="0.479"><properties><property name="score" value="0.9997634291648865" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i138]" time="0.441"><properties><property name="score" value="0.9886188507080078" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255138 (len: 184) is a human-generated sample, misclassified as AI-generated with confidence 0.98861885&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255138 (len: 184) is a human-generated sample, misclassified as AI-generated with confidence 0.98861885
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i139]" time="1.971"><properties><property name="score" value="0.9511344879865646" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i140]" time="0.448"><properties><property name="score" value="0.5366412997245789" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255140 (len: 142) is a human-generated sample, misclassified as AI-generated with confidence 0.5366413&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255140 (len: 142) is a human-generated sample, misclassified as AI-generated with confidence 0.5366413
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i141]" time="1.442"><properties><property name="score" value="0.8647339046001434" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i142]" time="2.468"><properties><property name="score" value="0.9962310791015625" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i143]" time="0.730"><properties><property name="score" value="0.8158917725086212" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i144]" time="2.431"><properties><property name="score" value="0.6019846796989441" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i145]" time="1.699"><properties><property name="score" value="0.1110960841178894" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i146]" time="0.919"><properties><property name="score" value="0.9628295600414276" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i147]" time="2.506"><properties><property name="score" value="0.9593395709991455" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i148]" time="2.523"><properties><property name="score" value="0.9812323053677877" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i149]" time="0.891"><properties><property name="score" value="0.00021088123321533203" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255149 (len: 363) is a human-generated sample, misclassified as AI-generated with confidence 0.00021088&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255149 (len: 363) is a human-generated sample, misclassified as AI-generated with confidence 0.00021088
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i150]" time="0.259"><properties><property name="score" value="0.9998137354850769" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i151]" time="0.835"><properties><property name="score" value="0.9971545040607452" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i152]" time="0.198"><properties><property name="score" value="0.8647211790084839" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i153]" time="0.202"><properties><property name="score" value="0.9997875094413757" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i154]" time="2.381"><properties><property name="score" value="0.8892017006874084" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i155]" time="2.477"><properties><property name="score" value="0.9852867245674133" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i156]" time="0.979"><properties><property name="score" value="0.9994515776634216" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i157]" time="1.362"><properties><property name="score" value="0.49514196316401166" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i158]" time="0.227"><properties><property name="score" value="0.9997168183326721" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i159]" time="1.877"><properties><property name="score" value="0.9998032301664352" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i160]" time="2.250"><properties><property name="score" value="0.9305805921554565" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i161]" time="1.004"><properties><property name="score" value="0.9475776851177216" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i162]" time="2.175"><properties><property name="score" value="0.5302713066339493" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i163]" time="2.443"><properties><property name="score" value="0.9980186462402344" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i164]" time="2.202"><properties><property name="score" value="0.9858382046222687" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i165]" time="1.267"><properties><property name="score" value="0.9624837239583334" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i166]" time="0.297"><properties><property name="score" value="0.9945158362388611" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i167]" time="0.163"><properties><property name="score" value="0.5262783765792847" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i168]" time="2.458"><properties><property name="score" value="0.6989508390426635" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i169]" time="2.580"><properties><property name="score" value="0.8227894504865011" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i170]" time="2.439"><properties><property name="score" value="0.9997873902320862" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i171]" time="2.096"><properties><property name="score" value="0.9991308897733688" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i172]" time="0.846"><properties><property name="score" value="0.9054249227046967" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i173]" time="1.720"><properties><property name="score" value="0.9601349433263143" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i174]" time="0.271"><properties><property name="score" value="0.999203622341156" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i175]" time="2.043"><properties><property name="score" value="0.9093873053789139" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i176]" time="1.067"><properties><property name="score" value="0.999671995639801" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i177]" time="0.719"><properties><property name="score" value="0.9155670404434204" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i178]" time="1.032"><properties><property name="score" value="0.444338063398997" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255178 (len: 423) is a human-generated sample, misclassified as AI-generated with confidence 0.44433806&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255178 (len: 423) is a human-generated sample, misclassified as AI-generated with confidence 0.44433806
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i179]" time="1.368"><properties><property name="score" value="0.8399171233177185" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i180]" time="0.894"><properties><property name="score" value="0.7615967094898224" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i181]" time="2.815"><properties><property name="score" value="0.5702897608280182" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255181 (len: 1023) is a human-generated sample, misclassified as AI-generated with confidence 0.57028976&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255181 (len: 1023) is a human-generated sample, misclassified as AI-generated with confidence 0.57028976
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i182]" time="0.954"><properties><property name="score" value="0.9005125164985657" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i183]" time="2.401"><properties><property name="score" value="0.9913132429122925" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i184]" time="0.611"><properties><property name="score" value="0.8495492935180664" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i185]" time="0.656"><properties><property name="score" value="0.09321713447570801" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i186]" time="0.454"><properties><property name="score" value="0.999782145023346" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i187]" time="1.182"><properties><property name="score" value="0.9998157918453217" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i188]" time="2.234"><properties><property name="score" value="0.7669108510017395" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i189]" time="1.240"><properties><property name="score" value="0.9607903361320496" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i190]" time="2.685"><properties><property name="score" value="0.9786059141159058" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i191]" time="2.804"><properties><property name="score" value="0.5965846061706543" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i192]" time="0.539"><properties><property name="score" value="0.9988358616828918" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i193]" time="3.008"><properties><property name="score" value="0.99976087808609" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i194]" time="2.525"><properties><property name="score" value="0.9869428157806397" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i195]" time="0.387"><properties><property name="score" value="0.9739251136779785" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255195 (len: 187) is a human-generated sample, misclassified as AI-generated with confidence 0.97392511&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255195 (len: 187) is a human-generated sample, misclassified as AI-generated with confidence 0.97392511
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i196]" time="0.735"><properties><property name="score" value="0.8736250698566437" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i197]" time="3.192"><properties><property name="score" value="0.4974241852760315" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i198]" time="1.693"><properties><property name="score" value="0.9814185301462809" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i199]" time="0.909"><properties><property name="score" value="0.815206915140152" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i200]" time="1.960"><properties><property name="score" value="0.49885278940200806" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255200 (len: 734) is a human-generated sample, misclassified as AI-generated with confidence 0.49885279&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255200 (len: 734) is a human-generated sample, misclassified as AI-generated with confidence 0.49885279
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i201]" time="2.937"><properties><property name="score" value="0.6596284866333008" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i202]" time="1.346"><properties><property name="score" value="0.9221746126810709" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i203]" time="0.156"><properties><property name="score" value="0.5073177814483643" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i204]" time="0.656"><properties><property name="score" value="0.760837197303772" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255204 (len: 228) is a human-generated sample, misclassified as AI-generated with confidence 0.7608372&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255204 (len: 228) is a human-generated sample, misclassified as AI-generated with confidence 0.7608372
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i205]" time="0.920"><properties><property name="score" value="0.9978813827037811" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i206]" time="2.286"><properties><property name="score" value="0.9136553168296814" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i207]" time="2.532"><properties><property name="score" value="0.9997764229774475" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i208]" time="0.662"><properties><property name="score" value="0.7979994118213654" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i209]" time="1.682"><properties><property name="score" value="0.929035946726799" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i210]" time="0.208"><properties><property name="score" value="0.9998146891593933" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i211]" time="1.122"><properties><property name="score" value="0.23154930273691812" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255211 (len: 447) is a human-generated sample, misclassified as AI-generated with confidence 0.2315493&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255211 (len: 447) is a human-generated sample, misclassified as AI-generated with confidence 0.2315493
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i212]" time="2.469"><properties><property name="score" value="0.9246889452139536" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i213]" time="0.231"><properties><property name="score" value="0.9992644190788269" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i214]" time="2.626"><properties><property name="score" value="0.9979465484619141" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i215]" time="1.032"><properties><property name="score" value="0.9839392602443695" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i216]" time="0.951"><properties><property name="score" value="0.2364269495010376" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i217]" time="2.342"><properties><property name="score" value="0.9997031837701797" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i218]" time="2.545"><properties><property name="score" value="0.6955123901367187" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i219]" time="3.460"><properties><property name="score" value="0.7214124898115793" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i220]" time="0.305"><properties><property name="score" value="0.9442246556282043" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i221]" time="0.435"><properties><property name="score" value="0.694334864616394" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i222]" time="2.649"><properties><property name="score" value="0.6693316698074341" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i223]" time="0.814"><properties><property name="score" value="0.9814737439155579" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i224]" time="1.348"><properties><property name="score" value="0.3791475296020508" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i225]" time="0.203"><properties><property name="score" value="0.7570915818214417" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i226]" time="0.822"><properties><property name="score" value="0.8827116787433624" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i227]" time="2.717"><properties><property name="score" value="0.6641123493512472" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i228]" time="2.902"><properties><property name="score" value="0.6547459959983826" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i229]" time="1.008"><properties><property name="score" value="0.9798719584941864" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i230]" time="0.406"><properties><property name="score" value="0.9874497652053833" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255230 (len: 145) is a human-generated sample, misclassified as AI-generated with confidence 0.98744977&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255230 (len: 145) is a human-generated sample, misclassified as AI-generated with confidence 0.98744977
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i231]" time="0.410"><properties><property name="score" value="0.9996953010559082" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255231 (len: 161) is a human-generated sample, misclassified as AI-generated with confidence 0.9996953&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255231 (len: 161) is a human-generated sample, misclassified as AI-generated with confidence 0.9996953
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i232]" time="0.993"><properties><property name="score" value="0.9986763894557953" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i233]" time="0.221"><properties><property name="score" value="0.9997331500053406" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i234]" time="0.184"><properties><property name="score" value="0.9986420273780823" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i235]" time="2.752"><properties><property name="score" value="0.9997337102890015" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i236]" time="2.762"><properties><property name="score" value="0.9425114393234253" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i237]" time="1.755"><properties><property name="score" value="0.29075706005096436" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i238]" time="0.288"><properties><property name="score" value="0.8549576997756958" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i239]" time="2.471"><properties><property name="score" value="0.9997880458831787" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i240]" time="2.217"><properties><property name="score" value="0.09285196661949158" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255240 (len: 809) is a human-generated sample, misclassified as AI-generated with confidence 0.09285197&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255240 (len: 809) is a human-generated sample, misclassified as AI-generated with confidence 0.09285197
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i241]" time="2.804"><properties><property name="score" value="0.1122020681699117" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255241 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.11220207&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255241 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.11220207
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i242]" time="2.485"><properties><property name="score" value="0.9635586142539978" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i243]" time="2.738"><properties><property name="score" value="0.7119329969088236" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i244]" time="2.132"><properties><property name="score" value="0.9928079098463058" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i245]" time="2.520"><properties><property name="score" value="0.9312599778175354" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i246]" time="0.212"><properties><property name="score" value="0.9706993103027344" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i247]" time="2.431"><properties><property name="score" value="0.12460075616836548" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255247 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.12460076&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255247 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.12460076
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i248]" time="1.304"><properties><property name="score" value="0.3659023642539978" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i249]" time="2.669"><properties><property name="score" value="0.18436744213104247" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255249 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.18436744&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255249 (len: 1024) is a human-generated sample, misclassified as AI-generated with confidence 0.18436744
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i250]" time="2.741"><properties><property name="score" value="0.9976030141115189" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i251]" time="3.074"><properties><property name="score" value="0.8787824988365174" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i252]" time="1.916"><properties><property name="score" value="0.9944134553273519" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i253]" time="0.874"><properties><property name="score" value="0.9971485137939453" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i254]" time="0.712"><properties><property name="score" value="0.8287554383277893" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i255]" time="2.903"><properties><property name="score" value="0.8915076375007629" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i256]" time="2.674"><properties><property name="score" value="0.9962790131568908" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i257]" time="0.183"><properties><property name="score" value="0.8452549576759338" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i258]" time="0.747"><properties><property name="score" value="0.9979833364486694" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i259]" time="2.827"><properties><property name="score" value="0.9433715581893921" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i260]" time="4.286"><properties><property name="score" value="0.5180151462554932" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i261]" time="6.368"><properties><property name="score" value="0.99968221783638" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i262]" time="3.786"><properties><property name="score" value="0.9961702704429627" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i263]" time="1.374"><properties><property name="score" value="0.9911569058895111" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i264]" time="3.144"><properties><property name="score" value="0.6651038328806559" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i265]" time="3.410"><properties><property name="score" value="0.9069875776767731" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i266]" time="0.396"><properties><property name="score" value="0.99982750415802" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i267]" time="2.909"><properties><property name="score" value="0.9449732542037964" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i268]" time="3.127"><properties><property name="score" value="0.9324176549911499" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i269]" time="2.650"><properties><property name="score" value="0.6932074785232544" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i270]" time="1.269"><properties><property name="score" value="0.9977475702762604" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i271]" time="0.844"><properties><property name="score" value="4.8279762268066406e-05" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i272]" time="2.468"><properties><property name="score" value="0.6068743228912353" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i273]" time="2.892"><properties><property name="score" value="0.9731637001037597" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i274]" time="0.927"><properties><property name="score" value="0.9990138113498688" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i275]" time="0.762"><properties><property name="score" value="0.0616203248500824" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255275 (len: 302) is a human-generated sample, misclassified as AI-generated with confidence 0.06162032&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255275 (len: 302) is a human-generated sample, misclassified as AI-generated with confidence 0.06162032
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i276]" time="0.759"><properties><property name="score" value="0.05853164196014404" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i277]" time="0.895"><properties><property name="score" value="0.9997464716434479" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i278]" time="0.809"><properties><property name="score" value="0.9940798580646515" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i279]" time="2.596"><properties><property name="score" value="0.9434233665466308" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i280]" time="1.108"><properties><property name="score" value="0.995629757642746" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i281]" time="0.698"><properties><property name="score" value="0.8067503571510315" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i282]" time="4.147"><properties><property name="score" value="0.6429866671562194" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i283]" time="4.006"><properties><property name="score" value="0.9407749056816102" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i284]" time="2.872"><properties><property name="score" value="0.983567214012146" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i285]" time="0.786"><properties><property name="score" value="0.16738349199295044" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i286]" time="2.754"><properties><property name="score" value="0.6698049187660218" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i287]" time="1.905"><properties><property name="score" value="0.9996131062507629" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i288]" time="1.355"><properties><property name="score" value="0.9870142936706543" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i289]" time="2.562"><properties><property name="score" value="0.9884985834360123" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i290]" time="2.795"><properties><property name="score" value="0.6000116467475891" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i291]" time="2.825"><properties><property name="score" value="0.2953288912773132" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i292]" time="1.080"><properties><property name="score" value="0.705899178981781" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i293]" time="2.246"><properties><property name="score" value="0.5072129368782043" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i294]" time="0.223"><properties><property name="score" value="0.929703950881958" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i295]" time="1.996"><properties><property name="score" value="0.9996638149023056" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i296]" time="0.757"><properties><property name="score" value="0.7567206919193268" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i297]" time="1.512"><properties><property name="score" value="0.1868351697921753" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i298]" time="0.847"><properties><property name="score" value="0.9939405918121338" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i299]" time="2.979"><properties><property name="score" value="0.9996690630912781" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i300]" time="2.683"><properties><property name="score" value="0.5157073438167572" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i301]" time="1.382"><properties><property name="score" value="0.9972848097483317" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i302]" time="2.494"><properties><property name="score" value="0.9997820258140564" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i303]" time="2.674"><properties><property name="score" value="0.999788224697113" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i304]" time="0.208"><properties><property name="score" value="0.9748216271400452" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i305]" time="0.232"><properties><property name="score" value="0.9998229146003723" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i306]" time="0.877"><properties><property name="score" value="0.9993984401226044" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i307]" time="0.884"><properties><property name="score" value="0.9122644364833832" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i308]" time="1.796"><properties><property name="score" value="0.6154283434152603" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i309]" time="0.955"><properties><property name="score" value="0.047338247299194336" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i310]" time="2.744"><properties><property name="score" value="0.9975978016853333" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i311]" time="2.900"><properties><property name="score" value="0.602080500125885" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i312]" time="1.674"><properties><property name="score" value="0.3146459261576335" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255312 (len: 641) is a human-generated sample, misclassified as AI-generated with confidence 0.31464593&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255312 (len: 641) is a human-generated sample, misclassified as AI-generated with confidence 0.31464593
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i313]" time="1.264"><properties><property name="score" value="0.9597663581371307" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i314]" time="1.062"><properties><property name="score" value="0.9916037619113922" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i315]" time="1.475"><properties><property name="score" value="0.9204014738400778" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i316]" time="1.114"><properties><property name="score" value="0.9517025351524353" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i317]" time="0.355"><properties><property name="score" value="0.7616866827011108" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i318]" time="4.314"><properties><property name="score" value="0.998293662071228" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i319]" time="2.508"><properties><property name="score" value="0.9994392246007919" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i320]" time="0.445"><properties><property name="score" value="0.9995920062065125" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i321]" time="2.126"><properties><property name="score" value="0.31485438346862793" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i322]" time="2.742"><properties><property name="score" value="0.9840458482503891" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i323]" time="2.784"><properties><property name="score" value="0.9941763043403625" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i324]" time="3.087"><properties><property name="score" value="0.99957355260849" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i325]" time="0.492"><properties><property name="score" value="0.9998257756233215" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i326]" time="0.431"><properties><property name="score" value="0.9994305968284607" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i327]" time="2.366"><properties><property name="score" value="0.6851132154464722" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i328]" time="3.055"><properties><property name="score" value="0.5983311295509338" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i329]" time="0.205"><properties><property name="score" value="0.9995595812797546" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i330]" time="2.349"><properties><property name="score" value="0.6394256830215455" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i331]" time="1.945"><properties><property name="score" value="0.9993870705366135" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i332]" time="1.689"><properties><property name="score" value="0.9997101227442423" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i333]" time="0.387"><properties><property name="score" value="0.8306254744529724" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i334]" time="0.589"><properties><property name="score" value="0.9998063445091248" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i335]" time="1.149"><properties><property name="score" value="0.9920057952404022" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i336]" time="0.603"><properties><property name="score" value="0.2071707844734192" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i337]" time="0.530"><properties><property name="score" value="0.9997798800468445" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i338]" time="0.679"><properties><property name="score" value="0.06858500838279724" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i339]" time="2.296"><properties><property name="score" value="0.9845941305160523" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i340]" time="2.518"><properties><property name="score" value="0.919108659029007" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i341]" time="0.953"><properties><property name="score" value="0.9579263031482697" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i342]" time="1.589"><properties><property name="score" value="0.9994104901949564" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i343]" time="0.252"><properties><property name="score" value="0.5920212864875793" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255343 (len: 82) is a human-generated sample, misclassified as AI-generated with confidence 0.59202129&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255343 (len: 82) is a human-generated sample, misclassified as AI-generated with confidence 0.59202129
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i344]" time="1.522"><properties><property name="score" value="0.999766488869985" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i345]" time="0.686"><properties><property name="score" value="0.8605888783931732" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i346]" time="1.384"><properties><property name="score" value="0.9972351789474487" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i347]" time="0.974"><properties><property name="score" value="0.9962185323238373" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i348]" time="2.465"><properties><property name="score" value="0.7411656479040781" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i349]" time="0.392"><properties><property name="score" value="0.9998263716697693" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i350]" time="1.192"><properties><property name="score" value="0.9944879114627838" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i351]" time="0.204"><properties><property name="score" value="0.8022099733352661" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i352]" time="1.214"><properties><property name="score" value="0.998905748128891" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i353]" time="2.837"><properties><property name="score" value="0.7068503002325693" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i354]" time="0.190"><properties><property name="score" value="0.6997880339622498" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255354 (len: 58) is a human-generated sample, misclassified as AI-generated with confidence 0.69978803&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255354 (len: 58) is a human-generated sample, misclassified as AI-generated with confidence 0.69978803
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i355]" time="2.571"><properties><property name="score" value="0.4201407730579376" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i356]" time="0.803"><properties><property name="score" value="0.010107249021530151" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i357]" time="2.727"><properties><property name="score" value="0.6670459985733033" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i358]" time="1.363"><properties><property name="score" value="0.9838470419247946" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i359]" time="0.717"><properties><property name="score" value="0.012442529201507568" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i360]" time="0.177"><properties><property name="score" value="0.6215845346450806" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255360 (len: 51) is a human-generated sample, misclassified as AI-generated with confidence 0.62158453&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255360 (len: 51) is a human-generated sample, misclassified as AI-generated with confidence 0.62158453
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i361]" time="2.429"><properties><property name="score" value="0.9551822185516358" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i362]" time="1.457"><properties><property name="score" value="0.194048543771108" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255362 (len: 489) is a human-generated sample, misclassified as AI-generated with confidence 0.19404854&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255362 (len: 489) is a human-generated sample, misclassified as AI-generated with confidence 0.19404854
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i363]" time="0.479"><properties><property name="score" value="0.9996726512908936" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i364]" time="2.457"><properties><property name="score" value="0.6010900020599366" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i365]" time="0.403"><properties><property name="score" value="0.9993131160736084" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i366]" time="2.674"><properties><property name="score" value="0.6697026133537293" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i367]" time="0.483"><properties><property name="score" value="0.9998308420181274" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i368]" time="2.313"><properties><property name="score" value="0.9934981107711792" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i369]" time="1.831"><properties><property name="score" value="0.9419496854146322" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i370]" time="2.495"><properties><property name="score" value="0.9713339924812316" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i371]" time="2.825"><properties><property name="score" value="0.6352494120597839" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i372]" time="1.661"><properties><property name="score" value="0.9967522223790487" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i373]" time="0.242"><properties><property name="score" value="0.9996663331985474" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i374]" time="2.453"><properties><property name="score" value="0.6213465929031372" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i375]" time="1.836"><properties><property name="score" value="0.9936741590499878" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i376]" time="2.517"><properties><property name="score" value="0.9920497179031372" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i377]" time="0.964"><properties><property name="score" value="1.919269561767578e-05" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i378]" time="0.536"><properties><property name="score" value="0.10967493057250977" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i379]" time="0.725"><properties><property name="score" value="0.036823123693466187" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i380]" time="0.216"><properties><property name="score" value="0.833694338798523" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i381]" time="1.495"><properties><property name="score" value="0.9183606306711832" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i382]" time="1.843"><properties><property name="score" value="0.9985030690828959" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i383]" time="0.970"><properties><property name="score" value="0.9985866248607635" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i384]" time="2.492"><properties><property name="score" value="0.6870572209358216" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i385]" time="0.236"><properties><property name="score" value="0.8542024493217468" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255385 (len: 59) is a human-generated sample, misclassified as AI-generated with confidence 0.85420245&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255385 (len: 59) is a human-generated sample, misclassified as AI-generated with confidence 0.85420245
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i386]" time="2.597"><properties><property name="score" value="0.9835602402687073" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i387]" time="2.714"><properties><property name="score" value="0.9300037264823914" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i388]" time="2.557"><properties><property name="score" value="0.9967865109443664" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i389]" time="0.234"><properties><property name="score" value="0.9995842576026917" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i390]" time="1.537"><properties><property name="score" value="0.10720725357532501" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i391]" time="0.223"><properties><property name="score" value="0.9949228167533875" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i392]" time="1.223"><properties><property name="score" value="0.7815245787302653" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i393]" time="2.290"><properties><property name="score" value="0.98918117582798" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i394]" time="3.103"><properties><property name="score" value="0.999371874332428" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i395]" time="4.618"><properties><property name="score" value="0.5557099431753159" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i396]" time="2.635"><properties><property name="score" value="0.8088749647140503" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255396 (len: 422) is a human-generated sample, misclassified as AI-generated with confidence 0.80887496&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255396 (len: 422) is a human-generated sample, misclassified as AI-generated with confidence 0.80887496
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i397]" time="2.945"><properties><property name="score" value="0.907037901878357" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i398]" time="2.684"><properties><property name="score" value="0.9978402614593506" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i399]" time="1.830"><properties><property name="score" value="0.338957945505778" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i400]" time="0.501"><properties><property name="score" value="0.9994282126426697" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i401]" time="2.315"><properties><property name="score" value="0.9875274777412415" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i402]" time="0.432"><properties><property name="score" value="0.999528169631958" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i403]" time="0.869"><properties><property name="score" value="0.9955301284790039" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i404]" time="1.927"><properties><property name="score" value="0.9984803994496664" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i405]" time="1.255"><properties><property name="score" value="0.39586130777994794" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i406]" time="1.755"><properties><property name="score" value="0.9996432065963745" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i407]" time="2.049"><properties><property name="score" value="0.3135975996653239" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i408]" time="1.991"><properties><property name="score" value="0.888505756855011" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i409]" time="1.873"><properties><property name="score" value="0.9997315406799316" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i410]" time="1.257"><properties><property name="score" value="0.9599198698997498" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i411]" time="0.719"><properties><property name="score" value="0.9743247628211975" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i412]" time="1.496"><properties><property name="score" value="0.99966428677241" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i413]" time="1.063"><properties><property name="score" value="0.9990972578525543" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i414]" time="1.384"><properties><property name="score" value="0.31924806038538617" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i415]" time="2.594"><properties><property name="score" value="0.3536559045314789" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i416]" time="0.868"><properties><property name="score" value="0.9862883985042572" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i417]" time="0.641"><properties><property name="score" value="0.9249402284622192" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i418]" time="0.425"><properties><property name="score" value="0.9964147806167603" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i419]" time="0.226"><properties><property name="score" value="0.9798453450202942" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i420]" time="1.528"><properties><property name="score" value="0.9978712995847067" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i421]" time="0.725"><properties><property name="score" value="0.9152970314025879" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i422]" time="0.433"><properties><property name="score" value="0.9945303797721863" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255422 (len: 178) is a human-generated sample, misclassified as AI-generated with confidence 0.99453038&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255422 (len: 178) is a human-generated sample, misclassified as AI-generated with confidence 0.99453038
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i423]" time="2.218"><properties><property name="score" value="0.6804011702537537" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i424]" time="2.110"><properties><property name="score" value="0.6843551874160767" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i425]" time="1.012"><properties><property name="score" value="0.9975864291191101" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i426]" time="0.871"><properties><property name="score" value="0.9910227060317993" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i427]" time="0.301"><properties><property name="score" value="0.9884185194969177" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255427 (len: 84) is a human-generated sample, misclassified as AI-generated with confidence 0.98841852&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255427 (len: 84) is a human-generated sample, misclassified as AI-generated with confidence 0.98841852
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i428]" time="1.981"><properties><property name="score" value="0.9995418041944504" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i429]" time="0.734"><properties><property name="score" value="0.985821396112442" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i430]" time="1.155"><properties><property name="score" value="0.999723345041275" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i431]" time="1.110"><properties><property name="score" value="0.013919532299041748" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i432]" time="0.272"><properties><property name="score" value="0.998404324054718" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i433]" time="0.663"><properties><property name="score" value="0.7635983228683472" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i434]" time="1.394"><properties><property name="score" value="0.8917379577954611" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i435]" time="0.215"><properties><property name="score" value="0.9867053627967834" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i436]" time="2.344"><properties><property name="score" value="0.9994548916816711" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i437]" time="1.277"><properties><property name="score" value="0.4769858717918396" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i438]" time="0.646"><properties><property name="score" value="0.9996871948242188" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i439]" time="1.855"><properties><property name="score" value="0.49987681210041046" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i440]" time="1.427"><properties><property name="score" value="0.984759251276652" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i441]" time="1.702"><properties><property name="score" value="0.8941818922758102" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i442]" time="0.753"><properties><property name="score" value="0.0753740668296814" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255442 (len: 230) is a human-generated sample, misclassified as AI-generated with confidence 0.07537407&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255442 (len: 230) is a human-generated sample, misclassified as AI-generated with confidence 0.07537407
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i443]" time="2.687"><properties><property name="score" value="0.9997035503387451" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i444]" time="1.856"><properties><property name="score" value="0.0953722894191742" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i445]" time="1.734"><properties><property name="score" value="0.8825029879808426" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i446]" time="2.197"><properties><property name="score" value="0.9064329266548157" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i447]" time="2.587"><properties><property name="score" value="0.9889422535896302" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i448]" time="2.066"><properties><property name="score" value="0.8935689926147461" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i449]" time="1.289"><properties><property name="score" value="0.37011416753133136" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i450]" time="0.212"><properties><property name="score" value="0.9281619787216187" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i451]" time="2.337"><properties><property name="score" value="0.9997156858444214" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i452]" time="2.858"><properties><property name="score" value="0.6894926309585572" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i453]" time="1.317"><properties><property name="score" value="0.9988916317621866" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i454]" time="2.526"><properties><property name="score" value="0.9844942092895508" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i455]" time="2.546"><properties><property name="score" value="0.9474755078554153" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i456]" time="2.711"><properties><property name="score" value="0.7351000209649404" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i457]" time="0.375"><properties><property name="score" value="0.9994434714317322" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i458]" time="0.318"><properties><property name="score" value="0.6687248349189758" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255458 (len: 102) is a human-generated sample, misclassified as AI-generated with confidence 0.66872483&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255458 (len: 102) is a human-generated sample, misclassified as AI-generated with confidence 0.66872483
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i459]" time="1.313"><properties><property name="score" value="0.8481436371803284" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i460]" time="0.478"><properties><property name="score" value="0.5197636485099792" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i461]" time="1.972"><properties><property name="score" value="0.6086466461420059" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i462]" time="0.576"><properties><property name="score" value="0.9997522234916687" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i463]" time="1.554"><properties><property name="score" value="0.9995172222455343" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i464]" time="1.376"><properties><property name="score" value="0.9888548851013184" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i465]" time="2.806"><properties><property name="score" value="0.9994724541902542" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i466]" time="1.822"><properties><property name="score" value="0.5833948701620102" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i467]" time="1.539"><properties><property name="score" value="0.9981088240941366" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i468]" time="2.001"><properties><property name="score" value="0.9900354892015457" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i469]" time="0.199"><properties><property name="score" value="0.891660213470459" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i470]" time="2.344"><properties><property name="score" value="0.5652628779411316" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i471]" time="0.738"><properties><property name="score" value="0.9997907876968384" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i472]" time="0.856"><properties><property name="score" value="0.9642249345779419" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i473]" time="0.779"><properties><property name="score" value="0.9894407391548157" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i474]" time="2.341"><properties><property name="score" value="0.9916878938674927" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i475]" time="1.982"><properties><property name="score" value="0.9134259521961212" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i476]" time="1.354"><properties><property name="score" value="0.4530043403307597" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i477]" time="1.244"><properties><property name="score" value="0.34958215554555255" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i478]" time="0.886"><properties><property name="score" value="0.9997910559177399" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i479]" time="1.884"><properties><property name="score" value="0.06362304091453552" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i480]" time="2.334"><properties><property name="score" value="0.65437912940979" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i481]" time="2.365"><properties><property name="score" value="0.6764323314030966" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i482]" time="2.396"><properties><property name="score" value="0.6260708212852478" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i483]" time="0.718"><properties><property name="score" value="0.2220645546913147" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i484]" time="2.104"><properties><property name="score" value="0.9325633883476258" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i485]" time="0.430"><properties><property name="score" value="0.9996918439865112" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i486]" time="2.457"><properties><property name="score" value="0.870807925860087" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i487]" time="1.593"><properties><property name="score" value="0.9333720207214355" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i488]" time="0.427"><properties><property name="score" value="0.8716211915016174" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i489]" time="2.749"><properties><property name="score" value="0.8105617880821228" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i490]" time="0.279"><properties><property name="score" value="0.999062716960907" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i491]" time="1.921"><properties><property name="score" value="9.277462959289551e-05" /></properties><failure message="AssertionError: samples/webtext.test.jsonl:255491 (len: 780) is a human-generated sample, misclassified as AI-generated with confidence 9.277e-05&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/webtext.test.jsonl:255491 (len: 780) is a human-generated sample, misclassified as AI-generated with confidence 9.277e-05
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i492]" time="1.133"><properties><property name="score" value="0.3606734077135722" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i493]" time="0.517"><properties><property name="score" value="0.9998014569282532" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i494]" time="3.070"><properties><property name="score" value="0.6186677694320679" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i495]" time="3.337"><properties><property name="score" value="0.9772914528846741" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i496]" time="1.948"><properties><property name="score" value="0.3307531674702962" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i497]" time="0.708"><properties><property name="score" value="0.9979871511459351" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i498]" time="3.467"><properties><property name="score" value="0.14825217922528586" /></properties></testcase><testcase classname="test_roberta_detect" name="test_human_jsonl[i499]" time="3.619"><properties><property name="score" value="0.6335857629776" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i0]" time="0.523"><properties><property name="score" value="0.9967435598373413" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255001 (text: What is this?  This is a neat little plugin to giv) is an LLM-generated sample, misclassified as human-generated with confidence 0.99674356&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255001 (text: What is this?  This is a neat little plugin to giv) is an LLM-generated sample, misclassified as human-generated with confidence 0.99674356
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i1]" time="1.808"><properties><property name="score" value="0.9909460544586182" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i2]" time="1.342"><properties><property name="score" value="0.010897129774093628" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i3]" time="3.991"><properties><property name="score" value="0.5829756498336792" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i4]" time="2.948"><properties><property name="score" value="0.9116669148206711" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i5]" time="0.747"><properties><property name="score" value="0.7837815284729004" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255006 (text: ISIS is breeding from among the refugees forced ou) is an LLM-generated sample, misclassified as human-generated with confidence 0.78378153&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255006 (text: ISIS is breeding from among the refugees forced ou) is an LLM-generated sample, misclassified as human-generated with confidence 0.78378153
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i1]" time="1.674"><properties><property name="score" value="0.9909460544586182" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i2]" time="1.075"><properties><property name="score" value="0.010897129774093628" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i3]" time="2.955"><properties><property name="score" value="0.5829756498336792" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i4]" time="3.202"><properties><property name="score" value="0.9116669148206711" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i5]" time="0.638"><properties><property name="score" value="0.7837815284729004" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255006 (text: ISIS is breeding from among the refugees forced ou) is an LLM-generated sample, misclassified as human-generated with confidence 0.78378153&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255006 (text: ISIS is breeding from among the refugees forced ou) is an LLM-generated sample, misclassified as human-generated with confidence 0.78378153
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i6]" time="2.946"><properties><property name="score" value="0.014344975352287292" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255007 (text: More than a third of population as of recently, Ma) is an LLM-generated sample, misclassified as human-generated with confidence 0.01434498&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255007 (text: More than a third of population as of recently, Ma) is an LLM-generated sample, misclassified as human-generated with confidence 0.01434498
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i6]" time="3.114"><properties><property name="score" value="0.014344975352287292" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255007 (text: More than a third of population as of recently, Ma) is an LLM-generated sample, misclassified as human-generated with confidence 0.01434498&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255007 (text: More than a third of population as of recently, Ma) is an LLM-generated sample, misclassified as human-generated with confidence 0.01434498
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i7]" time="2.064"><properties><property name="score" value="0.3342360059420268" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i8]" time="2.664"><properties><property name="score" value="0.9993244647979737" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i9]" time="1.597"><properties><property name="score" value="0.9996711413065592" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i10]" time="0.824"><properties><property name="score" value="0.8524591326713562" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i11]" time="1.389"><properties><property name="score" value="0.998836100101471" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i12]" time="2.508"><properties><property name="score" value="0.9285116553306579" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i13]" time="0.447"><properties><property name="score" value="0.9919939041137695" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i14]" time="2.402"><properties><property name="score" value="0.936974823474884" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i15]" time="2.808"><properties><property name="score" value="0.9996443271636963" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i16]" time="0.330"><properties><property name="score" value="0.98846834897995" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i17]" time="2.997"><properties><property name="score" value="0.951131135225296" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i18]" time="0.768"><properties><property name="score" value="0.9992277026176453" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i19]" time="2.609"><properties><property name="score" value="0.9996503114700317" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i20]" time="2.465"><properties><property name="score" value="0.9903939366340637" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i21]" time="2.730"><properties><property name="score" value="0.058394432067871094" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i22]" time="2.761"><properties><property name="score" value="0.5787507742643356" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i23]" time="3.720"><properties><property name="score" value="0.8225263237953186" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i24]" time="0.830"><properties><property name="score" value="0.06563839316368103" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i25]" time="2.268"><properties><property name="score" value="0.5425450205802917" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i26]" time="0.582"><properties><property name="score" value="0.9720883369445801" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i27]" time="2.914"><properties><property name="score" value="0.9198516011238098" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i28]" time="0.525"><properties><property name="score" value="0.9257940649986267" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i29]" time="0.817"><properties><property name="score" value="0.999783456325531" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i30]" time="1.998"><properties><property name="score" value="0.11633069813251495" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i31]" time="0.237"><properties><property name="score" value="0.9927876591682434" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i32]" time="1.444"><properties><property name="score" value="0.9460114439328512" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i33]" time="0.409"><properties><property name="score" value="0.9967895150184631" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255034 (text: Education K-12  World Affairs Middle East &amp; North ) is an LLM-generated sample, misclassified as human-generated with confidence 0.99678952&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255034 (text: Education K-12  World Affairs Middle East &amp; North ) is an LLM-generated sample, misclassified as human-generated with confidence 0.99678952
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i7]" time="1.783"><properties><property name="score" value="0.3342360059420268" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i8]" time="2.826"><properties><property name="score" value="0.9993244647979737" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i9]" time="1.685"><properties><property name="score" value="0.9996711413065592" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i10]" time="0.873"><properties><property name="score" value="0.8524591326713562" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i11]" time="1.574"><properties><property name="score" value="0.998836100101471" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i12]" time="3.115"><properties><property name="score" value="0.9285116553306579" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i13]" time="0.427"><properties><property name="score" value="0.9919939041137695" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i14]" time="1.469"><properties><property name="score" value="0.936974823474884" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i15]" time="2.614"><properties><property name="score" value="0.9996443271636963" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i16]" time="0.303"><properties><property name="score" value="0.98846834897995" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i17]" time="2.855"><properties><property name="score" value="0.951131135225296" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i18]" time="0.781"><properties><property name="score" value="0.9992277026176453" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i19]" time="2.783"><properties><property name="score" value="0.9996503114700317" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i20]" time="2.431"><properties><property name="score" value="0.9903939366340637" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i21]" time="2.673"><properties><property name="score" value="0.058394432067871094" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i22]" time="2.129"><properties><property name="score" value="0.5787507742643356" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i23]" time="3.064"><properties><property name="score" value="0.8225263237953186" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i24]" time="0.788"><properties><property name="score" value="0.06563839316368103" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i25]" time="2.440"><properties><property name="score" value="0.5425450205802917" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i26]" time="0.451"><properties><property name="score" value="0.9720883369445801" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i27]" time="2.476"><properties><property name="score" value="0.9198516011238098" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i28]" time="0.529"><properties><property name="score" value="0.9257940649986267" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i29]" time="0.996"><properties><property name="score" value="0.999783456325531" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i30]" time="2.041"><properties><property name="score" value="0.11633069813251495" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i31]" time="0.268"><properties><property name="score" value="0.9927876591682434" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i32]" time="1.380"><properties><property name="score" value="0.9460114439328512" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i33]" time="0.407"><properties><property name="score" value="0.9967895150184631" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255034 (text: Education K-12  World Affairs Middle East &amp; North ) is an LLM-generated sample, misclassified as human-generated with confidence 0.99678952&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255034 (text: Education K-12  World Affairs Middle East &amp; North ) is an LLM-generated sample, misclassified as human-generated with confidence 0.99678952
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i34]" time="1.394"><properties><property name="score" value="0.31498650709788006" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i35]" time="3.269"><properties><property name="score" value="0.27188462018966675" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i36]" time="2.644"><properties><property name="score" value="0.9997288435697556" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i37]" time="0.453"><properties><property name="score" value="0.9997928738594055" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i38]" time="0.677"><properties><property name="score" value="0.9608039259910583" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i39]" time="2.651"><properties><property name="score" value="0.997704267501831" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i40]" time="1.012"><properties><property name="score" value="0.9972692728042603" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i41]" time="2.884"><properties><property name="score" value="0.9979674100875855" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i42]" time="0.573"><properties><property name="score" value="0.9194316267967224" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i43]" time="3.238"><properties><property name="score" value="0.5372697114944458" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i44]" time="3.141"><properties><property name="score" value="0.5620985150337219" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i45]" time="1.552"><properties><property name="score" value="0.9996896187464396" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i46]" time="0.971"><properties><property name="score" value="0.8975739479064941" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255047 (text: 16 #geekyvimcovers submitted after watching this v) is an LLM-generated sample, misclassified as human-generated with confidence 0.89757395&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255047 (text: 16 #geekyvimcovers submitted after watching this v) is an LLM-generated sample, misclassified as human-generated with confidence 0.89757395
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i34]" time="1.213"><properties><property name="score" value="0.31498650709788006" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i35]" time="2.763"><properties><property name="score" value="0.27188462018966675" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i36]" time="2.469"><properties><property name="score" value="0.9997288435697556" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i37]" time="0.373"><properties><property name="score" value="0.9997928738594055" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i38]" time="0.708"><properties><property name="score" value="0.9608039259910583" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i39]" time="2.592"><properties><property name="score" value="0.997704267501831" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i40]" time="1.041"><properties><property name="score" value="0.9972692728042603" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i41]" time="2.858"><properties><property name="score" value="0.9979674100875855" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i42]" time="0.509"><properties><property name="score" value="0.9194316267967224" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i43]" time="2.867"><properties><property name="score" value="0.5372697114944458" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i44]" time="2.521"><properties><property name="score" value="0.5620985150337219" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i45]" time="1.297"><properties><property name="score" value="0.9996896187464396" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i46]" time="0.874"><properties><property name="score" value="0.8975739479064941" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255047 (text: 16 #geekyvimcovers submitted after watching this v) is an LLM-generated sample, misclassified as human-generated with confidence 0.89757395&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255047 (text: 16 #geekyvimcovers submitted after watching this v) is an LLM-generated sample, misclassified as human-generated with confidence 0.89757395
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i47]" time="2.642"><properties><property name="score" value="0.611871874332428" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i48]" time="0.919"><properties><property name="score" value="0.9992437362670898" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i49]" time="1.570"><properties><property name="score" value="0.8716740806897482" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Glacier Ridge Christian School\n\nGlacier Ridge Christian School is a private Christian school in Johnstown, Ohio. It was founded in the fall of 1999 by Gary and Tammy Smith.\n\nThe school started with just 13 students in grades 5-8. By the end of the first year, the school had grown to 65 students in those same grades. As the school has continued to grow, so has the number of teachers.\n\nBy the year 2009, Glacier Ridge had expanded to offer a preschool through 12th grade education, and the school had grown to almost 200 students. In the fall of 2013, the enrollment was over 300 students.\n\nThe school started with the first graduating class of 2002. They are now on their 11th graduating class, and the school's graduates continue to receive acceptance to universities across the United States. The school offers several Advanced Placement classes.\n\nGlacier Ridge has had its students participate in various competitions and competitions. Most notable are the sports teams, especially their basketball teams, which have qualified for state competition multiple times. The school has also had various music groups participate in state and regional competitions, and in the 2010-2011 school year, the school sent students to participate in an International Children's Games competition in Romania.\n\nIn the 2010-2011 school year, a group of the school's teachers were sent to Haiti to help establish a new orphanage, &quot;Eli's Children's House&quot;, which was dedicated to Eli Pierre, who had died in a motorcycle accident. They were instrumental in bringing the orphanage into the community by helping establish new programs for the children, including English as a Second Language, Computer Lab, and Physical Education classes. The teachers assisted in the education of the orphans as well as working with the staff on the property.\n\nThe school has also taken various trips over the years, including trips to Virginia and New York City. The school also hosted an educational trip to Washington D.C. for one of their graduating classes.\n\nThe school has participated in various community events and service projects. One of the most notable is their annual &quot;Twilight Prom&quot; that is held each spring. Each year, the school hosts a prom for local high school students who have special needs. The students' dates are typically their own siblings.\n\nOn March 21, 2011, a fire destroyed the school's auditorium, causing over $1 million in damages. After the school received an outpouring of support, the school was able to begin construction on a new auditorium in time for the next school year.\n\nIn 2015, Glacier Ridge Christian School became part of the Christian School Association of Northeast Ohio.\n\nThe School is affiliated with the Association of Christian Schools International, the Ohio Christian Education Association, and the Christian School Association of Northeast Ohio.\n\n]" time="1.808"><properties><property name="score" value="0.999787708123525" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[2003 NCAA Division I-AA football season\n\nThe 2003 NCAA Division I-AA football season, part of college football in the United States organized by the National Collegiate Athletic Association at the Division I Football Championship Subdivision level, began on September 6, 2003, and concluded with the National Championship Game of the NCAA Division I Football Championship on December 21, 2003, at Finley Stadium in Chattanooga, Tennessee. The Georgia Southern Eagles won their first I-AA championship, defeating the Furman Paladins by a final score of 27\u221210. \n\n&lt;nowiki&gt;*&lt;/nowiki&gt; &quot;Denotes host institution&quot;\n\n&lt;nowiki&gt;*&lt;/nowiki&gt; &quot;Denotes host institution&quot;\n\n&lt;nowiki&gt;*&lt;/nowiki&gt; &quot;Denotes host institution&quot;\n]" time="0.431"><properties><property name="score" value="0.9993891716003418" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99938917&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99938917
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i47]" time="2.488"><properties><property name="score" value="0.611871874332428" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i48]" time="0.920"><properties><property name="score" value="0.9992437362670898" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i49]" time="1.389"><properties><property name="score" value="0.8716740806897482" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i50]" time="0.459"><properties><property name="score" value="0.9127016663551331" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i51]" time="2.796"><properties><property name="score" value="0.9949163794517517" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i52]" time="2.930"><properties><property name="score" value="0.9297672152519226" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i53]" time="0.362"><properties><property name="score" value="0.9998155236244202" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i54]" time="1.010"><properties><property name="score" value="0.7459888259569804" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i55]" time="1.139"><properties><property name="score" value="0.9994727969169617" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i56]" time="1.378"><properties><property name="score" value="0.3374236226081848" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i57]" time="0.543"><properties><property name="score" value="0.9953859448432922" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i58]" time="0.905"><properties><property name="score" value="0.9942357540130615" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i59]" time="2.758"><properties><property name="score" value="0.9964746733506521" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i60]" time="0.465"><properties><property name="score" value="0.9993060827255249" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i61]" time="0.970"><properties><property name="score" value="0.999697208404541" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i62]" time="0.792"><properties><property name="score" value="0.8535504043102264" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i63]" time="0.764"><properties><property name="score" value="0.24081149697303772" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i64]" time="2.927"><properties><property name="score" value="0.6387940883636475" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i65]" time="2.489"><properties><property name="score" value="0.6422924876213074" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i66]" time="2.545"><properties><property name="score" value="0.9914404034614563" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i67]" time="2.653"><properties><property name="score" value="0.18822550773620605" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i68]" time="1.789"><properties><property name="score" value="0.996404305100441" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i69]" time="0.809"><properties><property name="score" value="0.9998021423816681" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i70]" time="1.672"><properties><property name="score" value="0.889485701918602" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i71]" time="1.756"><properties><property name="score" value="0.9890483468770981" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i72]" time="2.430"><properties><property name="score" value="0.5777090787887573" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i73]" time="0.792"><properties><property name="score" value="0.9827710390090942" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i74]" time="1.441"><properties><property name="score" value="0.9990652203559875" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i75]" time="1.208"><properties><property name="score" value="0.9991807341575623" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i76]" time="0.555"><properties><property name="score" value="0.8225718140602112" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i77]" time="0.290"><properties><property name="score" value="0.7510509490966797" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255078 (text: Hurst Productions' love of Ken Burns documentary s) is an LLM-generated sample, misclassified as human-generated with confidence 0.75105095&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255078 (text: Hurst Productions' love of Ken Burns documentary s) is an LLM-generated sample, misclassified as human-generated with confidence 0.75105095
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Adiadochokinesia\n\nAdiadochokinesia, or more fully dysdiadochokinesia, is a neurologic symptom, described as the inability to perform rapid alternating movements, as seen in ataxia.\n\nExamples of this are \n\nA classic test is &quot;the finger to nose test&quot;. The examiner repeatedly taps the patient's finger with the examiner's finger and asks the patient to reciprocate in the same rhythm.\n\nThis should be distinguished from dysdiadochokinesia, a progressive condition involving more or less the same symptoms.\n]" time="0.412"><properties><property name="score" value="0.9471513628959656" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.94715136&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.94715136
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i78]" time="0.193"><properties><property name="score" value="0.9997794032096863" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i79]" time="0.161"><properties><property name="score" value="0.968305766582489" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i80]" time="1.475"><properties><property name="score" value="0.36186615626017254" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i81]" time="1.505"><properties><property name="score" value="0.34135639667510986" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i82]" time="2.064"><properties><property name="score" value="0.907912403345108" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i83]" time="1.291"><properties><property name="score" value="0.913715640703837" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i84]" time="0.344"><properties><property name="score" value="0.735531210899353" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255085 (text: Let's go national! (CBC) The country's first-ever ) is an LLM-generated sample, misclassified as human-generated with confidence 0.73553121&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255085 (text: Let's go national! (CBC) The country's first-ever ) is an LLM-generated sample, misclassified as human-generated with confidence 0.73553121
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Akyem Kotoku\n\nAkyem Kotoku is one of the five traditional districts of the Ashanti Empire. The people are collectively known as Akyem Kotokus.\n\nAkyem Kotoku lies in the Eastern region of Ashanti and is bounded by the East Mamprusi to the East, Nkwantanan to the North, Ashanti to the West and Ahanta West to the South. It is part of Ashanti Region. The area is mainly mountainous with some rich valleys and hills. Akyem Kotoku is known as the &quot;Area of Plenty&quot;.\n\nThe people are mainly farmers and fishermen. Its capital is Akyem Kotoku. It has two paramountcies, Buokrom and Kumasi. Buokrom, the newest paramountcy, was established in 1975. The Akyem Kotoku Traditional Council has nine chiefs.\n\nThe major towns are Abokobi, Domenase, Akropong, and Nsuaem. Other settlements include Aduanful, Boankraful, Denkyemerem, Anhwiaso, Gyankroma, Kwabenya, Nananom, Obofram, Okwensu, Ofoase, Nkoranza, Nsutam, Aduaso, Adom, Ofram, Odomankoma, Asafo, Anyigbe, Obofram, Kwabenya, Denkyemerem, Odumase, Nsutam, Aduaso, Nsuaem, Nsuaso, Tafo-Nsuaem, Manhyia, Boaso, Adom, Ofram, Anhwiaso, Adoase, Anyigbe, Boankraful, Okwensu, Ofoase, Nkoranza, Akropong, Gyankroma, Domenase, Egyir, Afram, Denkyerem, Nananom, Akuapem South, Asafo, Aboadze and Koma.\n\nThe Akyem Kotoku Traditional Council has nine paramountcies.\n\nThe Akyem Kotoku people have two coats of arms, one to use at home and the other for international use.\n\nThe arms used at home is a gold shield with a black anchor in the center. On the shield is a white dove with a golden olive branch in its beak. The dove is supported by two black wild dogs facing each other.\n\nThe shield is surmounted by a gold mural crown with five points and black stripes.\n\nThe shield is supported by two black lions standing on a white scroll. The scroll bears the motto: &quot;MUSTER FAITHFULNESS&quot; (Cape Coast Latin: &quot;Fidelitas Summa Est&quot;).\n\nThe arms used outside the country is a gold shield with a black ram's head in the center. On the shield is a green laurel wreath.\n\nThe shield is surmounted by a gold mural crown with five points and black stripes.\n\nThe shield is supported by two black and white wild dogs standing on a white scroll. The scroll bears the motto: &quot;MUSTER FAITHFULNESS&quot; (Cape Coast Latin: &quot;Fidelitas Summa Est&quot;).\n\nThe two dogs represent vigilance. The two lions in the home arms represent fortitude and strength.\n\n]" time="1.969"><properties><property name="score" value="0.17941135168075562" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Aulus Atilius Calatinus\n\nAulus Atilius Calatinus was consul of the Roman Republic in 221 BC with Gaius Claudius Nero. The primary action of the consulship was the siege and capture of the Volscian city of Suessa Pometia.\n\nIn the same year he was sent to Tarentum with a fleet of 10,000 infantry and 800 cavalry to aid the Tarentines in the Second War with the Bruttians. Calatinus was captured in a naval action off the coast near Tarentum, and the Bruttian admiral Lucerius was rewarded with his command for this victory. The Roman troops under his command, however, escaped to Thurii.\n\nThe earlier Aulus Atilius Calatinus was elected consul in 222 BC, and killed in a battle with the Samnites.\n\n]" time="0.476"><properties><property name="score" value="0.9989840388298035" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Frederic Myers\n\nFrederic William Henry Myers (15 January 1843 \u2013 18 January 1901) was a classical scholar, poet, playwright, novelist, and the founder of the Society for Psychical Research, which he led until his death.\n\nFrederic Myers was born at Kew, Surrey, the eldest son of Frederic Myers, a surgeon. He was educated at King's College School, and at St Paul's School in London. After a year at Trinity College, Cambridge (1861\u20132), he was a schoolmaster at Rugby from 1862 to 1871. In 1872 he became classical lecturer at University College, London. He became a member of the Council of the College (now part of UCL), and was appointed professor of Latin there in 1880. In 1876 Myers joined the Metaphysical Society, along with Alfred Tennyson, J. M. Barrie and others. His friendship with Barrie lasted for the remainder of his life, and Myers even accompanied Barrie and his wife on their honeymoon. Myers and Tennyson became close friends and met frequently in later years, as seen in their correspondence.\n\nMyers wrote prose and verse for periodicals, including a parody of Gerard Manley Hopkins, and some detective stories, and contributed to books of poetry and belles lettres, such as &quot;Whistler and His Art&quot; and the &quot;Yellow Book&quot;. His poetry was praised by Henry James, who]" time="0.790"><properties><property name="score" value="0.9790789484977722" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.97907895&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.97907895
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i85]" time="2.679"><properties><property name="score" value="0.6074527055025101" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i86]" time="3.255"><properties><property name="score" value="0.8846002668142319" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i87]" time="3.088"><properties><property name="score" value="0.23578212261199952" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i88]" time="1.225"><properties><property name="score" value="0.920451819896698" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i89]" time="2.637"><properties><property name="score" value="0.9995527863502502" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i90]" time="0.175"><properties><property name="score" value="0.8738381862640381" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i91]" time="2.092"><properties><property name="score" value="0.1060546487569809" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i92]" time="2.626"><properties><property name="score" value="0.9997923493385314" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i93]" time="0.771"><properties><property name="score" value="0.17623576521873474" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i94]" time="1.713"><properties><property name="score" value="0.9986061652501425" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i95]" time="0.327"><properties><property name="score" value="0.5785720348358154" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i96]" time="2.600"><properties><property name="score" value="0.9411070108413696" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i97]" time="0.189"><properties><property name="score" value="0.9987768530845642" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i98]" time="2.155"><properties><property name="score" value="0.9998217970132828" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i99]" time="1.870"><properties><property name="score" value="0.999772846698761" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i100]" time="2.118"><properties><property name="score" value="0.6086667060852051" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i101]" time="0.177"><properties><property name="score" value="0.5439051389694214" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255102 (text: WHAT WE DO  We are your digital design and photogr) is an LLM-generated sample, misclassified as human-generated with confidence 0.54390514&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255102 (text: WHAT WE DO  We are your digital design and photogr) is an LLM-generated sample, misclassified as human-generated with confidence 0.54390514
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[His father was a professor and his mother was an economist. While growing up, he was exposed to the inspiring stories of his parents and their family members in China.\n\n\n\n\n\nKoh attended the University of Michigan where he received a bachelor of arts degree in economics, then went to the University of California, Berkeley where he earned a master of arts and doctor of philosophy degrees in economics.\n\n\n\n\n\nHe returned to Singapore in 1998 and joined the Singapore Civil Service.\n\n\n\n\n\nAs Deputy Secretary of the Ministry of Finance (2001-2007), he was involved in the design and implementation of the Singapore Savings Bond and the Central Provident Fund (CPF) investments.\n\n\n\n\n\nIn 2008, Koh was appointed as the Managing Director of the Monetary Authority of Singapore (MAS), where he oversees the development and implementation of macroeconomic policies for Singapore.\n\n\n\n\n\nHe is also the Chairman of the Government of Singapore Investment Corporation (GIC) which is responsible for the management of Singapore\u2019s foreign reserves and investment portfolio.]" time="0.692"><properties><property name="score" value="0.9581279754638672" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Little Tikes LTX Battery Operated Tricycle\n\nHe\u2019s been dreaming about his first bike for a while, and now he can hop on and ride around with friends and siblings alike on this bright red Little Tikes LTX Battery Operated Tricycle. It features an adjustable seat, a footrest and a chain guard so he\u2019ll stay safe while riding. Battery-operated, it requires 2 AA batteries (not included). This is a toy for children 3 years and older.]" time="0.339"><properties><property name="score" value="0.999829888343811" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99982989&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99982989
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i102]" time="2.123"><properties><property name="score" value="0.9688578695058823" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i103]" time="0.483"><properties><property name="score" value="0.9989340901374817" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i104]" time="0.194"><properties><property name="score" value="0.7036051750183105" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i105]" time="0.790"><properties><property name="score" value="0.037441521883010864" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i106]" time="2.448"><properties><property name="score" value="0.5476221323013306" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i107]" time="2.468"><properties><property name="score" value="0.9993246595064799" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i108]" time="1.307"><properties><property name="score" value="0.3506947159767151" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i109]" time="1.596"><properties><property name="score" value="0.4575774669647217" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i110]" time="1.447"><properties><property name="score" value="0.29678553342819214" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i111]" time="0.349"><properties><property name="score" value="0.9825385808944702" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i112]" time="2.382"><properties><property name="score" value="0.999797260761261" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i113]" time="1.269"><properties><property name="score" value="0.9991209705670675" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i114]" time="1.216"><properties><property name="score" value="0.9997122883796692" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i115]" time="2.464"><properties><property name="score" value="0.9997557163238525" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i116]" time="1.241"><properties><property name="score" value="0.9998210867245992" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i117]" time="2.276"><properties><property name="score" value="0.5388989567756652" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i118]" time="1.121"><properties><property name="score" value="0.01113554835319519" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i119]" time="0.761"><properties><property name="score" value="0.025812894105911255" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i120]" time="1.210"><properties><property name="score" value="0.98994579911232" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255121 (text: Flash Flood Watch issued October 3 at 3:46PM CDT e) is an LLM-generated sample, misclassified as human-generated with confidence 0.9899458&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255121 (text: Flash Flood Watch issued October 3 at 3:46PM CDT e) is an LLM-generated sample, misclassified as human-generated with confidence 0.9899458
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Trixie Textured - My cat licked my thumb.\n\n\n\nJoin Date: Nov 2006 Posts: 4,592\n\nI have not been to China, though I have been to Japan and love it.\n\n\n\nI have friends in China who have also been to Japan, and they both love it, and neither of them think of it as a &quot;poor&quot; country. In fact, a lot of people who have been there say that China is much nicer.\n\n\n\nChina is not a poor country. They have tons of money and resources. Their problem is that they do not spend enough on infrastructure and don't value education as highly as they could. So their people do not make as much money as they could. And they don't have nice things. It is much nicer than India, but not as good as Japan. I would like to visit it, because I do not think I would be disappointed, but I don't expect it to be better than Japan.\n\n\n\nThey have a different mindset than us, and do things differently. Their government is not a democracy. It is a communist government. I think that is not a bad thing, and has it's benefits. They have much fewer guns and gun crime than we do. They are, for the most part, more honest than we are. Their infrastructure, while not as nice as ours, is very good. They spend more on infrastructure than we do. They have much more public transportation, and better public transportation. They are working on their environmental problems, and they have made more progress than we have. They have fewer social problems, and less crime than we do. They have a very good education system, especially compared to India.\n\n\n\nThey spend more money on schools and education, and invest more in their future, than we do.\n\n\n\nChina is a country that is far more well developed than many people in the US would think.\n\n\n\nThe fact that they have fewer guns, and better education and fewer social problems is probably why they have fewer school shootings than we do. I think we have way too many guns in this country, and that is a big part of the problem.\n\n\n\nJust my thoughts.]" time="1.311"><properties><property name="score" value="0.9997957348823547" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[&quot;Say to them: 'As surely as I live, declares the Sovereign Lord, I take no pleasure in the death of the wicked, but rather that they turn from their ways and live. Turn! Turn from your evil ways! Why will you die, people of Israel?' declares the Sovereign Lord.&quot;\n\nEzekiel 33:11-11\n\nGod wants every sinner to be saved. He does not want people to perish but be saved. God hates sin, and sin destroys a person. But He loves people. He created them to have fellowship with Him and one day live in His Kingdom, and He does not want that for one person to be lost.\n\n\n\n\n\nGod's command to Ezekiel was that He was telling him to go and tell the people of Israel that they need to turn to God, because they were going to die if they did not. We know that Ezekiel told them the same thing, because God's word is truth. God's Word does not fail.\n\n\n\n\n\nWhat God is saying is that people are going to die if they do not turn to Him. People who reject God will die. The Bible says in Romans 1:32, &quot;Though they know God's righteous decree that those who do such things deserve death, they not only continue to do these very things but also approve of those who practice them.&quot; God says that those people who do these things deserve death, and yet people continue to do them and approve of them.\n\n\n\n\n\nGod will not tolerate the sinful acts of humanity, but He wants every person to be saved. There is only one way to be saved. The Bible says in John 14:6, &quot;Jesus answered, 'I am the way and the truth and the life. No one comes to the Father except through me.'\u201d\n\n\n\n\n\nNo one can be saved except through Jesus Christ, but when people put Jesus as their savior, they are saved. It's that simple.\n\n\n\n\n\nThere are many people who call themselves Christians who say that God wants everyone to be saved, and they quote Ezekiel 33:11 and John 3:16, but they don't realize that God's word is truth. God says in Revelation 21:8, &quot;But the cowardly, the unbelieving, the vile, the murderers, the sexually immoral, those who practice magic arts, the idolaters and all liars\u2014their place will be in the fiery lake of burning sulfur. This is the second death.&quot;\n\n\n\n\n\nThere is only one way to be saved, and it is not through anyone or anything but Jesus Christ. Jesus is the only way to heaven. God says that all the people who are going to perish are going to go to a place called the second death. That's hell.\n\n\n\n\n\nWhen people quote Ezekiel 33:11 and John 3:16, they don't realize that God's word is truth. They don't realize that when they say that God wants everyone to be saved, they are not telling the truth. When they say that God wants everyone to be saved, they are putting God in a box, because God's word is truth. God says that those who do not accept Jesus as their Savior are going to perish. God's word is truth, and people who do not put their faith in Jesus Christ are going to die.\n\n\n\n\n\nThe Bible says in John 14:6, &quot;Jesus answered, 'I am the way and the truth and the life. No one comes to the Father except through me.'\u201d\n\n\n\n\n\nJesus is the only way to the Father, and no one can be saved except through Him. He died on the cross for our sins, and He rose from the dead on the third day, and He sits at the right hand of God, interceding for us. It's that simple.\n\n\n\n\n\nLet's take a look at Ezekiel 33:11-11, and read it carefully.\n\n\n\n\n\n&quot;Say to them: 'As surely as I live, declares the Sovereign Lord, I take no pleasure in the death of the wicked, but rather that they turn from their ways and live. Turn! Turn from your evil ways! Why will you die, people of Israel?' declares the Sovereign Lord.&quot;\n\n\n\n\n\nGod says that He takes no pleasure in the death of the wicked. God does not like it when people die. But God hates sin, and sin leads to death. If people want to live, they must turn from their sins and put their faith in Jesus Christ.\n\n\n\n\n\nIf people don't put their faith in Jesus Christ, they are going to perish. God doesn't want that. He doesn't want people to perish, but to live.\n\n\n\n\n\nGod wants people to put their faith in Jesus Christ. If people put their faith in Jesus Christ, they are saved. If they don't put their faith in Jesus Christ, they are going to perish. That's the truth. That's the message of Ezekiel 33:11-11.\n\n\n\n\n\nPeople are going to die if they don't put their faith in Jesus Christ. People will perish. People will be lost. People will be doomed. People will go to hell.\n\n\n\n\n\nLet's read Ezekiel 33:11-12.\n\n\n\n\n\n&quot;For I take no pleasure in the death of anyone,&quot; declares the Sovereign Lord. &quot;Repent and live!&quot;\n\n\n\n\n\nGod takes no pleasure in the death of anyone. God does not want anyone to perish. He doesn't want anyone to die. He wants people to repent and live. God is saying that He is not going to stand back and watch as people perish. He's going to do something about it.\n\n\n\n\n\nGod says in Ezekiel 33:13-16, &quot;Turn! Turn from your evil ways! Why will you die, people of Israel? I have no pleasure in the death of the wicked, but rather that they turn from their ways and live. Do you think I take pleasure in the death of the wicked? declares the Sovereign Lord. Rather, am I not pleased when they turn from their ways and live? Don't presume that you can say to yourselves]" time="3.593"><properties><property name="score" value="0.9998182535171509" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Image caption Bailiffs removed protesters from Grangemouth plant in October 2013\n\nFive people have been charged in connection with an anti-fracking protest at an oil refinery in Grangemouth.\n\nThe protesters climbed on to the chimney at Ineos petrochemicals plant on 9 October.\n\nPolice Scotland said two men, aged 21 and 24, had been charged in connection with the protest and would appear at Falkirk Sheriff Court on Monday.\n\nThree women, aged 40, 44 and 48, were charged with breach of the peace and will also appear at court.\n\nThe 40-year-old was also charged with breach of the peace.\n\nThe other three were arrested in connection with a protest outside a hotel in Glasgow on 10 October.]" time="0.453"><properties><property name="score" value="0.5120607614517212" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.51206076&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.51206076
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i121]" time="2.262"><properties><property name="score" value="0.5016933530569077" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i122]" time="2.723"><properties><property name="score" value="0.2717686891555786" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i123]" time="0.195"><properties><property name="score" value="0.9970172643661499" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i124]" time="0.371"><properties><property name="score" value="0.9854787588119507" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255125 (text: Jackson has advised Clinton for years, and aides s) is an LLM-generated sample, misclassified as human-generated with confidence 0.98547876&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255125 (text: Jackson has advised Clinton for years, and aides s) is an LLM-generated sample, misclassified as human-generated with confidence 0.98547876
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Working with an example - rotating through an array\n\nFor a number of years, I've been fascinated by the idea of writing code that can rotate through an array. Let's say I have the following array of numbers:\n\nvar myNumbers = [ 1, 2, 3, 4, 5 ];\n\nThe following snippet of code would display the same numbers in reverse order:\n\nfor ( var i = myNumbers.length - 1; i &gt;= 0; i-- ) { console.log( myNumbers[i] ); }\n\nWhat's going on here?\n\nWe know that the index of an array can be used as a &quot;pointer&quot; to the location in memory that contains a particular item in an array. If the array myNumbers has five items, then myNumbers[0] , myNumbers[1] , myNumbers[2] , myNumbers[3] and myNumbers[4] will point to the values 1 , 2 , 3 , 4 and 5 respectively.\n\nIn the previous example, the code starts with the length of the array (5) and subtracts one from it. As we know that index 0 in an array contains the first item in the array, the previous code will execute the following steps:\n\nTake the length of the array (5) Subtract 1 (so we now have 4) Start at the index that contains the first item in the array (myNumbers[4] = 3) and run down the array until we reach the last item in the array (myNumbers[3] = 5)\n\nSo the above code will end up displaying the items in the array in reverse order, starting at index 4 and working backwards to index 0.\n\nAt this point, we have a good idea of how we can access the values in an array in a specific order, but how do we go the other way?\n\nIt would be great if we could simply use the same code as above, but add a &quot;--&quot; operator before the last number. That way, we could add a simple parameter to the code to control the range of numbers it will use.\n\nUnfortunately, we can't do that. While JavaScript will happily work with -- , it will also happily take -- in code and convert it to -1 , or the last number in an array. That's not going to do us any good, as we want the code to be flexible enough that we can work with a number that we specify as the range of numbers.\n\nSo let's look at a simple function that can give us the results we need.\n\nfunction rotate(numbers, direction, number) { numbers.reverse(); numbers.push(number); numbers.unshift(number); };\n\nThe rotate function above has three parameters, and will rotate the values in the numbers array.\n\nnumbers is the array we want to use.\n\ndirection is the parameter that allows us to choose whether we want the values to be rotated in the direction we specify. In this case, we've chosen direction to be either &quot;right&quot; or &quot;left&quot;\n\nnumber is the number we're using as the upper bound of the range that we want to use.\n\nLet's take a look at a simple example:\n\nvar numbers = [ 1, 2, 3, 4, 5 ]; rotate( numbers, &quot;right&quot; ); console.log( numbers ); //output: [ 1, 2, 3, 4, 5 ]\n\nAs we can see, the numbers array is simply rotated right by the number we specified. That's fine, but let's add in some code that we can use to control which numbers are displayed:\n\nvar numbers = [ 1, 2, 3, 4, 5 ]; rotate( numbers, &quot;right&quot;, 1 ); console.log( numbers ); //output: [ 1, 4, 3, 5, 2 ]\n\nWe can see that we've rotated the numbers so that the first number in the array is now number 1 , and the next number in the array is 4 .\n\nWe've also specified that we want to rotate in a right-handed manner, by passing &quot;right&quot; as the second parameter. This means that we have access to the following rotation values:\n\nrotate( numbers, &quot;left&quot; ); rotate( numbers, &quot;right&quot; ); rotate( numbers, &quot;right&quot; ); rotate( numbers, &quot;right&quot; ); rotate( numbers, &quot;left&quot; );\n\nIf we try to access the rotate function with &quot;left&quot; as the parameter, we'll see the following output:\n\nvar numbers = [ 1, 2, 3, 4, 5 ]; rotate( numbers, &quot;left&quot; ); console.log( numbers ); //output: [ 4, 3, 5, 2, 1 ]\n\nNot quite what we wanted, right?\n\nWhat we want is to be able to take an array, specify the number of numbers we want to display, and be able to choose the direction of rotation.\n\nWe also want to ensure that the values we are displaying are always within the array's bounds, and we're using them correctly.\n\nThe good news is that it's not too difficult to do this. We simply need to know what we want the code to look like, and write the code to do what we want it to do.\n\nLet's take a look at some simple code to display the numbers in an array in a given direction:\n\nfunction rotateArray( numbers, direction, startAt ) { numbers.reverse(); numbers.push( startAt ); numbers.unshift( startAt ); };\n\nrotateArray function\n\nThis is the code we'll use to display the numbers in the numbers array in the direction we specify. We'll start with a simple parameter:\n\nstartAt is the number we're displaying. It can either be the first number in the array, or any number within the array.\n\nLet's take a look at the code again:\n\nfunction rotateArray( numbers, direction, startAt ) { numbers.reverse(); numbers.push( startAt ); numbers.unshift( startAt ); };\n\nUsing this simple function, we can take any number of numbers, and display them in the direction we want. Let's look at a quick example:\n\nvar numbers = [ 1, 2, 3, 4, 5 ]; rotateArray( numbers, &quot;left&quot;, 3 ); console.log( numbers ); //output: [ 4, 3, 5, 2, 1 ]\n\nThis time, we've used &quot;left&quot; as the direction, and we've also used the parameter &quot;3&quot; as the startAt value. The result is that we now have the numbers displayed in the array starting at the number 3 .\n\nIf we use a number that isn't within the bounds of the array, we'll get the following result:\n\nvar numbers = [ 1, 2, 3, 4, 5 ]; rotateArray( numbers, &quot;right&quot;, 3 ); console.log( numbers ); //output: [ 5, 2, 4, 1, 3 ]\n\nAs you can see, the direction we've specified is &quot;right&quot;, and the value we've specified for startAt is 3 . Unfortunately, this value isn't within the bounds of the array, so the values in the array have been displayed starting from the last number in the array.\n\nLet's take a look at how we can use this code to work with a number that's outside the bounds of the array:\n\nvar numbers = [ 1, 2, 3, 4, 5 ]; rotateArray( numbers, &quot;right&quot;, 6 ); console.log( numbers ); //output: [ 5, 2, 4, 1, 3 ]\n\nThis time, we've specified &quot;right&quot; as the direction, and a value of 6 as the startAt parameter. The result is that we now have the values displayed in the array starting at the number 6 .\n\nThis time, the result we want is displayed in the array, but there's a problem. We don't want the number 6 to be the start of the array. Instead, we want it to be the last number in the array.\n\nThe good news is that we can easily handle this by modifying the code slightly:\n\nfunction rotateArray( numbers, direction, startAt ) { numbers.reverse(); numbers.push( startAt ); numbers.unshift( startAt ); };\n\nUsing the simple rotation code above, we can now take any array, specify the number we want to display, and choose the direction we want to use. This means we can write some simple code that can work with any number, regardless of its position in the array.\n\nThe next step is to write some code that can ensure the number we specify as the startAt is in the range that we expect.\n\nLet's take a look at how we can do that:\n\nfunction rotateArray( numbers, direction, startAt ) { if( startAt &gt;= numbers.length ) { throw new RangeError(&quot;Start at is outside of the]" time="5.567"><properties><property name="score" value="0.995492330619267" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Contents\n\n1. The Meaning of Love\n\n2. Our Attitude Toward Others\n\n3. The Two Great Commandments\n\n4. Developing Love\n\n5. How to Help Others\n\n6. The Golden Rule\n\n7. The Commandments in the New Testament\n\n8. Tolerance\n\n\nCHAPTER 1\n\nThe Meaning of Love\n\nBefore we can love others, we must know the meaning of love. It is important to love God and obey Him and to have a right relationship with Him. But to have a right relationship with Him we must love Him. That means that we must have a good relationship with all His laws.\n\nOf course we can't keep all the commandments all the time. Yet we must have a good relationship with them. Otherwise we will not know what God is like.\n\nLove is based on God's law.\n\nRomans 13:10. The commandments, You shall not commit adultery, You shall not steal, You shall not bear false witness, You shall not covet; and any other commandment, are summed up in this word, You shall love your neighbor as yourself.\n\nGalatians 5:14. For all the law is fulfilled in one word, even in this: You shall love your neighbor as yourself.\n\nLove is based on God's law. We should study God's commandments and know the laws He has given us. Then we will know how to have a good relationship with Him.\n\nOne reason we are to love God and obey Him is that it is important to have a good relationship with God. We need to know how much God loves us. Without knowing this we cannot have a right relationship with God. We need to have a good relationship with Him so we can be happy and serve Him better.\n\nLet us understand what love is.\n\n1 John 4:8,9. He who does not love does not know God, for God is love. In this the love of God was manifested toward us, that God has sent His only begotten Son into the world, that we might live through Him.\n\nIn these verses John explained that love is the opposite of lawlessness. Love is the opposite of murder, hatred, malice, evil, adultery, stealing, lying, and a lot of other bad things.\n\nRomans 13:8-10. Owe no man anything, but to love one another: for he that loves another has fulfilled the law. For this, You shall not commit adultery, You shall not kill, You shall not steal, You shall not covet, and if there be any other commandment, it is briefly comprehended in this saying, namely, You shall love your neighbor as yourself. Love does no harm to a neighbor; therefore love is the fulfillment of the law.\n\nWe learn from this that the opposite of love is lawlessness. The opposite of love is murder, hatred, envy, malice, evil, and all the other things mentioned in these verses.\n\nVerse 10 says, &quot;Love does no harm to a neighbor.&quot; In other words, love is a good thing. It is a good thing to have a good relationship with God.\n\nRomans 13:8 says, &quot;Love does no harm to a neighbor.&quot; Love helps and blesses others.\n\nLove is a good thing. The reason love is a good thing is because God is love.\n\n1 John 4:16. And we have known and believed the love that God has for us. God is love; and he who abides in love abides in God, and God in him.\n\nGod is love.\n\n1 John 4:7-11. Beloved, let us love one another: for love is of God; and every one that loves is born of God, and knows God. He that loves not knows not God; for God is love. In this was manifested the love of God toward us, because that God sent his only begotten Son into the world, that we might live through him. Herein is love, not that we loved God, but that he loved us, and sent his Son to be the propitiation for our sins. Beloved, if God so loved us, we ought also to love one another. No man has seen God at any time. The only begotten Son, which is in the bosom of the Father, he has declared him.\n\nVerse 8 says, &quot;God is love.&quot; God's name is Love.\n\nGod's name is Love.\n\nJohn 4:8. \u2026 the well was deep; and I said, &quot;How shall I get down?&quot;\n\nAnd he said, &quot;Come, and I will show you.&quot;\n\n9. And he said, &quot;Draw you [a] little water out of the well, and \u2026 drink.&quot;\n\n10. And I said, &quot;Sir, give me [a] little to drink.&quot;\n\n11. And he said, &quot;Drink, my [b] son, and \u2026 your eyes shall be opened.&quot;\n\n12. And I drank, and it was as if I had had new vision given to me; for I could see [c] through those walls of water.\n\n13. And he said to me, &quot;Go [a] down, for the water is [b] yet a little way below the earth. \u2026\n\n15. \u2026 This water is [c] that which I said \u2026 will make you come up hither.&quot;\n\n16. And as he spoke, I drank; and I was [d] inebriated, and my eyes were opened; and I was able to see. \u2026\n\n19. \u2026 &quot;Come,&quot; he said, &quot;follow me \u2026 and I will give you \u2026 rest.&quot;\n\nGod is the God of rest.\n\n1 Thessalonians 4:9. \u2026 God has not appointed us to wrath, but to obtain salvation by our Lord Jesus Christ.\n\nGod is our salvation. God is our rest. God is our peace.\n\nGod is our rest.\n\n1 Peter 2:6. \u2026 you may be \u2026 a kind of first-fruits of his creatures.\n\nGod is our first fruits.\n\n2 Timothy 2:13. If we are faithless, he remains faithful; for he cannot deny himself.\n\nGod is faithful.\n\nGod is the God of rest.\n\nGod is our salvation.\n\nGod is our peace.\n\nGod is our rest.\n\nHebrews 4:10-13. For he who has entered His rest has himself also ceased from his own works, as God did from His. Let us therefore be diligent to enter that rest \u2026\n\n12. \u2026 For the word of God is living and active, and sharper than any two-edged sword, and piercing even to the division of soul and spirit, and of joints and marrow, and is a discerner of the thoughts and intents of the heart.\n\n13. \u2026 neither is there any creature that is not manifest in his sight; but all things are naked and opened to the eyes of him with whom we have to do.\n\nGod is a discerner of the thoughts and intents of the heart.\n\nGod sees everything.\n\nGod sees everything.\n\n1 John 4:8,9. He who does not love does not know God, for God is love. In this the love of God was manifested toward us, that God has sent His only begotten Son into the world, that we might live through Him.\n\nGod is love.\n\n1 John 3:16,18. Hereby perceive we the love of God, because he laid down his life for us: and we ought to lay down our lives for the brethren.\n\n19. But whoso has this world's goods, and sees his brother in need, and shuts up his heart from him, how does the love of God abide in him?\n\nGod is love.\n\nGod is our rest.\n\nGod is our salvation.\n\nGod is our peace.\n\nGod is a discerner of the thoughts and intents of the heart.\n\nGod is love.\n\nGod is our rest.\n\nGod is our salvation.\n\nGod is our peace.\n\nGod is a discerner of the thoughts and intents of the heart.\n\nThe Bible says that God is love.\n\nWe can study love.\n\n1 John 4:7-11. Beloved, let us love one another: for love is of God; and every one that loves is born of God, and knows God. He that loves not knows not God; for God is love. In this was manifested the love of God toward us, because that God sent his only begotten Son into the world, that we might live through him. Herein is love, not that we loved God, but that he loved us, and sent his Son to be the propitiation for our sins. Beloved, if God so loved us, we ought also to love one another. No man has seen God at any time. The only begotten Son, which is in the bosom of the Father, he has declared him.\n\nVerse 11 says, &quot;Herein is love, not that we loved God, but that he loved us.&quot; The opposite of love is lawlessness. But the opposite of love is not lawlessness. The opposite of love is law.\n\nLoving God and obeying Him is loving Him.\n\nThe Bible says that God is love.\n\nLoving God and obeying Him is loving Him.\n\nLoving God and obeying Him is loving God.\n\nJohn 14:21. He that has my commandments,]" time="6.226"><properties><property name="score" value="0.5310911163687706" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Interviews\n\nInterview with Ariel Efraim (Soilwork)\n\nInterview with Steven Rosen (August Burns Red)\n\nInterview with Brent Hinds (Mastodon)\n\nInterview with Paul Kuhr (I See Stars)\n\nInterview with Christian Antidormi (Hatebreed)\n\nInterview with Jack Syperek (Structures)\n\nInterview with Ben Gordon (The Treatment)\n\nInterview with Doc Coyle (God Forbid)\n\nInterview with Chris Barnes (Six Feet Under)\n\nInterview with Mike D (The Sword)\n\nInterview with Russ Rankin (The Slackers)\n\nInterview with Claudio Sanchez (Coheed and Cambria)\n\nInterview with Udo Dirkschneider (Accept)\n\nInterview with Ben Weinman (Dillinger Escape Plan)\n\nInterview with Matt Pike (High On Fire)\n\nInterview with Joe Duplantier (Gojira)\n\nInterview with Johannes Eckerstr\xf6m (Evergrey)\n\nInterview with Pete Loose (Reel Big Fish)\n\nInterview with Pete Wentz (Fall Out Boy)\n\nInterview with Daniel Svensson (Raubtier)\n\nInterview with Chino Moreno (Deftones)\n\nInterview with Pierre Bouvier (Simple Plan)\n\nInterview with Jon Egan (The Ghost Of A Thousand)\n\nInterview with Mike D (Lack Of Afro)\n\nInterview with Michael Crafter (Sevendust)\n\nInterview with Charlie Benante (Anthrax)\n\nInterview with Rob Caggiano (Volbeat)\n\nInterview with Ronnie Atkins (Pretty Maids)\n\nInterview with Gee Anzalone (Exhorder)\n\nInterview with Micky Shirley (Five Finger Death Punch)\n\nInterview with Oderus Urungus (GWAR)\n\nInterview with Ryan Downey (Texas Hippie Coalition)\n\nInterview with Chris Chasse (Unearth)\n\nInterview with Lee Altus (Suffocation)\n\nInterview with Roy Mayorga (Soulfly)\n\nInterview with Gus G (Firewind)\n\nInterview with Paul Allender (In Flames)\n\nInterview with Francesco Artusato (All Shall Perish)\n\nInterview with Rody Walker (Pyogenesis)\n\nInterview with Mike Dean (Slipknot)\n\nInterview with Angela Gossow (Arch Enemy)\n\nInterview with Trevor Strnad (The Black Dahlia Murder)\n\nInterview with Mike Peters (The Alarm)\n\nInterview with Jonas Ekdahl (Freak Kitchen)\n\nInterview with Dave Ellefson (Megadeth)\n\nInterview with Corey Taylor (Slipknot)\n\nInterview with Keith Buckley (Every Time I Die)\n\nInterview with Jack Frost (Kill Devil Hill)\n\nInterview with Josh Rand (Temple of Brutality)\n\nInterview with Michael Amott (Arch Enemy)\n\nInterview with Scott Gorham (Thin Lizzy)\n\nInterview with Matt Heafy (Trivium)\n\nInterview with Joe Kresge (Unearth)\n\nInterview with Randy Blythe (Lamb of God)\n\nInterview with Benji Webbe (Dimmu Borgir)\n\nInterview with Max Cavalera (Soulfly)\n\nInterview with John Bush (Armored Saint)\n\nInterview with Davey Muise (Exhorder)\n\nInterview with Stephen Carpenter (Deftones)\n\nInterview with Dani Filth (Cradle Of Filth)\n\nInterview with Leif Cuzner (Darkane)\n\nInterview with George \u201cCorspegrinder\u201d Fisher (Cannibal Corpse)\n\nInterview with Steve Asheim (Hatebreed)\n\nInterview with Bobby Blitz Ellsworth (Overkill)\n\nInterview with Adam Dutkiewicz (Killswitch Engage)\n\nInterview with Mick Murphy (Exhumed)\n\nInterview with Doug Pinnick (King\u2019s X)\n\nInterview with Alexi Laiho (Children Of Bodom)\n\nInterview with M. Shadows (Avenged Sevenfold)\n\nInterview with Alex Holzwarth (Rhapsody Of Fire)\n\nInterview with Carlos Cruz (Soulfly)\n\nInterview with Glenn Hughes (Deep Purple/Black Country Communion)\n\nInterview with Oliver Palotai (Kamelot)\n\nInterview with Meegs Rascon (The Mars Volta)\n\nInterview with Joakim Brod\xe9n (Sabaton)\n\nInterview with Tim \u201cRipper\u201d Owens (Iced Earth/Yngwie Malmsteen)\n\nInterview with Matt Heafy (Trivium)\n\nInterview with Warrel Dane (Nevermore)\n\nInterview with Zach Myers (Shinedown)\n\nInterview with Michael Poulsen (Volbeat)\n\nInterview with Michael Amott (Arch Enemy)\n\nInterview with Mat Sinner (Primal Fear)\n\nInterview with Bj\xf6rn \u201cSpeed\u201d Strid (Soilwork)\n\nInterview with Matt Heafy (Trivium)\n\nInterview with Rob Flynn (Machine Head)\n\nInterview with Serj Tankian (System Of A Down)\n\nInterview with Ville Valo (HIM)\n\nInterview with Jimmy DeGrasso (Megadeth)\n\nInterview with Joey Jordison (Slipknot)\n\nInterview with Chris Jericho (Fozzy/Fozzy Osbourne)\n\nInterview with Jon Donais (Shadows Fall)\n\nInterview with Jason Hook (Five Finger Death Punch)\n\nInterview with Steve Mazur (Grave Digger)\n\nInterview with Bj\xf6rn \u201cSpeed\u201d Strid (Soilwork)\n\nInterview with Sonny Sandoval (P.O.D.)\n\nInterview with Serj Tankian (System Of A Down)\n\nInterview with Fredrik Thordendal (Meshuggah)\n\nInterview with Wes Borland (Limp Bizkit)\n\nInterview with Alexi Laiho (Children Of Bodom)\n\nInterview with Blaze Bayley (Iron Maiden)\n\nInterview with Mark Hunter (Chimaira)\n\nInterview with Max Cavalera (Soulfly/Sepultura)\n\nInterview with Dave Lombardo (Slayer)\n\nInterview with Bruce Dickinson (Iron Maiden)\n\nInterview with Morgan Lander (Fear Factory)\n\nInterview with Oliver Palotai (Kamelot)\n\nInterview with James Hetfield (Metallica)\n\nInterview with Al Jourgensen (Ministry)\n\nInterview with Corey Taylor (Stone Sour/Slipknot)\n\nInterview with Alex Skolnick (Testament)\n\nInterview with Charlie Benante (Anthrax)\n\nInterview with Chris Reifert (Autopsy)\n\nInterview with Lajon Witherspoon (Sevendust)\n\nInterview with Jason McMaster (Grave Robber)\n\nInterview with Jon Howard (Light The Torch)\n\nInterview with Neil Fallon (Clutch)\n\nInterview with Dino Cazares (Fear Factory)\n\nInterview with Oderus Urungus (GWAR)\n\nInterview with Rex Brown (Pantera)\n\nInterview with Dave Lombardo (Slayer)\n\nInterview with Serj Tankian (System Of A Down)\n\nInterview with John Moyer (Disturbed)\n\nInterview with Mike Portnoy (Adrenaline Mob)\n\nInterview with King Diamond\n\nInterview with Trevor Strnad (The Black Dahlia Murder)\n\nInterview with Mikael \xc5kerfeldt (Opeth)\n\nInterview with Scott Ian (Anthrax)\n\nInterview with Paul Gray (Slipknot)\n\nInterview with Casey Chaos (Amen)\n\nInterview with Opeth\u2019s Mikael \xc5kerfeldt\n\nInterview with Mark Morton (Lamb Of God)\n\nInterview with Shawn Drover (Megadeth)\n\nInterview with Chino Moreno (Deftones)\n\nInterview with Serj Tankian (System Of A Down)\n\nInterview with Bruce Corbitt (Corrosion Of Conformity)\n\nInterview with Ray Alder (Fates Warning)\n\nInterview with Rob \u201cBlasko\u201d Nicholson (Ozzy Osbourne)\n\nInterview with Tom Araya (Slayer)\n\nInterview with Max Cavalera (Soulfly)\n\nInterview with Ryan Martinie (Cynic)\n\nInterview with Max Cavalera (Soulfly)\n\nInterview with Jason Newsted (Voivod)\n\nInterview with Mick Murphy (Exhumed)\n\nInterview with Lars Frederiksen (Rancid)\n\nInterview with Mike Muir (Suicidal Tendencies)\n\nInterview with Dave Mustaine (Megadeth)\n\nInterview with Neil Fallon (Clutch)\n\nInterview with Bruce Dickinson (Iron Maiden)\n\nInterview with Maynard James Keenan (Tool)\n\nInterview with Mike \u201cRat\u201d Ratner (S.O.D.)\n\nInterview with Max Cavalera (Soulfly)\n\nInterview with Tom Araya (Slayer)\n\nInterview with Chuck Billy (Testament)\n\nInterview with Tom Araya (Slayer)\n\nInterview with Phil Anselmo (Down)\n\nInterview with King Diamond\n\nInterview with Corey Taylor (Stone Sour/Slipknot)\n\nInterview with Dave Mustaine (Megadeth)\n\nInterview with Lars Ulrich (Metallica)\n\nInterview with Jerry Cantrell]" time="5.283"><properties><property name="score" value="0.09577819279261998" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Apple\u2019s ResearchKit is off to a good start, but there\u2019s plenty of work to be done.\n\n\n\nResearchKit was launched in March as a way for Apple\u2019s developers to easily tap into medical data from the iPhone to help researchers track and collect information about a variety of health conditions. Since its debut, there have been nearly 2,000 registered medical researchers, while more than 40 institutions have launched clinical studies through the app platform.\n\n\n\nBut there are several hurdles for Apple and its partners. On Wednesday, Dr. Harvey Fineberg, president of the Institute of Medicine, will join at a panel discussion at Mount Sinai Hospital in New York to examine what works and what\u2019s missing in the fast-evolving field of mobile health.\n\n\n\nThe session, titled \u201cWhat Are the Future Directions for ResearchKit?\u201d comes after a first wave of successful projects on ResearchKit, including apps that help people manage diabetes, heart disease and asthma.\n\n\n\nThere are also some examples of the problems Apple is facing as it tries to get more partners to adopt the tool. Earlier this month, a major player in the diabetes community said it would not be taking part in the app\u2019s data collection. The Juvenile Diabetes Research Foundation said in a statement that its members were not interested in the iPhone platform and \u201cwould prefer to conduct research studies using our own patient data that is gathered on an Android platform.\u201d\n\n\n\nThat\u2019s likely an expensive prospect for the organization, given the popularity of iPhones in the United States. As of last year, Android devices accounted for just over half of all U.S. smartphone subscribers, while iOS devices made up around 41 percent. That\u2019s according to the Pew Research Center, which estimated last year that there were 121 million iPhone users in the United States.\n\n\n\nThe success of Apple\u2019s platform will hinge in part on how it handles data privacy and security concerns. Apple had to clear some hurdles in the early days of ResearchKit because many people were concerned about privacy when they heard it would rely on data from Apple\u2019s servers. Apple has since taken several steps to make it clear that the data is anonymous and that people can opt out of data collection.\n\n\n\nThe app program could also be a major part of the iPhone\u2019s future if it takes off. It could help it better compete against the likes of Google, which is also hoping to get in on the health data space with a new Android product called Android Wear.\n\n\n\nThe first public version of ResearchKit will be unveiled on Wednesday, the panel will also include representatives from Mount Sinai and Pfizer.\n\n\n\nThat event will be at 11:30 a.m. EDT in New York City, and it will be livestreamed on Fortune\u2019s website.]" time="1.779"><properties><property name="score" value="0.4640413721402486" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[\n\nSend this thread via email by filling out the fields below\n\n\n\nThread: Lost 10lbs in the last 3 weeks Send To:\n\nUse semi-colon to separate multiple addresses Your Name: Your Email: Add a personal message: Send this thread via email by filling out the fields below\n\n\n\n]" time="0.220"><properties><property name="score" value="0.8082296252250671" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[I don't think it is your attitude. I think it is just the way you are.\n\nIt isn't a criticism, just a statement of fact. It's like saying I don't think you are stupid.\n\nJust that you are naive, sweet, guileless, open. You have so much going for you, you don't have to be cynical.\n\nBut it isn't fair to the rest of us, not fair to expect us to feel your innocence and not to respond to it.\n\nYou need to toughen up, otherwise you will just be hurt. But don't ever be cynical.]" time="0.324"><properties><property name="score" value="0.9995108842849731" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Get the biggest daily news stories by email Subscribe Thank you for subscribing We have more newsletters Show me See our privacy notice Could not subscribe, try again later Invalid Email\n\nAs the Queen's cousin, Lady Amelia Windsor is more used to socialising with the royal family than you or I.\n\nBut even for her, it seems the sight of seeing Kate Middleton and Prince William in attendance at Pippa Middleton's wedding - as well as the newlyweds' own big day - was enough to make her feel &quot;uncomfortable&quot;.\n\nSpeaking at the wedding of one of her friend's brothers in Kent, the 28-year-old spoke about how it made her feel to have them there, reports the Daily Mail .\n\nShe said: &quot;Well I think it is a beautiful wedding, obviously, because it's Pippa and James.\n\n&quot;I've never been to a wedding before, so I don't know really what it should be like. But I do find it quite uncomfortable seeing them.&quot;\n\nThe Royal family came together to celebrate the wedding of Kate's sister to hedge fund manager James Matthews in the Berkshire village of Englefield.\n\n(Image: Getty Images Europe)\n\n(Image: PA)\n\n(Image: PA)\n\nAnd while Pippa and James' wedding was a private affair with just 200 guests, they were still surrounded by some royals as well as other notable names.\n\nLady Amelia's father is the Queen's first cousin, so it was no surprise to see the Queen herself attend the wedding, alongside Prince Harry and Duchess Meghan.\n\nThe Princesses Beatrice and Eugenie were also there, as well as Kate's sister, Pippa's brother James Middleton, David and Victoria Beckham and Sir Elton John.\n\nLady Amelia revealed that she even had a few words with Harry and Meghan, with the former asking her how she felt about her cousin's big day.\n\nShe said: &quot;I had a little conversation with Harry and Meghan, and they said how lovely the wedding was. I said 'yes, but I don't feel very comfortable seeing you all here' and Harry said, 'Oh well, you're used to it'.\n\n(Image: PA)\n\n(Image: PA)\n\n&quot;I said, 'Yes, but it is still very strange to see you here'. I suppose you can't blame me, I am a bit naive. But the whole thing was really lovely.&quot;\n\nShe added: &quot;I mean, James is my cousin, I grew up with him and I love him dearly, so it was really lovely to see him. It was a lovely wedding, they are such a lovely couple. I am so pleased they are so happy.&quot;\n\n(Image: Getty Images Europe)\n\n(Image: Getty Images Europe)\n\nAnd as for what she thought of her father's outfit - which caused much comment among fans - she admitted she wasn't sure.\n\nShe said: &quot;Well, he always wears an odd choice of clothes, but that is what he likes. It wasn't bad, I mean he didn't look like a big prawn.&quot;\n\nShe added: &quot;It was lovely to see him happy.&quot;]" time="1.929"><properties><property name="score" value="0.9994182189305624" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Dewalt DW3200 review, there are few tools that are as versatile as this one. You may have a thousand bucks of cordless tools, but there\u2019s nothing that will do as much as this one will do. As you might have noticed, the construction of this unit is incredibly sturdy, so that you don\u2019t have to worry about it breaking when you use it. Also, it is a bit heavy, and it will take a while to get used to using it, especially if you have not used any other heavy drills.\n\nPower\n\nWe also tested the power of this Dewalt DW3200 and we must say that it is very impressive. It delivers 3,300 RPM, which is more than enough to bore holes and drill in all kinds of materials. It can easily penetrate into wood, brick, or concrete and it can bore holes up to two inches. The trigger lock mechanism also works well, which makes this drill safer to use.\n\nCordless or corded?\n\nIt is important to remember that this drill can work either as a corded or a cordless unit, and this is very important for you to know. This Dewalt DW3200 uses the 20 volt lithium-ion batteries and they can last longer than you might expect. They are also very light and durable, so you will be able to use them for a long time. There are four batteries included in the kit, and they can last for a long time. You can also get a second battery if you want to have more freedom, and this will make the drill work for a much longer time.\n\nExtra Features\n\nYou may also like the extra features included in this Dewalt DW3200. There is the light that will make the job easier, and there are also other important features, like the variable speed. This is also a very durable drill, so it will last for a very long time. Also, the build quality is fantastic, so you don\u2019t have to worry about it breaking easily.\n\nOverall\n\nIf you have a lot of different drilling needs, this is the drill for you. This drill is perfect for the casual user, but it can also be used by professionals. It is also very durable, so you can use it for a very long time, and you will get the most for your money. You should also remember that this drill is very powerful, and it can work for a long time, without breaking down.\n\nIf you want to see what more we have, you can visit Amazon.com and check all of the available drills. You can also check our article on the best reciprocating saw, to see what the best product in this category is. Also, don\u2019t forget to like and share this article, if you liked it.]" time="1.788"><properties><property name="score" value="0.9771516521771749" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Source: Food\n\nAs I\u2019m sure you\u2019re well aware by now, it\u2019s officially National Grilled Cheese Sandwich Month. With that in mind, this week\u2019s Food Week: Hors D\u2019Oeuvre features a variety of hors d\u2019oeuvres you can make in your own home. After all, a good grilled cheese is just like a fine wine: it gets better with age.\n\nContinue reading]" time="0.375"><properties><property name="score" value="0.8901404738426208" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Jack and Justin are two of the hottest and most popular guys on the block. They make the perfect couple and know exactly what each other likes. But when they meet Tyler, things get complicated. This sizzling trio enjoys their time together and both Tyler and Jack crave more. But Justin is not so easily convinced. How will this all work out?\n\nWarning: This story contains three scorching hot guys and lots of hot and hard action.\n\n\n\n\xa92015 Scarlet Rose (P)2017 Scarlet Rose]" time="0.372"><properties><property name="score" value="0.9998045563697815" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[LAS VEGAS, NV - JANUARY 07: Recording artist Demi Lovato performs onstage during the 2017 DIRECTV NOW Super Saturday Night Concert at Park Theater at Park MGM on January 7, 2017 in Las Vegas, Nevada. (Photo by Kevin Winter/Getty Images for DIRECTV) (Photo: Kevin Winter, 2017 Getty Images)\n\nLAS VEGAS - Former Disney star Demi Lovato will release a new single from her sixth studio album on June 15th.\n\nThat's according to a video of Lovato, 26, posted on the YouTube channel of Las Vegas radio station KVEG.\n\nThe new album is the follow-up to 2015's &quot;Confident,&quot; which features the top 10 hits &quot;Cool for the Summer&quot; and &quot;Confident.&quot;\n\nMore: Everything you need to know about this year's 2017 Route 91 Harvest festival\n\nMore: Demi Lovato set for national tour with DJ Khaled\n\nMore: Demi Lovato re-engages with ex-boyfriend Wilmer Valderrama, so could there be more?\n\nThe pop singer also revealed she is launching a partnership with American Airlines to celebrate Hispanic Heritage Month, and she plans to go on a national tour with DJ Khaled.\n\nDemi Lovato to Release New Single, Announce Album Title on June 15th https://t.co/OBkKbXWySZ \u2014 Pop Crave (@PopCrave) June 1, 2017\n\nRead or Share this story: https://usat.ly/2rVY4mA]" time="0.985"><properties><property name="score" value="0.8391260802745819" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[RACISM IN BRITAIN HAS BECOME SO entrenched that it is becoming a bigger problem than it was at the time of the Birmingham pub bombings, when the threat of violent nationalism in the late 1970s prompted the first official inquiry into racial attacks, according to a report published today by the Equality and Human Rights Commission (EHRC).\n\nThe report, We\u2019re still struggling: the legacy of the Stephen Lawrence inquiry, says that despite years of measures to increase the representation of ethnic minorities in public life and protect them from discrimination, they are still \u201con the receiving end of racism\u201d, with black people twice as likely to suffer police stop and search tactics as whites.\n\nFigures from the Crown Prosecution Service (CPS) show that more than 40 per cent of cases involving charges of racially aggravated harassment are dropped, with charges reduced to more minor ones, or not pursued. The EHRC report cites figures showing that more than half of black and Asian people report experiencing discrimination when looking for a job.\n\nIn October, the Telegraph reported how]" time="0.652"><properties><property name="score" value="0.8731753528118134" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[# This Source Code Form is subject to the terms of the Mozilla Public\n\n# License, v. 2.0. If a copy of the MPL was not distributed with this\n\n# file, You can obtain one at http://mozilla.org/MPL/2.0/.\n\n# The following comment must be at the start of a new line.\n\n#\n\n#\n\n# The contents of this file are subject to the Mozilla Public\n\n# License Version 1.1 (the &quot;License&quot;); you may not use this file\n\n# except in compliance with the License. You may obtain a copy of\n\n# the License at http://www.mozilla.org/MPL/\n\n# Software distributed under the License is distributed on an &quot;AS\n\n# IS&quot; basis, WITHOUT WARRANTY OF ANY KIND, either express or implied.\n\n# See the License for the specific language governing rights and\n\n# limitations under the License.\n\n# The Original Code is Mozilla Communicator client code, released\n\n# March 31, 1998.\n\n# The Initial Developer of the Original Code is\n\n# Netscape Communications Corporation.\n\n# Portions created by the Initial Developer are Copyright (C) 1998\n\n# the Initial Developer. All Rights Reserved.\n\n# Contributor(s):\n\n#\n\n# Alternatively, the contents of this file may be used under the terms of\n\n# either of the GNU General Public License Version 2 or later (the &quot;GPL&quot;),\n\n# or the GNU Lesser General Public License Version 2.1 or later (the &quot;LGPL&quot;),\n\n# in which case the provisions of the GPL or the LGPL are applicable instead\n\n# of those above. If you wish to allow use of your version of this file only\n\n# under the terms of either the GPL or the LGPL, and not to allow others to\n\n# use your version of this file under the terms of the MPL, indicate your\n\n# decision by deleting the provisions above and replace them with the notice\n\n# and other provisions required by the GPL or the LGPL. If you do not delete\n\n# the provisions above, a recipient may use your version of this file under\n\n# the terms of any one of the MPL, the GPL or the LGPL.\n\n#\n\n# ***** END LICENSE BLOCK *****\n\nMOCHITEST_MANIFESTS += [\n\n' xpcshell.ini ' ,\n\n' mochitest-glue.ini ' ,\n\n' mochitest-plain.ini ' ,\n\n]\n\nDIST_DIRS += [\n\n' external/ ' ,\n\n' ipc/ ' ,\n\n' modules/ ' ,\n\n]\n\nEXTRA_COMPONENTS += [\n\n' mozLoop.js ' ,\n\n' webextensions.js ' ,\n\n]\n\ndeps = [\n\n'../base ' ,\n\n'../browser/base ' ,\n\n'../browser/components/migration ' ,\n\n'../browser/components/places ' ,\n\n'../browser/themes/shared ' ,\n\n'../content/base ' ,\n\n'../content/html ' ,\n\n'../content/xul ' ,\n\n'../embedding/browser ' ,\n\n'../embedding/components/browser ' ,\n\n'../embedding/components/toolkit ' ,\n\n'../extensions/browser/api ' ,\n\n'../extensions/common/moz.build ' ,\n\n'../gfx/2d ' ,\n\n'../gfx/layers ' ,\n\n'../intl ' ,\n\n'../netwerk ' ,\n\n'../printing ' ,\n\n'../services/sync ' ,\n\n'../toolkit/components/global ' ,\n\n'../toolkit/components/printing ' ,\n\n'../toolkit/components/satchel ' ,\n\n'../toolkit/crashreporter ' ,\n\n'../toolkit/library ' ,\n\n'../toolkit/mozapps/extensions ' ,\n\n'../toolkit/mozapps/extensions/extensions-types.js ' ,\n\n'../xpcom/base ' ,\n\n'../xpcom/build ' ,\n\n'../xpcom/build/shared-nspr ' ,\n\n'../xpcom/build/shared-nspr/src ' ,\n\n'../xpcom/ds ' ,\n\n'../xpcom/ds/io ' ,\n\n'../xpcom/ds/xpcom ' ,\n\n'../xpcom/glue ' ,\n\n'../xpcom/io ' ,\n\n'../xpcom/threads ' ,\n\n'../xpcom/threads/src ' ,\n\n'../xpcom/threads/xpcom ' ,\n\n' ../xpfe/appshell/nsAppShell.cpp ' ,\n\n' ../xpfe/appshell/nsAppShell.h ' ,\n\n' ../xpfe/appshell/nsAppsShellService.cpp ' ,\n\n' ../xpfe/appshell/nsAppsShellService.h ' ,\n\n' ../xpfe/appshell/nsNativeAppSupportWin.cpp ' ,\n\n' ../xpfe/appshell/nsNativeAppSupportWin.h ' ,\n\n' ../xpfe/appshell/nsNativeAppSupportXPConnect.cpp ' ,\n\n' ../xpfe/appshell/nsNativeAppSupportXPConnect.h ' ,\n\n' ../xpfe/appshell/nsWindowRoot.cpp ' ,\n\n' ../xpfe/appshell/nsWindowRoot.h ' ,\n\n' ../xpfe/appshell/nsXULAppAPI.h ' ,]" time="3.540"><properties><property name="score" value="0.03840629756450653" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The world of professional wrestling is always changing, but the superstars themselves can sometimes get lost in the shuffle. Sometimes, the changes within the industry are so drastic that a superstar's popularity can get left in the past. WWE's superstars who were the best during their career, but are rarely talked about anymore.\n\nWWE is still the leader of the wrestling world. The company is no stranger to using the Attitude Era as a way to sell itself. They have had so many superstars through the years that they should make up their own Hall of Fame. Some of those superstars are the ones who are mentioned the most.\n\nThe casual fan knows about Hulk Hogan, Stone Cold, and The Rock. That is what makes this so fun. Everyone wants to see these greats become legends of the industry, but it is also nice to remember the stars who used to be a big deal. Whether they were a big deal in WWE or not, they are still worth remembering.\n\nContinue scrolling to keep reading Click the button below to start this article in quick view Share Tweet Email Copy Link Copied\n\n10 Sin Cara\n\nvia trbimg.com\n\nSin Cara has become a joke of a superstar. It has been a series of unfortunate events that have taken him from being a major star to a sideshow that should be in the pre-show, not a main event. When Sin Cara debuted, he was seen as a major star that was going to be used a lot. It took some time, but it finally happened.\n\nThe problem was that he never delivered.\n\nHe was so unimpressive that he was booted from WWE in 2016. He is making an appearance at WrestleMania 33. For his sake, it would be best if he kept quiet and didn't interact with the crowd.\n\n9 Kelly Kelly\n\nvia tumblr.com\n\nKelly Kelly was a fan favorite during her time in WWE. She made her debut as a popular model who was wrestling on ECW. From there, she rose to the top as a valuable part of the roster. Kelly became a popular figure who had her own dedicated fans. She had the right amount of talent to be a decent worker, but it was her good looks that gave her the extra push.\n\nAfter leaving WWE in 2012, Kelly kept wrestling. She is back on the roster with TNA, but she has not appeared on a television episode of Impact Wrestling since December of 2015.\n\n8 Alberto Del Rio\n\nvia wrestlingrumors.net\n\nIt is a sad truth that Alberto Del Rio has fallen from his former position in the industry. He was a very big part of WWE. He was not only the United States Champion, but also a World Champion. Alberto had a unique presence as a superstar that could do things no one else could do. He also had the microphone skills to back up his moves.\n\nHis tenure in WWE was marred with incidents that made it clear he was not going to be there much longer. His time in Lucha Underground was more memorable, but it was a step down from what he used to be.\n\n7 The Great Khali\n\nvia cagesideseats.com\n\nThe Great Khali is not really a superstar that anyone really misses. His presence was laughable at times. The thing about Khali is that he had a lot of potential. He was very big and very strong, but he never seemed to have any clue about how to use those traits to his advantage.\n\nKhali is now an actor and a former wrestler. He is part of the reality show Khali Kisses, which involves him doing a lot of kisses on people. It is a far cry from his WWE days.\n\n6 Kelly Kelly's Boyfriend\n\nvia youtube.com\n\nWrestlers who are lucky enough to find a wife have a much easier time in the wrestling industry. The same could not be said for Mike Knox. He was only a superstar for a very short time, and he was a part of one of the most bizarre moments in wrestling history when he got the Attitude Adjustment from Kane. Knox is not talked about much these days, but there was a time when people were curious about him.\n\nHe married the popular WWE superstar Kelly Kelly. The couple got married in 2014. It was a marriage that was not meant to last, as the couple divorced in 2016.\n\n5 Santino Marella\n\nvia ringsidenews.com\n\nSantino Marella will forever be a part of WWE. The WWE Hall of Fame is home to some great wrestlers and some very bad wrestlers. Santino is among those who have made an impression for one reason or another. His career with the company was not as successful as he had hoped, but he made a huge impact. He was the guy who was hilarious because he was so bad.\n\nSantino's last match was in 2014. He has done some backstage work, but the star has pretty much disappeared from WWE.\n\n4 Tyson Kidd\n\nvia heavy.com\n\nTyson Kidd was a true performer in WWE. He had a unique look, and he could really do it all. The man was skilled on the mic, in the ring, and he had a good personality. Everything came together for Tyson, but that would all change during an unfortunate injury. Tyson was a mid-carder who could have risen to a top level, but that was not meant to be.\n\nKidd is now recovering from his neck injury. He is still hoping to get into the ring again, but it is looking unlikely.\n\n3 Christian\n\nvia bodybuilding.com\n\nChristian is not a name that is being mentioned much these days. It is a shame, as he is one of the best performers of his time. Christian has made a career of being known for his impressive in-ring work and his ability to entertain fans. Christian was a popular superstar during his time in WWE and he did some good things for TNA.\n\nChristian has not been in the ring since December of 2015. He is trying to get back into the ring, but there is a possibility that his neck injury may have ended his career.\n\n2 Dolph Ziggler\n\nvia wrestlingnewssource.com\n\nDolph Ziggler had the chance to become a very big star in WWE. Unfortunately, things have not been going well for Ziggler. He was once a main eventer, but he has fallen to the bottom of the card. He is one of the many guys who will be pushed as the underdog who is going to have the moment of their life.\n\nZiggler has not been the focus of WWE. It is likely that he will be entering a new chapter of his career in the near future. The most likely place would be in Ring of Honor, but WWE has shown a lot of interest in Ziggler.\n\n1 Wade Barrett\n\nvia wrestlingrumors.net\n\nIt is going to be very hard for the casual fan to remember Wade Barrett. He had a short run in WWE, but he did his part as a mid-carder. Barrett is a former Intercontinental Champion. He did well in WWE as a member of The Nexus, but his star power never quite took off.\n\nBarrett has been out of WWE since 2014. He has had some ups and downs in the independent circuit, but his career has largely gone downhill. It is possible that he could get another chance with WWE. That would probably be the only way that fans would get to see him again.\n\nBarrett has been trying to make a name for himself in the business world. He launched a website last year. It is not likely that he will become a success, but it is nice to see him try to make something of his life.]" time="4.135"><properties><property name="score" value="0.9997806463922773" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[University: The primary mission of the University is to provide education. For that purpose, the University has established different faculties and departments, of which the oldest and most important is the faculty of theology. The latter is a privileged entity, by virtue of its academic reputation, both in the local and the global sphere. The mission of the University is to educate the professors and the students in the science of theology, which means, it also needs to equip them with the specific professional skills, which make them able to exercise the profession of the ministry.]" time="0.419"><properties><property name="score" value="0.9997656941413879" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The brother-in-law of the Palestinian-American teenager who was filmed being brutally beaten by Israeli police at a West Bank checkpoint says the incident \u201chas put a wedge\u201d between the Palestinians and Israelis.\n\nYoussef Almogy said he has had more than 50 phone calls from news media outlets following the brutal arrest of 16-year-old Tariq Abu Khdeir, the Associated Press reported.\n\n\u201cWhen I saw it, I cried,\u201d Almogy said. \u201cI was so shocked, but after that, I started thinking about my brother, who was jailed for 11 years, and then released for nothing. And I said, \u2018If they can do that to him, then imagine what they can do to us.\u2019\u201d\n\n\u201cAnd this has put a wedge between Palestinians and Israelis,\u201d he added.\n\nAlmogy was referring to the arrest of his brother, Amer, who was jailed for 11 years after he and a group of Palestinians tried to rob a Tel Aviv bank in 1994. Almogy said Amer was never convicted of a crime and that he believes the former prisoner is being held because of his brother\u2019s connection to the teen, the AP said.\n\n&quot;We're used to seeing this sort of thing,&quot; he said.\n\nAbu Khdeir, who was born in Florida but living with his family in East Jerusalem, was arrested Thursday after he was identified as being involved in a clash between]" time="1.033"><properties><property name="score" value="0.9790650010108948" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[\n\nToday we\u2019re excited to announce that Bethesda Game Studios has joined Oculus as a developer for the Oculus Rift. The partnership will see the creation of two exclusive, made-for-VR games from Bethesda Game Studios; the first is currently being designed for virtual reality and the second is being built for traditional monitors.\n\n\n\n\u201cAt Oculus, we\u2019re dedicated to helping developers and partners to grow, and the support and dedication we\u2019ve received from Bethesda Game Studios has been outstanding,\u201d said Brendan Iribe, CEO of Oculus. \u201cJohn Carmack, founder of id Software and now CTO at Oculus, believes in our vision and has created a custom version of Doom 3 BFG Edition for Oculus. We look forward to continuing to work with Bethesda as we develop our first generation of virtual reality games and experiences.\u201d\n\n\n\n\u201cI have a lot of personal interest in making sure that the VR gaming space is successful because I really believe in the technology and its potential to change gaming,\u201d said Carmack. \u201cAs a developer, the Rift is something that I\u2019ve wanted to work with for a long time, but the existing technology wasn\u2019t quite there yet. I\u2019m happy to be able to help to make the Rift a success and contribute to the rebirth of VR gaming.\u201d\n\n\n\nAt Bethesda Game Studios, developers are working on a number of unannounced projects including new entries in major franchises. Those who have seen this year\u2019s E3 show floor, Gamescom or Quakecon presentations have seen teases of what the studio is up to and a few of the titles we have in development are also being designed for virtual reality. The two games announced today are entirely new IP and we can\u2019t wait to show them to you when they\u2019re ready for primetime.\n\n\n\nFor those of you unfamiliar with Bethesda Game Studios, the acclaimed developer is responsible for creating many of the industry\u2019s most popular games including the \u201cElder Scrolls\u201d series, the single-player add-on to the original \u201cDoom\u201d and the \u201cFallout\u201d series. In fact, \u201cFallout 3,\u201d \u201cFallout: New Vegas\u201d and \u201cThe Elder Scrolls V: Skyrim\u201d were each awarded game of the year at the 2008, 2010 and 2011 BAFTA awards, respectively.\n\n\n\n\u201cWe have made some long-term technology investments with VR in mind,\u201d said Todd Howard, Game Director and Executive Producer at Bethesda Game Studios. \u201cHardware is finally catching up to the visions developers have had for virtual reality. Now when you put on a VR headset, you\u2019re transported to new and amazing worlds.\u201d\n\n\n\nThis year, Bethesda Game Studios showcased \u201cDoom 3 BFG Edition\u201d for Oculus. \u201cDoom 3 BFG Edition\u201d is a complete overhaul of \u201cDoom 3\u201d that brings modern technology and a classic, storied franchise into the world of virtual reality. This is the first title designed for virtual reality.\n\n\n\n\u201cJohn Carmack\u2019s work in the early 90s helped usher in the era of modern 3D gaming, and his genius continues to inspire developers to innovate,\u201d said Brendan Iribe, CEO of Oculus. \u201cThe team at Bethesda Game Studios is made up of the most talented developers in the industry and we can\u2019t wait to bring their newest games, made for virtual reality, to life.\u201d\n\n\n\nOver the past 20 years, Bethesda Game Studios has earned its reputation as one of the industry\u2019s most respected and accomplished game development studios. Creators of the 2006 and 2008 Game of the Year, \u201cThe Elder Scrolls IV: Oblivion\u201d and 2010 Game of the Year, \u201cThe Elder Scrolls V: Skyrim,\u201d Bethesda Game Studios has earned a reputation for excellence and quality with each of its five releases.\n\n\n\nThis partnership is another step toward realizing the future of virtual reality. It\u2019s an exciting time to be a gamer with the release of these groundbreaking VR titles, the arrival of next generation consoles and the continued evolution of the PC.\n\n\n\nStay tuned for more news on Bethesda Game Studios\u2019 titles for Oculus.\n\n\n\n\u2013 The Oculus Team]" time="2.633"><properties><property name="score" value="0.9480832517147064" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[\n\nHi~ B een a long time no see. Thank you for the continued support on the blog. So, how have you all been doing?\n\nToday, I have a special post. I know most of you do not know but it is the 'World Animal Day'. The day is celebrated worldwide on October 4 to create awareness of animal rights and welfare. The day is a very special one for me because I am a vegetarian, animal lover and a dog-person.\n\n\n\n\n\nSo for the last one year, I have been trying to spread the awareness among the people. I have been trying to encourage people to stop using animals as fur coats, cosmetics and for food. I want people to know the feelings of these animals, the pains they suffer and what they go through while they are alive. Also, the environmental damage they cause to the planet. It is very important to spread this message because animals are treated as if they are some sort of toys that we can play with.\n\n\n\n\n\nToday, I would like to share with you some tips on how we can help the animals and make the planet a better place to live. I also have a small personal pledge to make. Also, if you have not signed the pledge yet, you can also make your own pledge to protect the animals and the planet.\n\n\n\n\n\n\n\n\n\nTips and pledges to help the animals\n\n(made by my friend!)\n\nOne way we can make a change is by making an effort to not use cosmetics that are tested on animals. We can choose to use all cruelty free products and make a conscious effort not to use anything that has been tested on animals.\n\nWays to reduce our carbon footprints and help protect the environment:\n\nAvoid eating meat, or even better, go vegetarian (dairy, eggs are not vegan but I still encourage you to stop using dairy products as well)\n\nUse eco-friendly products - Not just cosmetics but also in our everyday lives.\n\nGo cruelty free - Choose cruelty free products and wear clothes that have not been made out of leather, fur or wool.\n\nStop buying those animals that we do not need. Also, choose to buy second hand or recycled clothes. Avoid buying pets if you cannot provide the proper care for them.\n\nAdopt, don't buy! You can adopt a dog or cat from an animal shelter. Do not buy animals from pet stores or breeders. They are sold to you as &quot;breed-free&quot; animals but most of them have been bred in horrible conditions. Also, make sure you know the sex of the animal you are getting, because they usually are sold as 'un-neutered' or 'un-spayed' animals.\n\nMake a conscious effort to not use plastic bags. Use reusable bags or better still, try to avoid using bags at all.\n\nTry to use products that do not harm the environment and are biodegradable.\n\nWays to help the animals:\n\nSign the pledge - To show that you support the cause to protect the animals.\n\nShare the post - Let others know of the cruelty that the animals go through. Let them know how they can help the animals too.\n\nJoin or start a protest or campaign to help the animals.\n\nThe small pledge I am making is to help the animals. Every month, I will be donating some money to an animal shelter in my town to help feed the dogs and cats there.\n\nThese are just some of the ways we can help the animals and the planet. We can all make a change, and make this planet a better place to live.\n\n\n\n\n\nIf you want to know how to become a vegan, and which cosmetics are tested on animals, you can read this post: How to be a vegan?\n\nRemember, we are what we eat, so be what you want to be. Be a good person, who loves animals and helps to save the planet.\n\nWith lots of love,\n\nPriya]" time="2.347"><properties><property name="score" value="0.9960466921329498" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Top Customer Reviews\n\nWhere reviews refer to foods or cosmetic products, results may vary from person to person. Customer reviews are independent and do not represent the views of The Hut Group.\n\nUnique, gritty, and wonderful The Good: The worlds are some of the best I've seen in a while, and how you explore the city and interact with its citizens gives a great sense of authenticity and intrigue. As a huge comic fan, the references are abound. The story and narrative is one of the best I've ever experienced. The Bad: The game is very linear and the combat while satisfying can get old at times. Overall: This is one of the best games I've ever played, period. Everything about it is gritty, dark, atmospheric, and enticing. While you have a very linear path to follow in the main story, the side quests and interactions you can have with the citizens is one of the best I've seen. You are truly the villain in this game, so if you want to be good and helpful to everyone, this is not for you. But for the fans of dark, gritty stories, this is one of the best. The story and narrative are top notch. The villains are some of the most compelling I've ever seen in a video game, with even the main villains being so much more compelling than the good guys. I loved this game from start to finish, and can't wait to see where the series goes from here. The ending while great is somewhat confusing, so if you aren't]" time="0.976"><properties><property name="score" value="0.9971693456172943" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Mrs. Mahilani Konczey\n\n&quot;The tutors at HW have made a huge difference in my child's academic performance, as well as in her confidence. The tutors are friendly, professional and punctual. My child enjoys her sessions and she has learned a lot from them. HW has a good location and easy parking. I highly recommend them.&quot;]" time="0.414"><properties><property name="score" value="0.9992594122886658" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99925941&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99925941
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i125]" time="0.173"><properties><property name="score" value="0.9970126152038574" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i126]" time="2.322"><properties><property name="score" value="0.9601956248283386" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i127]" time="0.335"><properties><property name="score" value="0.9374892711639404" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i128]" time="2.667"><properties><property name="score" value="0.6360526204109191" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i129]" time="1.310"><properties><property name="score" value="0.4337228735287984" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i130]" time="2.495"><properties><property name="score" value="0.20112688541412355" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i131]" time="2.741"><properties><property name="score" value="0.11983842849731445" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255132 (text: Harry Kane has reminded us of the fantastically da) is an LLM-generated sample, misclassified as human-generated with confidence 0.11983843&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255132 (text: Harry Kane has reminded us of the fantastically da) is an LLM-generated sample, misclassified as human-generated with confidence 0.11983843
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Flavor Pairs\n\n(Home)\n\nPickled Meat\n\nBacon, Flank Steak, Pork Butt, Bacon, Hams, Sausage, Canned Ham\n\nFor Cooking/Grilling\n\nPeppers, Onions, and other Vegetables\n\nPotatoes, Onion, Green Peppers, Jalapenos, Garlic, and other Vegetables\n\nPastas\n\nPasta, Macaroni, Shells, Ramen Noodles, Ziti, Gnocchi, Fusilli, and other Pastas\n\nSoups and Stews\n\nSoups, Stews, Chili, and other One-Pot Meals\n\nCorn, Peas, and other Grains\n\nCorn, Peas, Risotto, Polenta, Grits, Quinoa, Rice, and other Grains\n\nBeverages\n\nBeer, Alcohol, Soda, Juices, Tea, Coffee, Soda, and other Beverages]" time="0.982"><properties><property name="score" value="0.9997889399528503" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[And now, ladies and gentlemen, I give you the breathless high-speed ranting I did when I saw the (potentially) wackadoo conclusion to this storyline.\n\nOh, my God, you guys, this episode has a freaking atomic bomb on the side of it.\n\nOkay, cool. It\u2019s the mid-season finale. I know there\u2019s going to be some serious plotlines.\n\nAnd seriously, what was up with the weird science in this episode?\n\nNo. No, I did not just watch an episode with the headline: \u201cOuch, My Butt!\u201d\n\nThat plotline was weird as hell.\n\nBut I can understand why they did it. He\u2019s trying to be a \u201csophisticated ladies man\u201d and not the boy they\u2019ve been dealing with all these years.\n\nOh, now this is hilarious. But it\u2019s also dangerous, if you\u2019ve watched any superhero movie ever.\n\nIn the brief moment that I was allowed to breathe, I actually enjoyed this part.\n\nNot only is this scientifically improbable, but also\u2026 how does a security camera shoot a laser?\n\nI know this is only in fiction, but I do believe that they\u2019re committing suicide.\n\nThis is also part of the reason why I wanted to talk about this episode. I know it\u2019s not a good look. It\u2019s not exactly smart of him to do this.\n\nI don\u2019t even know what to say here.\n\nWHY?\n\nHe\u2019s not afraid of bugs, but he\u2019s scared of that snake?\n\nI mean, yeah, I get that he\u2019s doing it for a different reason, but\u2026\n\nIf it weren\u2019t for that, then I would not have laughed so hard at this.\n\nOkay, look, I was done with this season for a while. I thought that it was becoming a boring clich\xe9d plot that they were sticking to, so I was ready to say goodbye to it. But then I watched this, and I just felt like I was watching a new show. There was so much going on, and they were throwing so much at us at once, and it was just such a good use of time. I\u2019m really happy that they had this show, because it was seriously pretty funny and it was also action-packed. There was a lot going on, but it was very clear, and it was fun to watch. I just loved how much was going on. I loved it, and I really did enjoy it. It was a bit much, and they took a lot of crazy chances, but it was just so fun to watch. That was some crazy stuff, and it was nice to see something so good after so much of the regular garbage we were getting. Overall, I loved it.]" time="2.062"><properties><property name="score" value="0.9058307210604349" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Specification:\n\nProduct Name Bags Model SA01A Material Canvas + Zipper Size 29*15.5*10.5cm Color Red, Green, Black, Blue, Gray, Purple, Brown, Khaki, Yellow, Orange Weight 10g Quantity 1 Piece(s) Pattern Solid Inner Pocket Yes Closure Drawstring\n\nPackage include:\n\n1 X Canvas Bags\n\n\n\n\n\nDescription:\n\nDimensions of the storage bag, cool way to store the supplies for travel and daily life.\n\nLightweight, durable and sturdy canvas construction.\n\nA simple pouch, plenty of storage space to put many small items.\n\nThere is a handle for easy carrying and a good choice for daily life.\n\nUsed for carrying, practical, suitable for outdoor, camping, traveling, adventure, daily life, perfect for storing jewelry, makeup, keys, etc.\n\n\n\n\n\nSize: 29*15.5*10.5cm\n\n\n\nWeight: 10g\n\nMaterial: Canvas + Zipper\n\n\n\nColor: Red, Green, Black, Blue, Gray, Purple, Brown, Khaki, Yellow, Orange\n\nPattern: Solid\n\nInner Pocket: Yes\n\n\n\nClosure: Drawstring\n\n\n\n\n\n\n\nPackage include:\n\n1 X Canvas Bags]" time="0.714"><properties><property name="score" value="0.9935429096221924" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Author:\n\nFollowing a stint as a park ranger in Wyoming, Dan Beal relocated to Seattle and received a degree in science writing from the University of Washington.\n\nFollowing a stint as a park ranger in Wyoming, Dan Beal relocated to Seattle and received a degree in science writing from the University of Washington.]" time="0.219"><properties><property name="score" value="0.9994940757751465" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[(Adapted from the 2014 Urban Tree Canopy Conference)\n\nPotholes are problematic for all of us, and for our city\u2019s trees they represent a real danger. Roots can become entangled in the small cracks between the asphalt and the cement to create the impression that the tree is growing out of the road! And when the roots do become entwined, the pressure can be so great that the roots are destroyed. We all know the danger this creates for the trees.\n\nThe solution to this problem is obvious: pave over the whole road surface. Not so fast! In the City of Chicago, for example, there is a practice that preserves the utility of the tree\u2019s roots and of the road surface itself. The practice is called \u201cmilling\u201d. Milling involves cutting the asphalt out in a shape that will not disturb the roots and adding back in a new surface. While the tree is protected, the roots are not disturbed and the road surface is left intact. When a tree is scheduled for removal, this is often the method of choice. Milling also involves improving the root zone around the tree. The roots of many trees can\u2019t penetrate the compacted soil that is found beneath pavement, so it is necessary to mill the area to promote root growth.\n\nMilling is less expensive than paving over a road surface and can also be less expensive than removal of a tree. Milling should be done whenever possible and certainly before a tree is scheduled for removal. There is no need to sacrifice a mature, healthy tree.\n\nAdvertisements]" time="1.835"><properties><property name="score" value="0.026988595724105835" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[How do i search for files with a specific suffix using Unix?\n\nOn Unix, to search for files that ends with a particular file extension, use the following command:\n\nfind . -name '*.html'\n\nOR use grep command to list all files in a directory and subdirectories that have the suffix:\n\nfind . -type f | xargs grep -l &lt;suffix&gt;\n\nSample outputs:\n\n./Ajax/contact-us/index.html ./test/index.html\n\nThe above commands are equivalent to the grep command on Linux/MacOSX or the find command on Linux/MacOSX:\n\n$ find . -name &quot;*.html&quot; ./test/index.html ./Ajax/contact-us/index.html $ grep -l -Ri -f -E &quot;.*\\.html&quot; ./ test/index.html Ajax/contact-us/index.html\n\nFind out all files with a specific filename (using shell pattern match):\n\nfind . -name '*.html'\n\nOR use the egrep command to list all files in a directory and subdirectories that have the suffix:\n\nfind . -type f | xargs egrep -l &lt;suffix&gt;\n\nSample outputs:\n\n./Ajax/contact-us/index.html ./test/index.html\n\nOR\n\n$ find . -name &quot;*.html&quot; ./test/index.html ./Ajax/contact-us/index.html $ egrep -i &quot;.*\\.html&quot; . Ajax/contact-us/index.html test/index.html\n\nFind files in a directory and subdirectories:\n\nfind . -name &quot;*.html&quot;\n\nOR\n\nfind . -type f | xargs grep -l &lt;suffix&gt;\n\nSample outputs:\n\n./Ajax/contact-us/index.html ./test/index.html\n\nOR\n\n$ find . -name &quot;*.html&quot; ./test/index.html ./Ajax/contact-us/index.html $ egrep -i &quot;.*\\.html&quot; . Ajax/contact-us/index.html test/index.html\n\nFind files in a directory and all]" time="1.245"><properties><property name="score" value="0.005786925554275513" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Russia has expanded its ban on U.S. pork to include all American meat imports as the diplomatic row over alleged election meddling and the situation in Syria rumbles on.\n\nLast week, the Russian government added five new countries to the blacklist: South Korea, Montenegro, Iceland, Liechtenstein and Ukraine. While the list included major agricultural producers such as Brazil and Canada, U.S. meat producers were noticeably absent. However, on Wednesday, the government added the U.S. to the list of banned meat importers.\n\nThe ban was announced by the state consumer watchdog Rospotrebnadzor. A government official told Russian media that American meat imports were now being kept out of Russia on a \u201ctemporary basis\u201d.\n\nRussia is the fourth largest importer of U.S. meat, buying $51 million worth of poultry and $22 million of pork in 2016. According to the National Pork Producers Council, the U.S. shipped 27,600 metric tons of pork to Russia in 2016, which made up 0.3 percent of total U.S. pork exports.\n\nRead more: Polls Say Few Americans Think Trump is Putin's Puppet\n\nThe ban comes at a particularly difficult time for the U.S. meat industry, which is still struggling to recover from a disastrous salmonella outbreak that began in 2015. China, Vietnam and Mexico have all banned U.S. pork imports, while the EU and South Korea are still testing shipments.\n\nThe bans come as U.S. meat exporters have had difficulty convincing importers to trust American meat following the 2015 crisis. The outbreak, which resulted in more than 60 percent of pork being recalled, was linked to at least 184 illnesses in 20 states. While U.S. authorities traced the outbreak back to a farm in Missouri, they were unable to identify the exact source of contamination.\n\n\u201cThis has been a major setback for us and it will take a while for the industry to recover,\u201d said National Pork Producers Council spokesman Dave Warner.\n\nIt's not clear when Russia will lift its ban on U.S. meat, but Russia's state consumer protection watchdog has said it would allow Ukrainian meat imports.]" time="1.249"><properties><property name="score" value="0.8646958470344543" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Is there anything you would change about this book?\n\nIt's too long. I got bored. It's the wrong length for a book about the Crusades. I know you want to make it as comprehensive as possible, but let's face it, the Muslim perspective is the only one that matters. You can understand why people fought for this or that minor piece of territory, but you can't understand why they fought in the first place.\n\nHow would you have changed the story to make it more enjoyable?\n\nI wish they had done a better job of establishing the atmosphere of the time, instead of just going for sensationalism.\n\nHow could the performance have been better?\n\nThe narrator had a good voice for this.\n\nWas Crusade the right choice for you? Why or why not?\n\nNo, I'd never read about the Crusades before.]" time="0.884"><properties><property name="score" value="0.9998273253440857" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Most Helpful Positive Review\n\nI would purchase this product again\n\nI have used a baby wrap before for a previous baby and loved it, but this one is my favorite. The material feels great and is breathable, the design is cute and easy to use, and I love...Read complete review\n\nI have used a baby wrap before for a previous baby and loved it, but this one is my favorite. The material feels great and is breathable, the design is cute and easy to use, and I love how it fits under my shirt to show a little but isn't quite as obvious as a sling. I use this for my 1 month old and we have never had any problems with the carrier not staying tight enough. I would definitely recommend this product!]" time="0.760"><properties><property name="score" value="0.9998278617858887" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[A northwest Houston woman is under arrest, accused of child abuse.\n\nPolice say they were called to the woman's home on Almeda School Road around 10:45 a.m. Thursday. They say her husband had to take their child to the emergency room after the child suffered a burn.\n\nPolice say the child suffered burns to the back of his neck, shoulder and upper back area.\n\nPolice say the wife, 31-year-old Mercedes Victoria Gonzalez, was arrested.]" time="0.377"><properties><property name="score" value="0.9994427561759949" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[(NC)\u2014Despite the odds, there are some consumers who decide to head to the stores this weekend to start Christmas shopping. For some, the reason is because they still have some of their shopping list to complete. Others, however, have an entirely different reason.\n\nAccording to Statistics Canada, the average Canadian adult will spend $473.12 on gifts this holiday season. That means the average consumer will spend almost $35 more this year than last year.\n\n\u201cA lot of consumers have gone overboard,\u201d says Ted Michalos, founder of CMG Financial. \u201cIt\u2019s not necessarily the spending that\u2019s the problem, it\u2019s what they\u2019re spending the money on.\u201d\n\nThat\u2019s because many shoppers are buying gifts for themselves. That doesn\u2019t necessarily mean they\u2019re spending on the latest fashions, but rather toys and electronics. And the majority of these gifts, at 41.8 per cent, are going to be purchased online.\n\nHowever, the reason why so many people are spending online is because they\u2019re trying to avoid the crowds and waiting in lines.\n\n\u201cSome consumers just get too stressed and they go online to get their stuff,\u201d says Michalos. \u201cAnd if they do that, they\u2019re likely going to be paying more for the gifts.\u201d\n\nThat\u2019s because while there are fewer people at the store, they\u2019re paying more per person. For example, if someone is paying $50 for a gift in the store, they\u2019re likely to pay more than $50 online.\n\n\u201cThis is really the time where you\u2019re going to have to do your best to have everything done by the 15th,\u201d says Michalos. \u201cAnd it\u2019s definitely a good idea to sit down with your spouse or significant other and look at your budget. And when you have a child, the best thing to do is to write down what you\u2019re going to spend. That\u2019s a real good thing to do, because if you don\u2019t, you\u2019ll overspend.\u201d\n\nAnd that could come with consequences, because if you\u2019re overspending, you could be limiting your ability to spend on your children or other family members in the future.\n\n\u201cYou can go overboard and give away everything you have to your kids and that could be a problem,\u201d says Michalos. \u201cThis could be an opportunity to teach your kids about moderation, and also the difference between \u2018I want\u2019 and \u2018I need.\u2019\u201d\n\nBecause with Christmas, most people are shopping with the mindset of \u2018I want,\u2019 not \u2018I need.\u2019 And that can be a problem, because it will prevent you from saving or investing.\n\n\u201cIf you\u2019re going to be shopping on a credit card, you\u2019re going to be in big trouble,\u201d says Michalos. \u201cIf you\u2019re not able to pay off that debt by the end of January, you\u2019re going to be in big trouble.\u201d\n\nWant to know more about debt? Try these articles:\n\nNew Christmas ad showcases true meaning of the holidays\n\nTake advantage of these holiday deals\n\nHow to get your kids to be thankful for what they have\n\nWhat do you think? Join the conversation in the comments below or contact us using the \u201cHave Your Say\u201d form on this page.]" time="4.805"><properties><property name="score" value="0.9997683167457581" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Developing in a Virtual Machine on your computer or laptop\n\nHello , Here we are going to learn how to create virtual machine on your computer for testing purpose. Before we start , it is better to know what is virtual machine. A virtual machine is a simulated hardware that allows you to run other operating systems (or other versions of the same OS) at the same time as your current OS, inside a window on your current OS. It can also be thought of as a computer inside a computer. With a virtual machine, you can: Run another operating system (or another version of the same operating system) at the same time as your current operating system. Try out a new operating system without having to install it on your hard disk. Run applications for one operating system on another, without having to reboot your computer.\n\nStep 1: Create a New Virtual Machine\n\nVirtualBox provides a wizard that will assist you in creating a virtual machine. Launch the program and select \u201cNew\u201d.\n\nStep 2: Create a Virtual Machine\n\nThe first window you see is the New Virtual Machine Wizard. Select \u201cCreate a virtual machine\u201d.\n\nStep 3: Create a Virtual Machine\n\nIn the next window, enter the name for your virtual machine. This name will appear in the main VirtualBox window next to the icon for the virtual machine.\n\nStep 4: Create a Virtual Machine\n\nThe next window is where you select the operating system you want to install.\n\nStep 5: Create a Virtual Machine\n\nIn the next window, select the amount of memory you want to allocate to your virtual machine. The more memory you have, the more applications you can run simultaneously. However, keep in mind that the more memory you allocate, the slower your computer will be.\n\nStep 6: Create a Virtual Machine\n\nSelect \u201cCreate a virtual hard disk\ufffd]" time="2.124"><properties><property name="score" value="0.9991336464881897" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The internet is a funny place, where people can get away with saying and doing the weirdest stuff that would otherwise be totally unacceptable.\n\nBut it turns out not even the internet is a free pass to sexually harass people.\n\nThat's what one woman learned when she was sent a totally inappropriate, NSFW photo from a stranger, that happened to be someone she knew in real life.\n\nTwitter user @kendtee, who describes herself as &quot;an arm candy&quot;, tweeted about her encounter with the man she knew.\n\nAccording to @kendtee, the incident occurred after she agreed to take a picture with a guy. She then deleted the picture from her camera roll when it was over, and the man asked if she had. She said yes, and then he said this:\n\n&quot;It's a shame we can't see the photo anymore because it was pretty good&quot;\n\nShe didn't know how to respond. He then tweeted her back, asking her to &quot;leave [the photo] up for a little while&quot;. She declined.\n\nApparently, he then texted her back, asking her to send him a copy of the picture.\n\nShocked by this behaviour, she shared the exchange on Twitter, and said the worst part was how &quot;unbothered&quot; he seemed about it.\n\nEventually she shared a screenshot of the exchange, and people were not impressed.\n\n&quot;Oh wow. 'I'd do ya if I could' and then the photo request. You should probably just block him and be done with it. 'Doing ya' or not,&quot; one person said.\n\nAnother person had an idea on how to deal with it, tweeting, &quot;You know what? Be a bigger person, send it to him. For posterity.&quot;\n\nMany others were in complete agreement with this sentiment.\n\n&quot;I don't know this man and I probably never will. But I 100% understand why he's going to be blocked. A man who says that is definitely not a good person. All women should know their boundaries with men and this seems to me to be clear cut,&quot; one person said.\n\nIn the end, @kendtee went back to him and asked him to explain why he'd sent her that photo.\n\nAccording to @kendtee, he sent her this message in response:\n\n&quot;How do you think I know that you're gonna see this tweet and block me or delete it.\n\n&quot;I'm not trying to be a creep you just said that you'd do me so I figured we had a shot lol\n\n&quot;I like your style and want to get to know you better\n\n&quot;If you send the picture I promise not to be creepy and stop bothering you\n\n&quot;I'm being serious here and I don't think I deserve the reaction I got\n\n&quot;All I wanted to do was take a picture with a nice girl and see if she'd let me\n\n&quot;I don't want to be a bad person or make you feel uncomfortable.\n\n&quot;If you like I'll just go away\n\n&quot;You're gorgeous and I'm sorry I don't deserve to meet you but I thought I'd try my luck\n\n&quot;I won't message you again if you don't want me to\n\n&quot;I'll delete the tweet and block you\n\n&quot;I just wanted to try\n\n&quot;Please&quot;\n\nThe rest of Twitter wasn't impressed, calling him out for still not getting it, and for trying to shift the blame.\n\n&quot;Like you have to explain what 'likes your style' means, like you're trying to shift the blame. But it's not ok. It's not ok for you to message someone for a photo. It's not ok for you to say 'I'd do you' and it's not ok for you to try and justify your actions,&quot; one person said.\n\nAnother pointed out that he was trying to pressure her into sending him the photo, and wasn't taking responsibility for his actions.\n\n&quot;He's a creep, and still wants to force you to send him a picture of your body. He'll delete the tweet if you do, otherwise you can block him. This is on him, not you. You have no responsibility to 'meet' him,&quot; another person said.\n\n&quot;Ok let's stop there with the 'I would've taken a photo with you anyways' thing. That's not what this is about. This is about someone sending unsolicited nudes,&quot; another person pointed out.\n\nIn response, @kendtee wrote that she wasn't bothered, because at the end of the day she was a &quot;pretty woman who didn't need to be respected&quot;.\n\n&quot;I didn't get mad or freak out or feel the need to drag his name through the mud. I just calmly informed him that I have a man in my life and that I'm flattered but I don't share photos like that and I'm taken. He didn't like it so he blocked me and I'm okay with that,&quot; she wrote.\n\nALSO ON HUFFPOST AUSTRALIA]" time="4.612"><properties><property name="score" value="0.9016384959220887" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Perhaps one of the most beautiful areas of Southern California is the Pacific Palisades, a neighborhood filled with high rises, luxury estates, and tons of culture. From the obvious to the under the radar, Pacific Palisades has some of the most fascinating museums, gardens, and activities you can visit. When you need to get out of the house and explore a new part of Los Angeles, the Pacific Palisades is a fantastic choice.\n\nThe Getty Villa\n\nThe Getty Villa is a sight to behold, and is a must see when you are in the Pacific Palisades. You are welcome to explore the museum for free on the first Thursday of every month. They offer a free trolley that will take you to the Getty Villa, so you don\u2019t have to worry about navigating through the Pacific Palisades on a crowded day. This is the perfect time to visit the Pacific Palisades.\n\nThere are two events going on at the Getty Villa on Thursdays: Story Time and Pacific Voices. Story time will give you a better understanding of ancient Greek and Roman culture. It is a good idea to purchase your tickets in advance, so you can get a discount. The Pacific Voices event is a singing performance, and a great way to enjoy your time at the Getty Villa.\n\nSanta Monica Mountains National Recreation Area\n\nThe Santa Monica Mountains are a beautiful way to spend a day in the Pacific Palisades. The National Recreation Area has a fantastic number of trails you can explore, and is a great way to get out of the house and spend time with your family. Take the trail to Leo Carillo State Beach, or go for a hike along Rustic Canyon. Whichever trail you take, you are sure to enjoy your time in the Pacific Palisades.\n\nSanta Monica Baykeeper\n\nThe Santa Monica Baykeeper is a fantastic organization that works hard to clean up the Santa Monica Bay. This is a great way to spend time in the Pacific Palisades and make a real difference. Check their website for upcoming volunteer opportunities. You will learn more about the Santa Monica Bay, and get a chance to spend time with some interesting people.\n\nPacific Palisades Center for the Arts\n\nThe Pacific Palisades Center for the Arts is an excellent way to spend an afternoon. This isn\u2019t a museum, so you don\u2019t need to book in advance. It is free to walk around the grounds, and to watch a performance. They have everything from jazz concerts to visual arts. If you want to learn more about the Pacific Palisades, then this is a great place to start.\n\nPacific Palisades Farmers Market\n\nThe Pacific Palisades Farmers Market is an excellent way to spend your Saturday in the Pacific Palisades. From fruits and vegetables, to homemade items and baked goods, there are plenty of delicious items to choose from. This is a great place to bring the kids, as they have tons of activities.\n\nPacific Palisades Art Walk\n\nThe Pacific Palisades Art Walk is a fantastic opportunity to view some of the beautiful artwork created by locals. It is a fun evening filled with laughter and good food. You are welcome to bring your own picnic basket and enjoy it in the beautiful Pacific Palisades.]" time="2.188"><properties><property name="score" value="0.9997794032096863" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[We have been very pleased with the support, workmanship and professionalism of the team at WeFix. They have been able to provide quick and effective solutions for any problems that have arisen. The attention to detail, willingness to understand and understand the importance of high quality work have been important to us. The flexibility of being able to accommodate and adapt to any situations has been very helpful in terms of our team and their working styles.\n\n- Rick Phillips, Site Manager \u2013 Maersk Oil]" time="0.323"><properties><property name="score" value="0.9998255372047424" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Mastering financial knowledge and planning for a good life is a lifelong process. But when it comes to money, how many people have the knowledge to plan for a good life? And what if you had this information, would you use it? How would you live? Would your life change for the better? Would your relationships improve? Would your life be happier?\n\nIn most cases, we don\u2019t have enough financial knowledge to make the right decisions. We don\u2019t have the knowledge and therefore, we end up making bad decisions. The result of these bad decisions is that we lose more than we gain. We lose money and we also lose our peace of mind.\n\nI have experienced both sides of the coin and I know the emotional, mental and physical impact of being financially unstable. Money makes and breaks people. It has broken me in the past and it took me years to recover. But I am better now.\n\nIf you too want to make peace with your money]" time="0.633"><properties><property name="score" value="0.9997904896736145" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Tom Shanks (writer)\n\nTom Shanks (born 1961) is a Scottish writer, journalist, and musician. He is the author of eight novels, including &quot;The Hidden&quot;, the &quot;Vampyrrhic Times&quot; trilogy, and &quot;Eve&quot;, an &quot;Eve&quot; (comics) spin-off novel.\n\nShanks has been living in France since 1983 and is married to the French writer Marie H\xe9l\xe8ne Poitras.\n\nShanks was born in Edinburgh, Scotland in 1961.\n\nShanks has been writing since the mid-1990s and has published several novels.\n\nShanks has also been writing for the British comic book magazine &quot;2000 AD&quot;. His first script for them was the short story &quot;Trevor's Tale&quot;, drawn by Simon Harrison, in &quot;2000 AD&quot; #1704, published in March 2008.\n\nThe comic book &quot;Eve&quot; was written by Shanks. It was released by Black House Comics in 2011 and the second volume was published in 2012. In 2012, IDW Publishing published a comic book prequel to &quot;Eve&quot;, &quot;Eve: The Awakening&quot;. The book was written by Shanks and Marie H\xe9l\xe8ne Poitras and illustrated by Enrica Angiolini.\n\nShanks has also written the script for &quot;Apocalypse Cow&quot;, which was originally published in French and German as &quot;La Vache Apocalypse&quot;. It was adapted as a comic book by Tony Lee.\n\nIn March 2010, the novel &quot;Vampyrrhic Times&quot; was published by Gollancz in the UK and by Roc in the US. It was illustrated by Tony Harwood. The second volume of the trilogy, &quot;The Wisdom of Dead Men&quot;, was published in the UK in 2011 and the US in 2012. The third volume, &quot;The Trade of Queens&quot;, was published in the UK in 2013 and in the US in 2014.\n\n\n\n\n\n\n\nShanks and Poitras run the Black House Comics publishing house.\n\nShanks is a member of the band The Bassholes, along with co-writer of Eve, Marie H\xe9l\xe8ne Poitras. The band is composed of Shanks and Poitras on vocals and bass, Poitras' husband, French artist St\xe9phane Paitreau, on guitar, and drummer &quot;Chino&quot; Gonzales. They released their first album, &quot;No More Mr Nice Guy&quot;, on May 1, 2010, on Stuck on a Name Records. The album includes covers of classic rock songs, such as &quot;Jailhouse Rock&quot; and &quot;Born to be Wild&quot;. In 2013, the band released a second album, &quot;Sex Lies &amp; Basslines&quot;.\n]" time="2.593"><properties><property name="score" value="0.4642360806465149" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Rick Lake\n\nRichard G. Lake II (born April 7, 1979) is an American mixed martial artist who most recently competed in the Welterweight division of the Ultimate Fighting Championship. A professional competitor since 2004, he has also competed for the Quad Cities Silverbacks of the IFL, and was a competitor on Spike TV's &quot;&quot;.\n\nBorn and raised in the Quad Cities area, Lake attended East Moline High School.\n\nAfter graduating from high school, Lake worked for two years as a full-time welder, before deciding to pursue a career in mixed martial arts.\n\nLake made his professional mixed martial arts debut on October 30, 2004 for the Quad Cities Silverbacks of the International Fight League. He won his first three fights for the promotion, before suffering his first loss to Ian Loveland on April 8, 2005.\n\nAfter picking up three wins in the IFL, Lake was signed by the UFC. He was expected to make his debut at UFC Fight Night 7 against Jason Gilliam. However, Gilliam was forced from the card with an injury, and was replaced by Jorge Rivera. Lake lost to Rivera via first-round KO on August 23, 2007.\n\nFor his second fight in the promotion, Lake dropped down to Lightweight and faced Kyle Bradley at UFC Fight Night 11 on November 17, 2007. He won the fight via unanimous decision (30-27, 30-27, 29-28).\n\nIn his third fight for the promotion, Lake dropped to Featherweight to take on Mike Brown at UFC Fight Night 13 on June 16, 2008. He lost the fight via unanimous decision (29-28, 29-28, 29-28).\n\nFor his fourth fight in the promotion, Lake dropped to Bantamweight and faced Brad Pickett at UFC 95 on February 7, 2009. He lost the fight via TKO in the first round.\n\nFor his fifth fight in the promotion, Lake faced Tommy Hayden at UFC 100 on July 11, 2009. He won the fight via submission in the first round, earning &quot;Submission of the Night&quot; honors.\n\nFor his sixth fight in the promotion, Lake faced George Roop on March 21, 2010 at . He won the fight via unanimous decision.\n\nFor his seventh fight in the promotion, Lake fought Roland Delorme on June 19, 2010 at . He lost the fight via submission in the first round.\n\nFor his eighth fight in the promotion, Lake faced John Gunderson on November 20, 2010 at UFC 122. He won the fight via split decision.\n\nFor his ninth fight in the promotion, Lake fought Jason Reinhardt on February 5, 2011 at UFC 126. He won the fight via TKO in the first round.\n\nFor his tenth fight in the promotion, Lake faced newcomer Darren Uyenoyama on June 11, 2011 at . He won the fight via unanimous decision (30-27, 30-27, 30-27).\n\nFor his eleventh fight in the promotion, Lake faced former WEC Featherweight Champion Mike Brown on October 29, 2011 at UFC on Versus 6. He lost the fight via unanimous decision (29-28, 29-28, 29-28).\n\nFor his twelfth fight in the promotion, Lake faced Nam Phan on May 15, 2012 at UFC on FX 4. He lost the fight via unanimous decision.\n\nFor his thirteenth fight in the promotion, Lake faced Sam Sicilia on July 11, 2012 at UFC 149. He won the fight via unanimous decision.\n\nFor his fourteenth fight in the promotion, Lake faced Ivan Menjivar on December 15, 2012 at UFC 155. He won the fight via split decision.\n\nFor his fifteenth fight in the promotion, Lake faced Jeremy Larsen on June 15, 2013 at UFC 161. He won the fight via TKO in the second round.\n\nFor his sixteenth fight in the promotion, Lake faced promotional newcomer Justin Salas on November 16, 2013 at UFC 167. He lost the fight via split decision.\n\nFor his seventeenth fight in the promotion, Lake faced fellow newcomer Max Holloway on February 15, 2014 at UFC 169. He lost the fight via TKO in the first round.\n\nAfter his release from the UFC, Lake dropped down to Bantamweight and made his debut on April 11, 2015 at IFC: International Fighting Championship 2 in Des Moines, Iowa against Rafael &quot;Morcego&quot; da Silva. He lost the fight via knockout in the first round.\n\nLake faced Gabriel Checco on May 28, 2015 at CXF 5. He lost the fight via split decision.\n\nLake faced Daniel Weichel on August 14, 2015 at Superior Challenge 12. He lost the fight via submission in the first round.\n\nIn his fourth fight since returning from the UFC, Lake faced Ryan Quinn on October 23, 2015 at CXF 8. He won the fight via knockout in the first round.\n\nLake faced Jim Alers on January 8, 2016 at CXF 10. He won the fight via TKO in the second round.\n\nLake faced Martin Stapleton on February 19, 2016 at MFW: A Night of Champions. He lost the fight via split decision.\n\nLake next faced Dylan Logan on September 2, 2016 at Cage Warriors 82. He won the fight via TKO in the second round.\n\nOn February 28, 2017, it was announced that Lake had signed with the UFC and was expected to face Alessio Di Chirico on March 18, 2017 at UFC Fight Night 107. However, on March 7, Di Chirico pulled out from the event, citing injury, and was replaced by promotional newcomer Marcin Prachnio. Lake won the fight via TKO in the second round.\n\nLake faced Gregor Gillespie on June 25, 2017 at UFC Fight Night 112. He lost the fight via split decision.\n\nLake faced Jarred Brooks on December 9, 2017 at UFC Fight Night 123. He lost the fight via submission in the first round.\n\nLake faced Drakkar Klose on February 24, 2018 at UFC on Fox 28. He lost the fight via TKO in the second round.\n\nLake faced Andre Fili on August 25, 2018 at UFC Fight Night 135. He lost the fight via TKO in the first round.\n\nLake faced Desmond Green on February 17, 2019 at UFC on ESPN 2. He lost the fight via technical knock out.\n\nIn July 2007, Lake was diagnosed with Crohn's disease, an inflammatory bowel disease that causes ulcers in the digestive tract. He began experiencing symptoms shortly after moving to California to train full-time. His Crohn's was brought under control through medication and diet, and it is currently in remission.\n\n\n\n]" time="3.695"><properties><property name="score" value="0.9707017838954926" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Fenix Funk\n\nFenix Funk is the debut studio album by German electronic music group Telepopmusik. It was released on 9 October 2000 by Columbia Records. It was a success in Japan and peaked at #50 on the Oricon Albums Chart.\n]" time="0.197"><properties><property name="score" value="0.9423989653587341" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i0]" time="1.256"><properties><property name="score" value="0.33803772926330566" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i1]" time="1.555"><properties><property name="score" value="0.9909436305363973" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i2]" time="0.738"><properties><property name="score" value="0.21650105714797974" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i3]" time="0.871"><properties><property name="score" value="0.9852544367313385" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i4]" time="0.839"><properties><property name="score" value="0.9996868371963501" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i5]" time="0.920"><properties><property name="score" value="0.15916651487350464" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i6]" time="2.074"><properties><property name="score" value="0.8950265645980835" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i7]" time="1.289"><properties><property name="score" value="0.9805915355682373" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i8]" time="2.303"><properties><property name="score" value="0.4221643606821696" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i9]" time="1.739"><properties><property name="score" value="0.9949692885080973" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i10]" time="0.869"><properties><property name="score" value="0.9889686703681946" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i11]" time="1.667"><properties><property name="score" value="0.9995559453964233" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i12]" time="0.782"><properties><property name="score" value="0.9032338857650757" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i13]" time="1.417"><properties><property name="score" value="0.4955202341079712" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i14]" time="0.796"><properties><property name="score" value="0.004489243030548096" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i15]" time="2.309"><properties><property name="score" value="0.9933535854021708" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i16]" time="1.859"><properties><property name="score" value="0.9751461346944174" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i17]" time="1.473"><properties><property name="score" value="0.9997092485427856" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i18]" time="2.184"><properties><property name="score" value="0.9994335969289144" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i19]" time="2.708"><properties><property name="score" value="0.9996739625930786" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i20]" time="0.712"><properties><property name="score" value="0.9998303651809692" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i21]" time="1.669"><properties><property name="score" value="0.9996842940648397" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i22]" time="0.700"><properties><property name="score" value="0.9996777772903442" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i23]" time="1.820"><properties><property name="score" value="0.930377463499705" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i24]" time="0.719"><properties><property name="score" value="0.24683469533920288" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i25]" time="1.785"><properties><property name="score" value="0.9998238682746887" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i26]" time="2.070"><properties><property name="score" value="0.9997778534889221" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i27]" time="1.696"><properties><property name="score" value="0.999743382136027" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i28]" time="0.940"><properties><property name="score" value="0.9996185302734375" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i29]" time="0.731"><properties><property name="score" value="0.16794809699058533" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i30]" time="1.051"><properties><property name="score" value="0.9776257276535034" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i31]" time="0.500"><properties><property name="score" value="0.9997774958610535" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i32]" time="1.277"><properties><property name="score" value="0.999674121538798" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i33]" time="1.649"><properties><property name="score" value="0.5434567332267761" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i34]" time="1.808"><properties><property name="score" value="0.996519406636556" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i35]" time="1.708"><properties><property name="score" value="0.9756788015365601" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i36]" time="1.476"><properties><property name="score" value="0.9997512896855673" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i37]" time="1.954"><properties><property name="score" value="0.9998273650805155" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i38]" time="1.491"><properties><property name="score" value="0.8932798902193705" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i39]" time="1.903"><properties><property name="score" value="0.9887590805689493" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i40]" time="3.211"><properties><property name="score" value="0.9987603425979614" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i41]" time="1.347"><properties><property name="score" value="0.10070693492889404" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i42]" time="2.469"><properties><property name="score" value="0.9277148842811584" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i43]" time="1.268"><properties><property name="score" value="0.9995073080062866" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i44]" time="2.480"><properties><property name="score" value="0.37132346630096436" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i45]" time="1.976"><properties><property name="score" value="0.8899807532628378" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i46]" time="1.565"><properties><property name="score" value="0.3044572671254476" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i47]" time="1.915"><properties><property name="score" value="0.9976347287495931" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i48]" time="1.736"><properties><property name="score" value="0.8850578864415487" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i49]" time="1.232"><properties><property name="score" value="0.9250558316707611" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i0]" time="1.312"><properties><property name="score" value="0.9984716375668844" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i1]" time="1.403"><properties><property name="score" value="0.3971804579099019" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.39718046&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.39718046
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i132]" time="0.229"><properties><property name="score" value="0.9983857870101929" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i133]" time="1.316"><properties><property name="score" value="0.987224280834198" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i134]" time="1.164"><properties><property name="score" value="0.9617966711521149" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i135]" time="2.379"><properties><property name="score" value="0.9376503109931946" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i136]" time="2.369"><properties><property name="score" value="0.49767154455184937" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i137]" time="0.951"><properties><property name="score" value="0.8600474298000336" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i138]" time="1.788"><properties><property name="score" value="0.9997978955507278" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i139]" time="1.767"><properties><property name="score" value="0.9528961628675461" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i140]" time="0.876"><properties><property name="score" value="0.9940315186977386" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i141]" time="1.924"><properties><property name="score" value="0.9997547566890717" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i142]" time="2.519"><properties><property name="score" value="0.9997892260551453" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i143]" time="0.749"><properties><property name="score" value="0.0908764898777008" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255144 (text: Don't trust fake news  Help journalists by signing) is an LLM-generated sample, misclassified as human-generated with confidence 0.09087649&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255144 (text: Don't trust fake news  Help journalists by signing) is an LLM-generated sample, misclassified as human-generated with confidence 0.09087649
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i2]" time="1.047"><properties><property name="score" value="0.21446457505226135" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.21446458&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.21446458
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i144]" time="0.327"><properties><property name="score" value="0.9998300075531006" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i145]" time="2.405"><properties><property name="score" value="0.9998059272766113" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i146]" time="0.230"><properties><property name="score" value="0.9998277425765991" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i147]" time="1.804"><properties><property name="score" value="0.9989557862281799" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i148]" time="1.912"><properties><property name="score" value="0.563925489783287" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i149]" time="2.897"><properties><property name="score" value="0.9793100595474243" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i150]" time="0.863"><properties><property name="score" value="0.9103664755821228" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255151 (text: Keyslinging. Driving fast on one side of a state l) is an LLM-generated sample, misclassified as human-generated with confidence 0.91036648&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255151 (text: Keyslinging. Driving fast on one side of a state l) is an LLM-generated sample, misclassified as human-generated with confidence 0.91036648
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i3]" time="1.107"><properties><property name="score" value="0.01565384864807129" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01565385&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01565385
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i151]" time="0.669"><properties><property name="score" value="0.7917794585227966" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i152]" time="2.347"><properties><property name="score" value="0.6380691647529602" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i153]" time="2.395"><properties><property name="score" value="0.6419726014137268" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i154]" time="2.381"><properties><property name="score" value="0.9997654557228088" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i155]" time="2.291"><properties><property name="score" value="0.9944765090942382" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i156]" time="0.239"><properties><property name="score" value="0.9938088059425354" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255157 (text: Lethe crocodiles are commonly found in Australia. ) is an LLM-generated sample, misclassified as human-generated with confidence 0.99380881&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255157 (text: Lethe crocodiles are commonly found in Australia. ) is an LLM-generated sample, misclassified as human-generated with confidence 0.99380881
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i4]" time="1.245"><properties><property name="score" value="0.008036643266677856" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i5]" time="1.523"><properties><property name="score" value="0.005597591400146484" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00559759&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00559759
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i157]" time="2.074"><properties><property name="score" value="0.9997461885213852" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i158]" time="0.549"><properties><property name="score" value="0.9988151788711548" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i159]" time="2.590"><properties><property name="score" value="0.1939717173576355" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255160 (text: Revealed: The Men Behind the Pain Dispensers You D) is an LLM-generated sample, misclassified as human-generated with confidence 0.19397172&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255160 (text: Revealed: The Men Behind the Pain Dispensers You D) is an LLM-generated sample, misclassified as human-generated with confidence 0.19397172
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i6]" time="2.445"><properties><property name="score" value="0.9818304777145386" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i7]" time="1.682"><properties><property name="score" value="0.9998103976249695" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i8]" time="1.592"><properties><property name="score" value="0.9982912540435791" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i9]" time="1.952"><properties><property name="score" value="0.11578366160392761" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i10]" time="1.064"><properties><property name="score" value="0.9109793305397034" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i11]" time="1.933"><properties><property name="score" value="0.9996596574783325" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i12]" time="1.434"><properties><property name="score" value="0.986138105392456" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i13]" time="1.583"><properties><property name="score" value="0.16550901532173157" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i14]" time="0.967"><properties><property name="score" value="0.9984333515167236" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i15]" time="1.307"><properties><property name="score" value="0.9917864998181661" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i16]" time="1.302"><properties><property name="score" value="0.32962626218795776" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i17]" time="1.626"><properties><property name="score" value="0.9974329173564911" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i18]" time="1.339"><properties><property name="score" value="0.9989084998766581" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.9989085&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.9989085
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i160]" time="3.263"><properties><property name="score" value="0.943439245223999" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i161]" time="0.300"><properties><property name="score" value="0.9633482098579407" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255162 (text: At the April 23, 2016, state, federal and local he) is an LLM-generated sample, misclassified as human-generated with confidence 0.96334821&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255162 (text: At the April 23, 2016, state, federal and local he) is an LLM-generated sample, misclassified as human-generated with confidence 0.96334821
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i19]" time="1.946"><properties><property name="score" value="0.5452145487070084" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i20]" time="0.551"><properties><property name="score" value="0.9980182647705078" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i21]" time="1.564"><properties><property name="score" value="0.33240509033203125" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i22]" time="0.792"><properties><property name="score" value="0.8025331497192383" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i23]" time="1.449"><properties><property name="score" value="0.4001869360605876" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i24]" time="0.590"><properties><property name="score" value="0.9768226444721222" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i25]" time="1.131"><properties><property name="score" value="0.9832465946674347" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i26]" time="1.283"><properties><property name="score" value="0.004813969135284424" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00481397&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00481397
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i162]" time="1.785"><properties><property name="score" value="0.4274246543645859" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i163]" time="2.955"><properties><property name="score" value="0.9993236660957336" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i164]" time="2.550"><properties><property name="score" value="0.5668031722307205" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i165]" time="1.274"><properties><property name="score" value="0.9808626572291056" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i166]" time="2.495"><properties><property name="score" value="0.6265226125717163" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i167]" time="2.693"><properties><property name="score" value="0.9997748374938965" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i168]" time="1.426"><properties><property name="score" value="0.23346378405888876" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i169]" time="2.971"><properties><property name="score" value="0.9981573462486267" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i170]" time="1.331"><properties><property name="score" value="0.2878863016764323" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i171]" time="2.346"><properties><property name="score" value="0.9503672361373902" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i172]" time="2.395"><properties><property name="score" value="0.9995421648025513" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i173]" time="2.399"><properties><property name="score" value="0.9980881810188293" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i174]" time="0.755"><properties><property name="score" value="0.9996338784694672" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i175]" time="0.736"><properties><property name="score" value="0.999573826789856" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i176]" time="0.472"><properties><property name="score" value="0.9500802159309387" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i177]" time="2.452"><properties><property name="score" value="0.9648275375366211" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i178]" time="0.973"><properties><property name="score" value="0.999620109796524" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i179]" time="2.728"><properties><property name="score" value="0.9997826933860778" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i180]" time="2.727"><properties><property name="score" value="0.9017046213150024" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i181]" time="0.366"><properties><property name="score" value="0.9679163694381714" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255182 (text: In the predawn darkness a massive world breathes f) is an LLM-generated sample, misclassified as human-generated with confidence 0.96791637&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255182 (text: In the predawn darkness a massive world breathes f) is an LLM-generated sample, misclassified as human-generated with confidence 0.96791637
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i27]" time="1.623"><properties><property name="score" value="0.9993326663970947" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i28]" time="0.748"><properties><property name="score" value="0.009102582931518555" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i29]" time="0.764"><properties><property name="score" value="0.9952973425388336" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i30]" time="1.100"><properties><property name="score" value="0.9985516369342804" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.99855164&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.99855164
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i182]" time="2.341"><properties><property name="score" value="0.48077039420604706" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i183]" time="2.787"><properties><property name="score" value="0.9714486002922058" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i184]" time="1.006"><properties><property name="score" value="0.7967279851436615" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i185]" time="1.107"><properties><property name="score" value="0.9840319156646729" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255186 (text: Whistleblower vant bracelet  It is a warm day. The) is an LLM-generated sample, misclassified as human-generated with confidence 0.98403192&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255186 (text: Whistleblower vant bracelet  It is a warm day. The) is an LLM-generated sample, misclassified as human-generated with confidence 0.98403192
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i31]" time="0.494"><properties><property name="score" value="0.8445727229118347" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.84457272&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.84457272
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i186]" time="2.723"><properties><property name="score" value="0.9997856418291727" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i187]" time="1.394"><properties><property name="score" value="0.09830552339553833" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255188 (text: In an effort to rectify the problem, the FA has no) is an LLM-generated sample, misclassified as human-generated with confidence 0.09830552&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255188 (text: In an effort to rectify the problem, the FA has no) is an LLM-generated sample, misclassified as human-generated with confidence 0.09830552
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i32]" time="1.413"><properties><property name="score" value="0.9981215000152588" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i33]" time="1.569"><properties><property name="score" value="0.9975699782371521" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i34]" time="1.438"><properties><property name="score" value="0.3351157307624817" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i35]" time="0.796"><properties><property name="score" value="0.9845669567584991" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.98456696&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.98456696
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i188]" time="4.199"><properties><property name="score" value="0.9997950077056885" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i189]" time="3.553"><properties><property name="score" value="0.9998015761375427" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i190]" time="0.211"><properties><property name="score" value="0.9903724193572998" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i191]" time="0.183"><properties><property name="score" value="0.8735101819038391" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i192]" time="2.378"><properties><property name="score" value="0.6104792952537537" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i193]" time="1.702"><properties><property name="score" value="0.9998200535774231" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i194]" time="0.941"><properties><property name="score" value="0.9570085108280182" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i195]" time="1.980"><properties><property name="score" value="0.5899756550788879" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i196]" time="3.708"><properties><property name="score" value="0.9997909545898438" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i197]" time="0.412"><properties><property name="score" value="0.9989433884620667" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255198 (text: Overall 7 Story 6 Animation 8 Sound 10 Character 7) is an LLM-generated sample, misclassified as human-generated with confidence 0.99894339&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255198 (text: Overall 7 Story 6 Animation 8 Sound 10 Character 7) is an LLM-generated sample, misclassified as human-generated with confidence 0.99894339
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i36]" time="1.466"><properties><property name="score" value="0.38238511482874554" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i37]" time="1.654"><properties><property name="score" value="0.3328786293665568" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i38]" time="1.127"><properties><property name="score" value="0.4947515328725179" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.49475153&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.49475153
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i198]" time="2.407"><properties><property name="score" value="0.9483169436454773" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i199]" time="1.625"><properties><property name="score" value="0.9997892181078593" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i200]" time="2.210"><properties><property name="score" value="0.31340405344963074" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i201]" time="0.645"><properties><property name="score" value="0.09789961576461792" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i202]" time="2.530"><properties><property name="score" value="0.9997805833816529" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i203]" time="0.924"><properties><property name="score" value="0.9996819496154785" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i204]" time="0.845"><properties><property name="score" value="0.011714041233062744" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i205]" time="2.438"><properties><property name="score" value="0.9973636070887247" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i206]" time="0.807"><properties><property name="score" value="0.9994505643844604" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i207]" time="1.224"><properties><property name="score" value="0.880280077457428" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i208]" time="2.394"><properties><property name="score" value="0.5489491701126099" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i209]" time="0.821"><properties><property name="score" value="0.9443168044090271" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i210]" time="1.909"><properties><property name="score" value="0.33543288707733154" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i211]" time="0.217"><properties><property name="score" value="0.9676945209503174" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255212 (text: n.  1550s, from Latin aegaeius &quot;he surmounted,&quot; of) is an LLM-generated sample, misclassified as human-generated with confidence 0.96769452&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255212 (text: n.  1550s, from Latin aegaeius "he surmounted," of) is an LLM-generated sample, misclassified as human-generated with confidence 0.96769452
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i39]" time="1.188"><properties><property name="score" value="0.2151894768079122" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i40]" time="1.879"><properties><property name="score" value="0.45227333903312683" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i41]" time="1.194"><properties><property name="score" value="0.05955237150192261" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i42]" time="1.030"><properties><property name="score" value="0.9998224973678589" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i43]" time="0.974"><properties><property name="score" value="0.9994767606258392" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i44]" time="1.514"><properties><property name="score" value="0.9659350713094076" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i45]" time="1.453"><properties><property name="score" value="0.2908060948053996" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i46]" time="1.405"><properties><property name="score" value="0.9991037249565125" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i47]" time="1.702"><properties><property name="score" value="0.9741453925768534" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i48]" time="1.230"><properties><property name="score" value="0.07056388258934021" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07056388&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07056388
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i212]" time="1.809"><properties><property name="score" value="0.9998127669095993" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i213]" time="2.559"><properties><property name="score" value="0.9871624410152435" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i214]" time="0.317"><properties><property name="score" value="0.8781769871711731" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i215]" time="2.039"><properties><property name="score" value="0.496430367231369" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i216]" time="2.560"><properties><property name="score" value="0.9924725413322448" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i217]" time="1.107"><properties><property name="score" value="0.4522971709569295" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i218]" time="0.507"><properties><property name="score" value="0.912216305732727" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255219 (text: Dilla, the late hip-hop star who passed away last ) is an LLM-generated sample, misclassified as human-generated with confidence 0.91221631&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255219 (text: Dilla, the late hip-hop star who passed away last ) is an LLM-generated sample, misclassified as human-generated with confidence 0.91221631
     assert 'Human' == 'AI'
       - AI
-      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i49]" time="1.032"><properties><property name="score" value="0.9994921684265137" /></properties></testcase></testsuite></testsuites>
\ No newline at end of file
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i219]" time="2.511"><properties><property name="score" value="0.6019073486328125" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i220]" time="2.656"><properties><property name="score" value="0.6397075533866883" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i221]" time="2.522"><properties><property name="score" value="0.8198766589164734" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i222]" time="2.896"><properties><property name="score" value="0.9996848255395889" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i223]" time="0.467"><properties><property name="score" value="0.9936820268630981" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i224]" time="0.853"><properties><property name="score" value="0.9597063660621643" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i225]" time="0.781"><properties><property name="score" value="0.7497139275074005" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i226]" time="2.892"><properties><property name="score" value="0.999777638912201" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i227]" time="0.209"><properties><property name="score" value="0.999313235282898" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i228]" time="1.979"><properties><property name="score" value="0.9997290074825287" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i229]" time="1.612"><properties><property name="score" value="0.9804417689641317" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i230]" time="0.415"><properties><property name="score" value="0.9998247027397156" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i231]" time="1.439"><properties><property name="score" value="0.8327985207239786" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i232]" time="1.013"><properties><property name="score" value="0.9934880435466766" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i233]" time="1.273"><properties><property name="score" value="0.38267238934834796" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i234]" time="3.044"><properties><property name="score" value="0.2462162733078003" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i235]" time="0.365"><properties><property name="score" value="0.7097145915031433" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255236 (text: No, I don't hate kitchen memberships. Bread buns i) is an LLM-generated sample, misclassified as human-generated with confidence 0.70971459&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255236 (text: No, I don't hate kitchen memberships. Bread buns i) is an LLM-generated sample, misclassified as human-generated with confidence 0.70971459
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i236]" time="2.671"><properties><property name="score" value="0.6861971974372864" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i237]" time="2.137"><properties><property name="score" value="0.9997666329145432" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i238]" time="1.578"><properties><property name="score" value="0.2202665607134501" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i239]" time="2.776"><properties><property name="score" value="0.9461084485054017" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i240]" time="1.329"><properties><property name="score" value="0.9977206389109293" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i241]" time="0.524"><properties><property name="score" value="0.05372738838195801" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i242]" time="0.410"><properties><property name="score" value="0.9998160004615784" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i243]" time="2.034"><properties><property name="score" value="0.49255727231502533" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i244]" time="1.347"><properties><property name="score" value="0.9996853073438009" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i245]" time="2.641"><properties><property name="score" value="0.998547206322352" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i246]" time="0.271"><properties><property name="score" value="0.9079344272613525" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i247]" time="2.509"><properties><property name="score" value="0.9997824549674987" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i248]" time="2.035"><properties><property name="score" value="0.008699730038642883" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i249]" time="2.588"><properties><property name="score" value="0.6232220729192098" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i250]" time="2.122"><properties><property name="score" value="0.928134098649025" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i251]" time="0.382"><properties><property name="score" value="0.9995962977409363" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i252]" time="0.258"><properties><property name="score" value="0.984184741973877" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255253 (text: And so the pages turn, and I find myself at a loss) is an LLM-generated sample, misclassified as human-generated with confidence 0.98418474&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255253 (text: And so the pages turn, and I find myself at a loss) is an LLM-generated sample, misclassified as human-generated with confidence 0.98418474
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i253]" time="0.444"><properties><property name="score" value="0.7527053952217102" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255254 (text: fb 256×176 with Sheena Shaw and Darryl Jenkins. Fl) is an LLM-generated sample, misclassified as human-generated with confidence 0.7527054&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255254 (text: fb 256×176 with Sheena Shaw and Darryl Jenkins. Fl) is an LLM-generated sample, misclassified as human-generated with confidence 0.7527054
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i254]" time="2.391"><properties><property name="score" value="0.9998144507408142" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i255]" time="2.876"><properties><property name="score" value="0.2369483232498169" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i256]" time="2.321"><properties><property name="score" value="0.46763116121292114" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i257]" time="2.403"><properties><property name="score" value="0.5864988803863526" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i258]" time="2.423"><properties><property name="score" value="0.6036865472793579" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i259]" time="0.807"><properties><property name="score" value="0.9657982587814331" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i260]" time="2.401"><properties><property name="score" value="0.9565617084503174" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i261]" time="1.427"><properties><property name="score" value="0.374630610148112" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i262]" time="2.278"><properties><property name="score" value="0.9464545726776123" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i263]" time="2.794"><properties><property name="score" value="0.26969200372695923" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i264]" time="4.295"><properties><property name="score" value="0.5721692204475403" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i265]" time="3.974"><properties><property name="score" value="0.5307288765907288" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i266]" time="0.330"><properties><property name="score" value="0.9958112239837646" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i267]" time="0.948"><properties><property name="score" value="0.8579023778438568" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i268]" time="1.995"><properties><property name="score" value="0.999785840511322" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i269]" time="0.717"><properties><property name="score" value="0.21814966201782227" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i270]" time="0.376"><properties><property name="score" value="0.9998165965080261" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i271]" time="3.102"><properties><property name="score" value="0.9997609376907348" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i272]" time="1.624"><properties><property name="score" value="0.07995724678039551" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i273]" time="2.897"><properties><property name="score" value="0.9985660076141357" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i274]" time="0.453"><properties><property name="score" value="0.7341034412384033" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i275]" time="0.925"><properties><property name="score" value="0.9996973276138306" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255276 (text: Tracklist  01. Virtual Riot – VLA (extended cut)  ) is an LLM-generated sample, misclassified as human-generated with confidence 0.99969733&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255276 (text: Tracklist  01. Virtual Riot – VLA (extended cut)  ) is an LLM-generated sample, misclassified as human-generated with confidence 0.99969733
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i276]" time="0.248"><properties><property name="score" value="0.9998045563697815" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i277]" time="2.774"><properties><property name="score" value="0.9961873650550842" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i278]" time="2.904"><properties><property name="score" value="0.36395447254180907" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i279]" time="2.775"><properties><property name="score" value="0.2966157913208008" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i280]" time="2.735"><properties><property name="score" value="0.9792868256568908" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i281]" time="1.099"><properties><property name="score" value="0.12133944034576416" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255282 (text: The ATG-ACE 1000-F was an upgrade of the ATG-M10A/) is an LLM-generated sample, misclassified as human-generated with confidence 0.12133944&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255282 (text: The ATG-ACE 1000-F was an upgrade of the ATG-M10A/) is an LLM-generated sample, misclassified as human-generated with confidence 0.12133944
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i282]" time="2.210"><properties><property name="score" value="0.5194753259420395" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i283]" time="0.267"><properties><property name="score" value="0.9936036467552185" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i284]" time="2.686"><properties><property name="score" value="0.9519603610038757" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i285]" time="1.069"><properties><property name="score" value="0.9997805953025818" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i286]" time="0.153"><properties><property name="score" value="0.8191166520118713" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i287]" time="2.973"><properties><property name="score" value="0.9997820615768432" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i288]" time="2.550"><properties><property name="score" value="0.957938838005066" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i289]" time="2.663"><properties><property name="score" value="0.9991058856248856" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i290]" time="2.814"><properties><property name="score" value="0.6003148317337036" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i291]" time="2.632"><properties><property name="score" value="0.16980974674224852" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255292 (text: and so far, they've done a pretty amazing job.  La) is an LLM-generated sample, misclassified as human-generated with confidence 0.16980975&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255292 (text: and so far, they've done a pretty amazing job.  La) is an LLM-generated sample, misclassified as human-generated with confidence 0.16980975
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i292]" time="1.958"><properties><property name="score" value="0.9998079339663187" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i293]" time="2.901"><properties><property name="score" value="0.968584394454956" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i294]" time="0.377"><properties><property name="score" value="0.9983787536621094" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i295]" time="0.954"><properties><property name="score" value="0.9671914875507355" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i296]" time="2.324"><properties><property name="score" value="0.6072497367858887" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i297]" time="2.624"><properties><property name="score" value="0.6278205633163452" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i298]" time="0.580"><properties><property name="score" value="0.9920861124992371" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i299]" time="2.908"><properties><property name="score" value="0.9998195767402649" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i300]" time="0.601"><properties><property name="score" value="0.9989314675331116" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i301]" time="0.197"><properties><property name="score" value="0.5097333788871765" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i302]" time="0.797"><properties><property name="score" value="0.7438998222351074" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i303]" time="2.827"><properties><property name="score" value="0.9997888088226319" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i304]" time="0.686"><properties><property name="score" value="0.9955058097839355" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i305]" time="0.867"><properties><property name="score" value="0.7500880360603333" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i306]" time="2.655"><properties><property name="score" value="0.9928252220153808" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i307]" time="2.678"><properties><property name="score" value="0.9997792482376099" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i308]" time="2.796"><properties><property name="score" value="0.5994310855865479" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i309]" time="2.668"><properties><property name="score" value="0.9415273547172547" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i310]" time="1.718"><properties><property name="score" value="0.9997747540473938" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i311]" time="2.244"><properties><property name="score" value="0.5688995718955994" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i312]" time="2.365"><properties><property name="score" value="0.9963664889335633" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i313]" time="3.048"><properties><property name="score" value="0.9145281076431274" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i314]" time="2.742"><properties><property name="score" value="0.999805235862732" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i315]" time="1.515"><properties><property name="score" value="0.9998070597648621" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i316]" time="2.752"><properties><property name="score" value="0.9997460603713989" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i317]" time="0.271"><properties><property name="score" value="0.9651115536689758" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i318]" time="2.717"><properties><property name="score" value="0.999773633480072" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i319]" time="1.822"><properties><property name="score" value="0.9956425825754801" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i320]" time="2.694"><properties><property name="score" value="0.9996456384658814" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i321]" time="2.632"><properties><property name="score" value="0.9849258661270142" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i322]" time="2.626"><properties><property name="score" value="0.6162702679634094" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i323]" time="0.973"><properties><property name="score" value="0.9899247884750366" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i324]" time="0.424"><properties><property name="score" value="0.9899457097053528" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255325 (text: The only reason you might want to create a PDF wit) is an LLM-generated sample, misclassified as human-generated with confidence 0.98994571&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255325 (text: The only reason you might want to create a PDF wit) is an LLM-generated sample, misclassified as human-generated with confidence 0.98994571
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i325]" time="2.731"><properties><property name="score" value="0.2716928243637085" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i326]" time="2.640"><properties><property name="score" value="0.9105450272560119" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i327]" time="2.640"><properties><property name="score" value="0.9997518301010132" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i328]" time="2.159"><properties><property name="score" value="0.49858940839767457" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i329]" time="2.688"><properties><property name="score" value="0.9791828513145446" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i330]" time="2.334"><properties><property name="score" value="0.9998061060905457" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i331]" time="0.032"><failure message="RuntimeError: The expanded size of the tensor (999) must match the existing size (514) at non-singleton dimension 1.  Target sizes: [1, 999].  Tensor sizes: [1, 514]">E   RuntimeError: The expanded size of the tensor (999) must match the existing size (514) at non-singleton dimension 1.  Target sizes: [1, 999].  Tensor sizes: [1, 514]</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i332]" time="0.277"><properties><property name="score" value="0.9959239959716797" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i333]" time="0.917"><properties><property name="score" value="0.9998233318328857" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i334]" time="1.771"><properties><property name="score" value="0.8984413743019104" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i335]" time="2.298"><properties><property name="score" value="0.9205249190330506" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i336]" time="2.307"><properties><property name="score" value="0.5517356991767883" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i337]" time="0.819"><properties><property name="score" value="0.960362046957016" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i338]" time="1.101"><properties><property name="score" value="0.8785881102085114" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i339]" time="0.471"><properties><property name="score" value="0.9998230338096619" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i340]" time="0.211"><properties><property name="score" value="0.9940239787101746" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i341]" time="2.792"><properties><property name="score" value="0.9997658610343934" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i342]" time="0.876"><properties><property name="score" value="0.0016689896583557129" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i343]" time="1.095"><properties><property name="score" value="0.9960366487503052" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i344]" time="0.497"><properties><property name="score" value="0.9987975358963013" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i345]" time="2.865"><properties><property name="score" value="0.9996509253978729" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i346]" time="2.544"><properties><property name="score" value="0.9048105001449585" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i347]" time="0.508"><properties><property name="score" value="0.8847047686576843" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i348]" time="1.545"><properties><property name="score" value="0.993595818678538" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i349]" time="2.811"><properties><property name="score" value="0.9767492890357972" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i350]" time="0.923"><properties><property name="score" value="0.9965750277042389" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i351]" time="0.812"><properties><property name="score" value="0.08613896369934082" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i352]" time="2.693"><properties><property name="score" value="0.605113399028778" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i353]" time="0.613"><properties><property name="score" value="0.9989084005355835" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i354]" time="1.558"><properties><property name="score" value="0.9998050729433695" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i355]" time="1.467"><properties><property name="score" value="0.316456397374471" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255356 (text: Last season was one that really showed the charact) is an LLM-generated sample, misclassified as human-generated with confidence 0.3164564&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255356 (text: Last season was one that really showed the charact) is an LLM-generated sample, misclassified as human-generated with confidence 0.3164564
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i356]" time="0.439"><properties><property name="score" value="0.9997888207435608" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i357]" time="0.259"><properties><property name="score" value="0.99970942735672" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i358]" time="0.420"><properties><property name="score" value="0.9196286797523499" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i359]" time="3.009"><properties><property name="score" value="0.9052213549613952" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i360]" time="2.989"><properties><property name="score" value="0.9356784820556641" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i361]" time="0.467"><properties><property name="score" value="0.9880213737487793" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i362]" time="2.592"><properties><property name="score" value="0.5268798768520355" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i363]" time="1.987"><properties><property name="score" value="0.8674966891606649" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i364]" time="3.059"><properties><property name="score" value="0.9997783303260803" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i365]" time="2.673"><properties><property name="score" value="0.9971312433481216" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i366]" time="2.535"><properties><property name="score" value="0.6016250133514405" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i367]" time="0.310"><properties><property name="score" value="0.9998151659965515" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i368]" time="2.537"><properties><property name="score" value="0.9971346378326416" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i369]" time="2.267"><properties><property name="score" value="0.9998226404190064" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i370]" time="2.436"><properties><property name="score" value="0.9272860646247864" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i371]" time="2.378"><properties><property name="score" value="0.9997862100601196" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i372]" time="0.906"><properties><property name="score" value="0.9997847080230713" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i373]" time="1.278"><properties><property name="score" value="0.8570610682169596" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i374]" time="2.652"><properties><property name="score" value="0.9995952606201172" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i375]" time="1.580"><properties><property name="score" value="0.8204321463902792" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i376]" time="0.386"><properties><property name="score" value="0.9988147020339966" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i377]" time="0.898"><properties><property name="score" value="0.9239495396614075" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i378]" time="0.733"><properties><property name="score" value="0.7635477781295776" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i379]" time="2.356"><properties><property name="score" value="0.12390589714050293" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i380]" time="0.914"><properties><property name="score" value="0.9998095333576202" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i381]" time="0.674"><properties><property name="score" value="0.8703460097312927" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i382]" time="2.405"><properties><property name="score" value="0.9974981069564819" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i383]" time="0.732"><properties><property name="score" value="0.9962541460990906" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i384]" time="0.212"><properties><property name="score" value="0.9626439213752747" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255385 (text: Ginaldo Rivera, president of the Miami-Dade County) is an LLM-generated sample, misclassified as human-generated with confidence 0.96264392&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255385 (text: Ginaldo Rivera, president of the Miami-Dade County) is an LLM-generated sample, misclassified as human-generated with confidence 0.96264392
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i385]" time="2.427"><properties><property name="score" value="0.9997683644294739" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i386]" time="0.145"><properties><property name="score" value="0.9329478144645691" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255387 (text: BE INITIATED  Today is our annual pre-500th showdo) is an LLM-generated sample, misclassified as human-generated with confidence 0.93294781&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255387 (text: BE INITIATED  Today is our annual pre-500th showdo) is an LLM-generated sample, misclassified as human-generated with confidence 0.93294781
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i387]" time="0.363"><properties><property name="score" value="0.999808132648468" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i388]" time="2.414"><properties><property name="score" value="0.8842490553855896" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i389]" time="0.670"><properties><property name="score" value="0.9997976422309875" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i390]" time="0.515"><properties><property name="score" value="0.9998201727867126" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i391]" time="0.469"><properties><property name="score" value="0.9998268485069275" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i392]" time="2.247"><properties><property name="score" value="0.6144865900278091" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i393]" time="0.947"><properties><property name="score" value="0.9993368089199066" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i394]" time="0.909"><properties><property name="score" value="0.009654849767684937" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i395]" time="0.429"><properties><property name="score" value="0.9994434714317322" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i396]" time="0.408"><properties><property name="score" value="0.9998114705085754" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i397]" time="2.364"><properties><property name="score" value="0.9988821506500244" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i398]" time="1.686"><properties><property name="score" value="0.9914633830388387" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i399]" time="1.782"><properties><property name="score" value="0.9989510178565979" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i400]" time="1.271"><properties><property name="score" value="0.2638833522796631" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255401 (text: A Baki National Park official is covered in a plas) is an LLM-generated sample, misclassified as human-generated with confidence 0.26388335&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255401 (text: A Baki National Park official is covered in a plas) is an LLM-generated sample, misclassified as human-generated with confidence 0.26388335
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i401]" time="1.382"><properties><property name="score" value="0.44896437724431354" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i402]" time="0.408"><properties><property name="score" value="0.9997561573982239" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i403]" time="2.352"><properties><property name="score" value="0.999804417292277" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i404]" time="2.599"><properties><property name="score" value="0.9996833801269531" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i405]" time="0.235"><properties><property name="score" value="0.9996873140335083" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i406]" time="0.408"><properties><property name="score" value="0.9998157620429993" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i407]" time="0.661"><properties><property name="score" value="0.01550382375717163" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i408]" time="0.744"><properties><property name="score" value="0.9904817342758179" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i409]" time="0.260"><properties><property name="score" value="0.9997020363807678" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i410]" time="0.564"><properties><property name="score" value="0.983874499797821" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i411]" time="2.455"><properties><property name="score" value="0.993937861919403" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i412]" time="0.678"><properties><property name="score" value="0.9997487366199493" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i413]" time="0.366"><properties><property name="score" value="0.7568855285644531" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i414]" time="3.023"><properties><property name="score" value="0.9164233058691025" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i415]" time="1.371"><properties><property name="score" value="0.00849241018295288" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i416]" time="3.198"><properties><property name="score" value="0.9986894965171814" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i417]" time="0.241"><properties><property name="score" value="0.7175073623657227" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255418 (text: There are a lot of rumours going around about the ) is an LLM-generated sample, misclassified as human-generated with confidence 0.71750736&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255418 (text: There are a lot of rumours going around about the ) is an LLM-generated sample, misclassified as human-generated with confidence 0.71750736
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i418]" time="1.153"><properties><property name="score" value="0.40948204199473065" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i419]" time="2.531"><properties><property name="score" value="0.6415911674499511" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i420]" time="0.704"><properties><property name="score" value="0.9994630217552185" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i421]" time="2.417"><properties><property name="score" value="0.9997971057891846" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i422]" time="1.497"><properties><property name="score" value="0.9997176975011826" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i423]" time="2.465"><properties><property name="score" value="0.9599440544843674" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i424]" time="2.488"><properties><property name="score" value="0.4935627381006877" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i425]" time="0.754"><properties><property name="score" value="0.963879406452179" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255426 (text: The real question at this point is, can China drop) is an LLM-generated sample, misclassified as human-generated with confidence 0.96387941&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255426 (text: The real question at this point is, can China drop) is an LLM-generated sample, misclassified as human-generated with confidence 0.96387941
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i426]" time="1.598"><properties><property name="score" value="0.497709721326828" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i427]" time="0.818"><properties><property name="score" value="0.9962062835693359" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i428]" time="2.530"><properties><property name="score" value="0.5974959015846253" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i429]" time="2.789"><properties><property name="score" value="0.973056948184967" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i430]" time="2.679"><properties><property name="score" value="0.9421170204877853" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i431]" time="2.667"><properties><property name="score" value="0.9596379041671753" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i432]" time="0.640"><properties><property name="score" value="0.8552316427230835" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i433]" time="0.783"><properties><property name="score" value="0.14095017313957214" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i434]" time="0.211"><properties><property name="score" value="0.941279411315918" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255435 (text: 2015-08-20T13:38:59Z (GMT) William Frommer by (GMT) is an LLM-generated sample, misclassified as human-generated with confidence 0.94127941&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255435 (text: 2015-08-20T13:38:59Z (GMT) William Frommer by (GMT) is an LLM-generated sample, misclassified as human-generated with confidence 0.94127941
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i435]" time="2.465"><properties><property name="score" value="0.9997938573360443" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i436]" time="2.395"><properties><property name="score" value="0.6520919998486837" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i437]" time="2.333"><properties><property name="score" value="0.5293837547302246" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i438]" time="0.793"><properties><property name="score" value="0.9998171329498291" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i439]" time="1.269"><properties><property name="score" value="0.9860149820645651" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i440]" time="0.879"><properties><property name="score" value="0.999770998954773" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i441]" time="0.137"><properties><property name="score" value="0.8451054692268372" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255442 (text: Doesn't the fear just indicate that he's aggravate) is an LLM-generated sample, misclassified as human-generated with confidence 0.84510547&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255442 (text: Doesn't the fear just indicate that he's aggravate) is an LLM-generated sample, misclassified as human-generated with confidence 0.84510547
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i442]" time="0.983"><properties><property name="score" value="0.0002257227897644043" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i443]" time="0.222"><properties><property name="score" value="0.9971256852149963" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255444 (text: Adrian Peterson is one of the greatest running bac) is an LLM-generated sample, misclassified as human-generated with confidence 0.99712569&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255444 (text: Adrian Peterson is one of the greatest running bac) is an LLM-generated sample, misclassified as human-generated with confidence 0.99712569
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i444]" time="2.471"><properties><property name="score" value="0.9436723232269287" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i445]" time="2.327"><properties><property name="score" value="0.9689479827880859" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i446]" time="0.745"><properties><property name="score" value="0.002742975950241089" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i447]" time="0.376"><properties><property name="score" value="0.9975087642669678" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i448]" time="2.546"><properties><property name="score" value="0.9982641935348511" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i449]" time="1.130"><properties><property name="score" value="0.9095407128334045" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i450]" time="1.235"><properties><property name="score" value="0.3990402619043986" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i451]" time="0.155"><properties><property name="score" value="0.9750450253486633" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i452]" time="1.592"><properties><property name="score" value="0.8511840403079987" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i453]" time="0.409"><properties><property name="score" value="0.9997560381889343" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i454]" time="2.368"><properties><property name="score" value="0.9982591152191163" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i455]" time="1.002"><properties><property name="score" value="0.1610485315322876" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i456]" time="2.824"><properties><property name="score" value="0.9997801184654236" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i457]" time="1.614"><properties><property name="score" value="0.9945527116457621" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i458]" time="1.627"><properties><property name="score" value="0.39350426197052" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i459]" time="0.695"><properties><property name="score" value="0.9992922842502594" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i460]" time="1.682"><properties><property name="score" value="0.9997546275456747" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i461]" time="2.351"><properties><property name="score" value="0.1693040370941162" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255462 (text: Defection versus providing cover  Machiavelli dire) is an LLM-generated sample, misclassified as human-generated with confidence 0.16930404&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255462 (text: Defection versus providing cover  Machiavelli dire) is an LLM-generated sample, misclassified as human-generated with confidence 0.16930404
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i462]" time="0.399"><properties><property name="score" value="0.999813973903656" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i463]" time="2.362"><properties><property name="score" value="0.9961378931999206" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i464]" time="2.285"><properties><property name="score" value="0.6257474064826966" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i465]" time="0.194"><properties><property name="score" value="0.7319552898406982" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255466 (text: Winning a gold medal is a huge accomplishment in o) is an LLM-generated sample, misclassified as human-generated with confidence 0.73195529&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255466 (text: Winning a gold medal is a huge accomplishment in o) is an LLM-generated sample, misclassified as human-generated with confidence 0.73195529
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i466]" time="2.277"><properties><property name="score" value="0.9983927249908447" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i467]" time="0.721"><properties><property name="score" value="0.07397279143333435" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i468]" time="2.128"><properties><property name="score" value="0.9845006585121154" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i469]" time="2.459"><properties><property name="score" value="0.999148964881897" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i470]" time="0.349"><properties><property name="score" value="0.7805707454681396" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255471 (text: DISCORD WHO ARE YOU?  Who Am I?  Are You Still For) is an LLM-generated sample, misclassified as human-generated with confidence 0.78057075&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255471 (text: DISCORD WHO ARE YOU?  Who Am I?  Are You Still For) is an LLM-generated sample, misclassified as human-generated with confidence 0.78057075
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i471]" time="0.271"><properties><property name="score" value="0.5815640687942505" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i472]" time="0.164"><properties><property name="score" value="0.9996732473373413" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i473]" time="2.655"><properties><property name="score" value="0.925062096118927" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i474]" time="2.751"><properties><property name="score" value="0.5409912705421448" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i475]" time="2.412"><properties><property name="score" value="0.999428129196167" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i476]" time="2.425"><properties><property name="score" value="0.9997761964797973" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i477]" time="2.472"><properties><property name="score" value="0.9997556328773498" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i478]" time="1.191"><properties><property name="score" value="0.9997907082239786" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i479]" time="1.953"><properties><property name="score" value="0.8715042024850845" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i480]" time="0.795"><properties><property name="score" value="0.8485717177391052" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i481]" time="0.978"><properties><property name="score" value="0.9307777881622314" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255482 (text: Doors: 6:00 pm / Show: 7:00 pm  Show Tickets (upon) is an LLM-generated sample, misclassified as human-generated with confidence 0.93077779&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255482 (text: Doors: 6:00 pm / Show: 7:00 pm  Show Tickets (upon) is an LLM-generated sample, misclassified as human-generated with confidence 0.93077779
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i482]" time="0.298"><properties><property name="score" value="0.7093763947486877" /></properties><failure message="AssertionError: samples/xl-1542M.test.jsonl:255483 (text: Description  The first track of the Deadshot EFX G) is an LLM-generated sample, misclassified as human-generated with confidence 0.70937639&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/xl-1542M.test.jsonl:255483 (text: Description  The first track of the Deadshot EFX G) is an LLM-generated sample, misclassified as human-generated with confidence 0.70937639
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i483]" time="0.878"><properties><property name="score" value="0.9962012469768524" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i484]" time="1.753"><properties><property name="score" value="0.9997693300247192" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i485]" time="2.184"><properties><property name="score" value="0.5244623720645905" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i486]" time="2.445"><properties><property name="score" value="0.892016065120697" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i487]" time="2.746"><properties><property name="score" value="0.9996933460235595" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i488]" time="2.419"><properties><property name="score" value="0.44578053057193756" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i489]" time="0.220"><properties><property name="score" value="0.998384952545166" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i490]" time="1.426"><properties><property name="score" value="0.999548594156901" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i491]" time="0.471"><properties><property name="score" value="0.8805862069129944" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i492]" time="2.563"><properties><property name="score" value="0.9997556805610657" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i493]" time="2.616"><properties><property name="score" value="0.9997692823410034" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i494]" time="2.622"><properties><property name="score" value="0.9989821314811707" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i495]" time="1.707"><properties><property name="score" value="0.49954646825790405" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i496]" time="2.512"><properties><property name="score" value="0.9984587669372559" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i497]" time="0.625"><properties><property name="score" value="0.805763304233551" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i498]" time="2.052"><properties><property name="score" value="0.9980788975954056" /></properties></testcase><testcase classname="test_roberta_detect" name="test_llm_jsonl[i499]" time="2.648"><properties><property name="score" value="0.9875650405883789" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Glacier Ridge Christian School\n\nGlacier Ridge Christian School is a private Christian school in Johnstown, Ohio. It was founded in the fall of 1999 by Gary and Tammy Smith.\n\nThe school started with just 13 students in grades 5-8. By the end of the first year, the school had grown to 65 students in those same grades. As the school has continued to grow, so has the number of teachers.\n\nBy the year 2009, Glacier Ridge had expanded to offer a preschool through 12th grade education, and the school had grown to almost 200 students. In the fall of 2013, the enrollment was over 300 students.\n\nThe school started with the first graduating class of 2002. They are now on their 11th graduating class, and the school's graduates continue to receive acceptance to universities across the United States. The school offers several Advanced Placement classes.\n\nGlacier Ridge has had its students participate in various competitions and competitions. Most notable are the sports teams, especially their basketball teams, which have qualified for state competition multiple times. The school has also had various music groups participate in state and regional competitions, and in the 2010-2011 school year, the school sent students to participate in an International Children's Games competition in Romania.\n\nIn the 2010-2011 school year, a group of the school's teachers were sent to Haiti to help establish a new orphanage, &quot;Eli's Children's House&quot;, which was dedicated to Eli Pierre, who had died in a motorcycle accident. They were instrumental in bringing the orphanage into the community by helping establish new programs for the children, including English as a Second Language, Computer Lab, and Physical Education classes. The teachers assisted in the education of the orphans as well as working with the staff on the property.\n\nThe school has also taken various trips over the years, including trips to Virginia and New York City. The school also hosted an educational trip to Washington D.C. for one of their graduating classes.\n\nThe school has participated in various community events and service projects. One of the most notable is their annual &quot;Twilight Prom&quot; that is held each spring. Each year, the school hosts a prom for local high school students who have special needs. The students' dates are typically their own siblings.\n\nOn March 21, 2011, a fire destroyed the school's auditorium, causing over $1 million in damages. After the school received an outpouring of support, the school was able to begin construction on a new auditorium in time for the next school year.\n\nIn 2015, Glacier Ridge Christian School became part of the Christian School Association of Northeast Ohio.\n\nThe School is affiliated with the Association of Christian Schools International, the Ohio Christian Education Association, and the Christian School Association of Northeast Ohio.\n\n]" time="1.354"><properties><property name="score" value="0.999787708123525" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[2003 NCAA Division I-AA football season\n\nThe 2003 NCAA Division I-AA football season, part of college football in the United States organized by the National Collegiate Athletic Association at the Division I Football Championship Subdivision level, began on September 6, 2003, and concluded with the National Championship Game of the NCAA Division I Football Championship on December 21, 2003, at Finley Stadium in Chattanooga, Tennessee. The Georgia Southern Eagles won their first I-AA championship, defeating the Furman Paladins by a final score of 27\u221210. \n\n&lt;nowiki&gt;*&lt;/nowiki&gt; &quot;Denotes host institution&quot;\n\n&lt;nowiki&gt;*&lt;/nowiki&gt; &quot;Denotes host institution&quot;\n\n&lt;nowiki&gt;*&lt;/nowiki&gt; &quot;Denotes host institution&quot;\n]" time="0.358"><properties><property name="score" value="0.9993891716003418" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99938917&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99938917
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Adiadochokinesia\n\nAdiadochokinesia, or more fully dysdiadochokinesia, is a neurologic symptom, described as the inability to perform rapid alternating movements, as seen in ataxia.\n\nExamples of this are \n\nA classic test is &quot;the finger to nose test&quot;. The examiner repeatedly taps the patient's finger with the examiner's finger and asks the patient to reciprocate in the same rhythm.\n\nThis should be distinguished from dysdiadochokinesia, a progressive condition involving more or less the same symptoms.\n]" time="0.315"><properties><property name="score" value="0.9471513628959656" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.94715136&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.94715136
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Akyem Kotoku\n\nAkyem Kotoku is one of the five traditional districts of the Ashanti Empire. The people are collectively known as Akyem Kotokus.\n\nAkyem Kotoku lies in the Eastern region of Ashanti and is bounded by the East Mamprusi to the East, Nkwantanan to the North, Ashanti to the West and Ahanta West to the South. It is part of Ashanti Region. The area is mainly mountainous with some rich valleys and hills. Akyem Kotoku is known as the &quot;Area of Plenty&quot;.\n\nThe people are mainly farmers and fishermen. Its capital is Akyem Kotoku. It has two paramountcies, Buokrom and Kumasi. Buokrom, the newest paramountcy, was established in 1975. The Akyem Kotoku Traditional Council has nine chiefs.\n\nThe major towns are Abokobi, Domenase, Akropong, and Nsuaem. Other settlements include Aduanful, Boankraful, Denkyemerem, Anhwiaso, Gyankroma, Kwabenya, Nananom, Obofram, Okwensu, Ofoase, Nkoranza, Nsutam, Aduaso, Adom, Ofram, Odomankoma, Asafo, Anyigbe, Obofram, Kwabenya, Denkyemerem, Odumase, Nsutam, Aduaso, Nsuaem, Nsuaso, Tafo-Nsuaem, Manhyia, Boaso, Adom, Ofram, Anhwiaso, Adoase, Anyigbe, Boankraful, Okwensu, Ofoase, Nkoranza, Akropong, Gyankroma, Domenase, Egyir, Afram, Denkyerem, Nananom, Akuapem South, Asafo, Aboadze and Koma.\n\nThe Akyem Kotoku Traditional Council has nine paramountcies.\n\nThe Akyem Kotoku people have two coats of arms, one to use at home and the other for international use.\n\nThe arms used at home is a gold shield with a black anchor in the center. On the shield is a white dove with a golden olive branch in its beak. The dove is supported by two black wild dogs facing each other.\n\nThe shield is surmounted by a gold mural crown with five points and black stripes.\n\nThe shield is supported by two black lions standing on a white scroll. The scroll bears the motto: &quot;MUSTER FAITHFULNESS&quot; (Cape Coast Latin: &quot;Fidelitas Summa Est&quot;).\n\nThe arms used outside the country is a gold shield with a black ram's head in the center. On the shield is a green laurel wreath.\n\nThe shield is surmounted by a gold mural crown with five points and black stripes.\n\nThe shield is supported by two black and white wild dogs standing on a white scroll. The scroll bears the motto: &quot;MUSTER FAITHFULNESS&quot; (Cape Coast Latin: &quot;Fidelitas Summa Est&quot;).\n\nThe two dogs represent vigilance. The two lions in the home arms represent fortitude and strength.\n\n]" time="1.653"><properties><property name="score" value="0.17941135168075562" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Aulus Atilius Calatinus\n\nAulus Atilius Calatinus was consul of the Roman Republic in 221 BC with Gaius Claudius Nero. The primary action of the consulship was the siege and capture of the Volscian city of Suessa Pometia.\n\nIn the same year he was sent to Tarentum with a fleet of 10,000 infantry and 800 cavalry to aid the Tarentines in the Second War with the Bruttians. Calatinus was captured in a naval action off the coast near Tarentum, and the Bruttian admiral Lucerius was rewarded with his command for this victory. The Roman troops under his command, however, escaped to Thurii.\n\nThe earlier Aulus Atilius Calatinus was elected consul in 222 BC, and killed in a battle with the Samnites.\n\n]" time="0.493"><properties><property name="score" value="0.9989840388298035" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Frederic Myers\n\nFrederic William Henry Myers (15 January 1843 \u2013 18 January 1901) was a classical scholar, poet, playwright, novelist, and the founder of the Society for Psychical Research, which he led until his death.\n\nFrederic Myers was born at Kew, Surrey, the eldest son of Frederic Myers, a surgeon. He was educated at King's College School, and at St Paul's School in London. After a year at Trinity College, Cambridge (1861\u20132), he was a schoolmaster at Rugby from 1862 to 1871. In 1872 he became classical lecturer at University College, London. He became a member of the Council of the College (now part of UCL), and was appointed professor of Latin there in 1880. In 1876 Myers joined the Metaphysical Society, along with Alfred Tennyson, J. M. Barrie and others. His friendship with Barrie lasted for the remainder of his life, and Myers even accompanied Barrie and his wife on their honeymoon. Myers and Tennyson became close friends and met frequently in later years, as seen in their correspondence.\n\nMyers wrote prose and verse for periodicals, including a parody of Gerard Manley Hopkins, and some detective stories, and contributed to books of poetry and belles lettres, such as &quot;Whistler and His Art&quot; and the &quot;Yellow Book&quot;. His poetry was praised by Henry James, who]" time="0.762"><properties><property name="score" value="0.9790789484977722" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.97907895&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.97907895
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[His father was a professor and his mother was an economist. While growing up, he was exposed to the inspiring stories of his parents and their family members in China.\n\n\n\n\n\nKoh attended the University of Michigan where he received a bachelor of arts degree in economics, then went to the University of California, Berkeley where he earned a master of arts and doctor of philosophy degrees in economics.\n\n\n\n\n\nHe returned to Singapore in 1998 and joined the Singapore Civil Service.\n\n\n\n\n\nAs Deputy Secretary of the Ministry of Finance (2001-2007), he was involved in the design and implementation of the Singapore Savings Bond and the Central Provident Fund (CPF) investments.\n\n\n\n\n\nIn 2008, Koh was appointed as the Managing Director of the Monetary Authority of Singapore (MAS), where he oversees the development and implementation of macroeconomic policies for Singapore.\n\n\n\n\n\nHe is also the Chairman of the Government of Singapore Investment Corporation (GIC) which is responsible for the management of Singapore\u2019s foreign reserves and investment portfolio.]" time="0.549"><properties><property name="score" value="0.9581279754638672" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Little Tikes LTX Battery Operated Tricycle\n\nHe\u2019s been dreaming about his first bike for a while, and now he can hop on and ride around with friends and siblings alike on this bright red Little Tikes LTX Battery Operated Tricycle. It features an adjustable seat, a footrest and a chain guard so he\u2019ll stay safe while riding. Battery-operated, it requires 2 AA batteries (not included). This is a toy for children 3 years and older.]" time="0.247"><properties><property name="score" value="0.999829888343811" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99982989&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99982989
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Trixie Textured - My cat licked my thumb.\n\n\n\nJoin Date: Nov 2006 Posts: 4,592\n\nI have not been to China, though I have been to Japan and love it.\n\n\n\nI have friends in China who have also been to Japan, and they both love it, and neither of them think of it as a &quot;poor&quot; country. In fact, a lot of people who have been there say that China is much nicer.\n\n\n\nChina is not a poor country. They have tons of money and resources. Their problem is that they do not spend enough on infrastructure and don't value education as highly as they could. So their people do not make as much money as they could. And they don't have nice things. It is much nicer than India, but not as good as Japan. I would like to visit it, because I do not think I would be disappointed, but I don't expect it to be better than Japan.\n\n\n\nThey have a different mindset than us, and do things differently. Their government is not a democracy. It is a communist government. I think that is not a bad thing, and has it's benefits. They have much fewer guns and gun crime than we do. They are, for the most part, more honest than we are. Their infrastructure, while not as nice as ours, is very good. They spend more on infrastructure than we do. They have much more public transportation, and better public transportation. They are working on their environmental problems, and they have made more progress than we have. They have fewer social problems, and less crime than we do. They have a very good education system, especially compared to India.\n\n\n\nThey spend more money on schools and education, and invest more in their future, than we do.\n\n\n\nChina is a country that is far more well developed than many people in the US would think.\n\n\n\nThe fact that they have fewer guns, and better education and fewer social problems is probably why they have fewer school shootings than we do. I think we have way too many guns in this country, and that is a big part of the problem.\n\n\n\nJust my thoughts.]" time="1.113"><properties><property name="score" value="0.9997957348823547" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[&quot;Say to them: 'As surely as I live, declares the Sovereign Lord, I take no pleasure in the death of the wicked, but rather that they turn from their ways and live. Turn! Turn from your evil ways! Why will you die, people of Israel?' declares the Sovereign Lord.&quot;\n\nEzekiel 33:11-11\n\nGod wants every sinner to be saved. He does not want people to perish but be saved. God hates sin, and sin destroys a person. But He loves people. He created them to have fellowship with Him and one day live in His Kingdom, and He does not want that for one person to be lost.\n\n\n\n\n\nGod's command to Ezekiel was that He was telling him to go and tell the people of Israel that they need to turn to God, because they were going to die if they did not. We know that Ezekiel told them the same thing, because God's word is truth. God's Word does not fail.\n\n\n\n\n\nWhat God is saying is that people are going to die if they do not turn to Him. People who reject God will die. The Bible says in Romans 1:32, &quot;Though they know God's righteous decree that those who do such things deserve death, they not only continue to do these very things but also approve of those who practice them.&quot; God says that those people who do these things deserve death, and yet people continue to do them and approve of them.\n\n\n\n\n\nGod will not tolerate the sinful acts of humanity, but He wants every person to be saved. There is only one way to be saved. The Bible says in John 14:6, &quot;Jesus answered, 'I am the way and the truth and the life. No one comes to the Father except through me.'\u201d\n\n\n\n\n\nNo one can be saved except through Jesus Christ, but when people put Jesus as their savior, they are saved. It's that simple.\n\n\n\n\n\nThere are many people who call themselves Christians who say that God wants everyone to be saved, and they quote Ezekiel 33:11 and John 3:16, but they don't realize that God's word is truth. God says in Revelation 21:8, &quot;But the cowardly, the unbelieving, the vile, the murderers, the sexually immoral, those who practice magic arts, the idolaters and all liars\u2014their place will be in the fiery lake of burning sulfur. This is the second death.&quot;\n\n\n\n\n\nThere is only one way to be saved, and it is not through anyone or anything but Jesus Christ. Jesus is the only way to heaven. God says that all the people who are going to perish are going to go to a place called the second death. That's hell.\n\n\n\n\n\nWhen people quote Ezekiel 33:11 and John 3:16, they don't realize that God's word is truth. They don't realize that when they say that God wants everyone to be saved, they are not telling the truth. When they say that God wants everyone to be saved, they are putting God in a box, because God's word is truth. God says that those who do not accept Jesus as their Savior are going to perish. God's word is truth, and people who do not put their faith in Jesus Christ are going to die.\n\n\n\n\n\nThe Bible says in John 14:6, &quot;Jesus answered, 'I am the way and the truth and the life. No one comes to the Father except through me.'\u201d\n\n\n\n\n\nJesus is the only way to the Father, and no one can be saved except through Him. He died on the cross for our sins, and He rose from the dead on the third day, and He sits at the right hand of God, interceding for us. It's that simple.\n\n\n\n\n\nLet's take a look at Ezekiel 33:11-11, and read it carefully.\n\n\n\n\n\n&quot;Say to them: 'As surely as I live, declares the Sovereign Lord, I take no pleasure in the death of the wicked, but rather that they turn from their ways and live. Turn! Turn from your evil ways! Why will you die, people of Israel?' declares the Sovereign Lord.&quot;\n\n\n\n\n\nGod says that He takes no pleasure in the death of the wicked. God does not like it when people die. But God hates sin, and sin leads to death. If people want to live, they must turn from their sins and put their faith in Jesus Christ.\n\n\n\n\n\nIf people don't put their faith in Jesus Christ, they are going to perish. God doesn't want that. He doesn't want people to perish, but to live.\n\n\n\n\n\nGod wants people to put their faith in Jesus Christ. If people put their faith in Jesus Christ, they are saved. If they don't put their faith in Jesus Christ, they are going to perish. That's the truth. That's the message of Ezekiel 33:11-11.\n\n\n\n\n\nPeople are going to die if they don't put their faith in Jesus Christ. People will perish. People will be lost. People will be doomed. People will go to hell.\n\n\n\n\n\nLet's read Ezekiel 33:11-12.\n\n\n\n\n\n&quot;For I take no pleasure in the death of anyone,&quot; declares the Sovereign Lord. &quot;Repent and live!&quot;\n\n\n\n\n\nGod takes no pleasure in the death of anyone. God does not want anyone to perish. He doesn't want anyone to die. He wants people to repent and live. God is saying that He is not going to stand back and watch as people perish. He's going to do something about it.\n\n\n\n\n\nGod says in Ezekiel 33:13-16, &quot;Turn! Turn from your evil ways! Why will you die, people of Israel? I have no pleasure in the death of the wicked, but rather that they turn from their ways and live. Do you think I take pleasure in the death of the wicked? declares the Sovereign Lord. Rather, am I not pleased when they turn from their ways and live? Don't presume that you can say to yourselves]" time="2.862"><properties><property name="score" value="0.9998182535171509" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Image caption Bailiffs removed protesters from Grangemouth plant in October 2013\n\nFive people have been charged in connection with an anti-fracking protest at an oil refinery in Grangemouth.\n\nThe protesters climbed on to the chimney at Ineos petrochemicals plant on 9 October.\n\nPolice Scotland said two men, aged 21 and 24, had been charged in connection with the protest and would appear at Falkirk Sheriff Court on Monday.\n\nThree women, aged 40, 44 and 48, were charged with breach of the peace and will also appear at court.\n\nThe 40-year-old was also charged with breach of the peace.\n\nThe other three were arrested in connection with a protest outside a hotel in Glasgow on 10 October.]" time="0.377"><properties><property name="score" value="0.5120607614517212" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.51206076&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.51206076
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Working with an example - rotating through an array\n\nFor a number of years, I've been fascinated by the idea of writing code that can rotate through an array. Let's say I have the following array of numbers:\n\nvar myNumbers = [ 1, 2, 3, 4, 5 ];\n\nThe following snippet of code would display the same numbers in reverse order:\n\nfor ( var i = myNumbers.length - 1; i &gt;= 0; i-- ) { console.log( myNumbers[i] ); }\n\nWhat's going on here?\n\nWe know that the index of an array can be used as a &quot;pointer&quot; to the location in memory that contains a particular item in an array. If the array myNumbers has five items, then myNumbers[0] , myNumbers[1] , myNumbers[2] , myNumbers[3] and myNumbers[4] will point to the values 1 , 2 , 3 , 4 and 5 respectively.\n\nIn the previous example, the code starts with the length of the array (5) and subtracts one from it. As we know that index 0 in an array contains the first item in the array, the previous code will execute the following steps:\n\nTake the length of the array (5) Subtract 1 (so we now have 4) Start at the index that contains the first item in the array (myNumbers[4] = 3) and run down the array until we reach the last item in the array (myNumbers[3] = 5)\n\nSo the above code will end up displaying the items in the array in reverse order, starting at index 4 and working backwards to index 0.\n\nAt this point, we have a good idea of how we can access the values in an array in a specific order, but how do we go the other way?\n\nIt would be great if we could simply use the same code as above, but add a &quot;--&quot; operator before the last number. That way, we could add a simple parameter to the code to control the range of numbers it will use.\n\nUnfortunately, we can't do that. While JavaScript will happily work with -- , it will also happily take -- in code and convert it to -1 , or the last number in an array. That's not going to do us any good, as we want the code to be flexible enough that we can work with a number that we specify as the range of numbers.\n\nSo let's look at a simple function that can give us the results we need.\n\nfunction rotate(numbers, direction, number) { numbers.reverse(); numbers.push(number); numbers.unshift(number); };\n\nThe rotate function above has three parameters, and will rotate the values in the numbers array.\n\nnumbers is the array we want to use.\n\ndirection is the parameter that allows us to choose whether we want the values to be rotated in the direction we specify. In this case, we've chosen direction to be either &quot;right&quot; or &quot;left&quot;\n\nnumber is the number we're using as the upper bound of the range that we want to use.\n\nLet's take a look at a simple example:\n\nvar numbers = [ 1, 2, 3, 4, 5 ]; rotate( numbers, &quot;right&quot; ); console.log( numbers ); //output: [ 1, 2, 3, 4, 5 ]\n\nAs we can see, the numbers array is simply rotated right by the number we specified. That's fine, but let's add in some code that we can use to control which numbers are displayed:\n\nvar numbers = [ 1, 2, 3, 4, 5 ]; rotate( numbers, &quot;right&quot;, 1 ); console.log( numbers ); //output: [ 1, 4, 3, 5, 2 ]\n\nWe can see that we've rotated the numbers so that the first number in the array is now number 1 , and the next number in the array is 4 .\n\nWe've also specified that we want to rotate in a right-handed manner, by passing &quot;right&quot; as the second parameter. This means that we have access to the following rotation values:\n\nrotate( numbers, &quot;left&quot; ); rotate( numbers, &quot;right&quot; ); rotate( numbers, &quot;right&quot; ); rotate( numbers, &quot;right&quot; ); rotate( numbers, &quot;left&quot; );\n\nIf we try to access the rotate function with &quot;left&quot; as the parameter, we'll see the following output:\n\nvar numbers = [ 1, 2, 3, 4, 5 ]; rotate( numbers, &quot;left&quot; ); console.log( numbers ); //output: [ 4, 3, 5, 2, 1 ]\n\nNot quite what we wanted, right?\n\nWhat we want is to be able to take an array, specify the number of numbers we want to display, and be able to choose the direction of rotation.\n\nWe also want to ensure that the values we are displaying are always within the array's bounds, and we're using them correctly.\n\nThe good news is that it's not too difficult to do this. We simply need to know what we want the code to look like, and write the code to do what we want it to do.\n\nLet's take a look at some simple code to display the numbers in an array in a given direction:\n\nfunction rotateArray( numbers, direction, startAt ) { numbers.reverse(); numbers.push( startAt ); numbers.unshift( startAt ); };\n\nrotateArray function\n\nThis is the code we'll use to display the numbers in the numbers array in the direction we specify. We'll start with a simple parameter:\n\nstartAt is the number we're displaying. It can either be the first number in the array, or any number within the array.\n\nLet's take a look at the code again:\n\nfunction rotateArray( numbers, direction, startAt ) { numbers.reverse(); numbers.push( startAt ); numbers.unshift( startAt ); };\n\nUsing this simple function, we can take any number of numbers, and display them in the direction we want. Let's look at a quick example:\n\nvar numbers = [ 1, 2, 3, 4, 5 ]; rotateArray( numbers, &quot;left&quot;, 3 ); console.log( numbers ); //output: [ 4, 3, 5, 2, 1 ]\n\nThis time, we've used &quot;left&quot; as the direction, and we've also used the parameter &quot;3&quot; as the startAt value. The result is that we now have the numbers displayed in the array starting at the number 3 .\n\nIf we use a number that isn't within the bounds of the array, we'll get the following result:\n\nvar numbers = [ 1, 2, 3, 4, 5 ]; rotateArray( numbers, &quot;right&quot;, 3 ); console.log( numbers ); //output: [ 5, 2, 4, 1, 3 ]\n\nAs you can see, the direction we've specified is &quot;right&quot;, and the value we've specified for startAt is 3 . Unfortunately, this value isn't within the bounds of the array, so the values in the array have been displayed starting from the last number in the array.\n\nLet's take a look at how we can use this code to work with a number that's outside the bounds of the array:\n\nvar numbers = [ 1, 2, 3, 4, 5 ]; rotateArray( numbers, &quot;right&quot;, 6 ); console.log( numbers ); //output: [ 5, 2, 4, 1, 3 ]\n\nThis time, we've specified &quot;right&quot; as the direction, and a value of 6 as the startAt parameter. The result is that we now have the values displayed in the array starting at the number 6 .\n\nThis time, the result we want is displayed in the array, but there's a problem. We don't want the number 6 to be the start of the array. Instead, we want it to be the last number in the array.\n\nThe good news is that we can easily handle this by modifying the code slightly:\n\nfunction rotateArray( numbers, direction, startAt ) { numbers.reverse(); numbers.push( startAt ); numbers.unshift( startAt ); };\n\nUsing the simple rotation code above, we can now take any array, specify the number we want to display, and choose the direction we want to use. This means we can write some simple code that can work with any number, regardless of its position in the array.\n\nThe next step is to write some code that can ensure the number we specify as the startAt is in the range that we expect.\n\nLet's take a look at how we can do that:\n\nfunction rotateArray( numbers, direction, startAt ) { if( startAt &gt;= numbers.length ) { throw new RangeError(&quot;Start at is outside of the]" time="4.527"><properties><property name="score" value="0.995492330619267" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Contents\n\n1. The Meaning of Love\n\n2. Our Attitude Toward Others\n\n3. The Two Great Commandments\n\n4. Developing Love\n\n5. How to Help Others\n\n6. The Golden Rule\n\n7. The Commandments in the New Testament\n\n8. Tolerance\n\n\nCHAPTER 1\n\nThe Meaning of Love\n\nBefore we can love others, we must know the meaning of love. It is important to love God and obey Him and to have a right relationship with Him. But to have a right relationship with Him we must love Him. That means that we must have a good relationship with all His laws.\n\nOf course we can't keep all the commandments all the time. Yet we must have a good relationship with them. Otherwise we will not know what God is like.\n\nLove is based on God's law.\n\nRomans 13:10. The commandments, You shall not commit adultery, You shall not steal, You shall not bear false witness, You shall not covet; and any other commandment, are summed up in this word, You shall love your neighbor as yourself.\n\nGalatians 5:14. For all the law is fulfilled in one word, even in this: You shall love your neighbor as yourself.\n\nLove is based on God's law. We should study God's commandments and know the laws He has given us. Then we will know how to have a good relationship with Him.\n\nOne reason we are to love God and obey Him is that it is important to have a good relationship with God. We need to know how much God loves us. Without knowing this we cannot have a right relationship with God. We need to have a good relationship with Him so we can be happy and serve Him better.\n\nLet us understand what love is.\n\n1 John 4:8,9. He who does not love does not know God, for God is love. In this the love of God was manifested toward us, that God has sent His only begotten Son into the world, that we might live through Him.\n\nIn these verses John explained that love is the opposite of lawlessness. Love is the opposite of murder, hatred, malice, evil, adultery, stealing, lying, and a lot of other bad things.\n\nRomans 13:8-10. Owe no man anything, but to love one another: for he that loves another has fulfilled the law. For this, You shall not commit adultery, You shall not kill, You shall not steal, You shall not covet, and if there be any other commandment, it is briefly comprehended in this saying, namely, You shall love your neighbor as yourself. Love does no harm to a neighbor; therefore love is the fulfillment of the law.\n\nWe learn from this that the opposite of love is lawlessness. The opposite of love is murder, hatred, envy, malice, evil, and all the other things mentioned in these verses.\n\nVerse 10 says, &quot;Love does no harm to a neighbor.&quot; In other words, love is a good thing. It is a good thing to have a good relationship with God.\n\nRomans 13:8 says, &quot;Love does no harm to a neighbor.&quot; Love helps and blesses others.\n\nLove is a good thing. The reason love is a good thing is because God is love.\n\n1 John 4:16. And we have known and believed the love that God has for us. God is love; and he who abides in love abides in God, and God in him.\n\nGod is love.\n\n1 John 4:7-11. Beloved, let us love one another: for love is of God; and every one that loves is born of God, and knows God. He that loves not knows not God; for God is love. In this was manifested the love of God toward us, because that God sent his only begotten Son into the world, that we might live through him. Herein is love, not that we loved God, but that he loved us, and sent his Son to be the propitiation for our sins. Beloved, if God so loved us, we ought also to love one another. No man has seen God at any time. The only begotten Son, which is in the bosom of the Father, he has declared him.\n\nVerse 8 says, &quot;God is love.&quot; God's name is Love.\n\nGod's name is Love.\n\nJohn 4:8. \u2026 the well was deep; and I said, &quot;How shall I get down?&quot;\n\nAnd he said, &quot;Come, and I will show you.&quot;\n\n9. And he said, &quot;Draw you [a] little water out of the well, and \u2026 drink.&quot;\n\n10. And I said, &quot;Sir, give me [a] little to drink.&quot;\n\n11. And he said, &quot;Drink, my [b] son, and \u2026 your eyes shall be opened.&quot;\n\n12. And I drank, and it was as if I had had new vision given to me; for I could see [c] through those walls of water.\n\n13. And he said to me, &quot;Go [a] down, for the water is [b] yet a little way below the earth. \u2026\n\n15. \u2026 This water is [c] that which I said \u2026 will make you come up hither.&quot;\n\n16. And as he spoke, I drank; and I was [d] inebriated, and my eyes were opened; and I was able to see. \u2026\n\n19. \u2026 &quot;Come,&quot; he said, &quot;follow me \u2026 and I will give you \u2026 rest.&quot;\n\nGod is the God of rest.\n\n1 Thessalonians 4:9. \u2026 God has not appointed us to wrath, but to obtain salvation by our Lord Jesus Christ.\n\nGod is our salvation. God is our rest. God is our peace.\n\nGod is our rest.\n\n1 Peter 2:6. \u2026 you may be \u2026 a kind of first-fruits of his creatures.\n\nGod is our first fruits.\n\n2 Timothy 2:13. If we are faithless, he remains faithful; for he cannot deny himself.\n\nGod is faithful.\n\nGod is the God of rest.\n\nGod is our salvation.\n\nGod is our peace.\n\nGod is our rest.\n\nHebrews 4:10-13. For he who has entered His rest has himself also ceased from his own works, as God did from His. Let us therefore be diligent to enter that rest \u2026\n\n12. \u2026 For the word of God is living and active, and sharper than any two-edged sword, and piercing even to the division of soul and spirit, and of joints and marrow, and is a discerner of the thoughts and intents of the heart.\n\n13. \u2026 neither is there any creature that is not manifest in his sight; but all things are naked and opened to the eyes of him with whom we have to do.\n\nGod is a discerner of the thoughts and intents of the heart.\n\nGod sees everything.\n\nGod sees everything.\n\n1 John 4:8,9. He who does not love does not know God, for God is love. In this the love of God was manifested toward us, that God has sent His only begotten Son into the world, that we might live through Him.\n\nGod is love.\n\n1 John 3:16,18. Hereby perceive we the love of God, because he laid down his life for us: and we ought to lay down our lives for the brethren.\n\n19. But whoso has this world's goods, and sees his brother in need, and shuts up his heart from him, how does the love of God abide in him?\n\nGod is love.\n\nGod is our rest.\n\nGod is our salvation.\n\nGod is our peace.\n\nGod is a discerner of the thoughts and intents of the heart.\n\nGod is love.\n\nGod is our rest.\n\nGod is our salvation.\n\nGod is our peace.\n\nGod is a discerner of the thoughts and intents of the heart.\n\nThe Bible says that God is love.\n\nWe can study love.\n\n1 John 4:7-11. Beloved, let us love one another: for love is of God; and every one that loves is born of God, and knows God. He that loves not knows not God; for God is love. In this was manifested the love of God toward us, because that God sent his only begotten Son into the world, that we might live through him. Herein is love, not that we loved God, but that he loved us, and sent his Son to be the propitiation for our sins. Beloved, if God so loved us, we ought also to love one another. No man has seen God at any time. The only begotten Son, which is in the bosom of the Father, he has declared him.\n\nVerse 11 says, &quot;Herein is love, not that we loved God, but that he loved us.&quot; The opposite of love is lawlessness. But the opposite of love is not lawlessness. The opposite of love is law.\n\nLoving God and obeying Him is loving Him.\n\nThe Bible says that God is love.\n\nLoving God and obeying Him is loving Him.\n\nLoving God and obeying Him is loving God.\n\nJohn 14:21. He that has my commandments,]" time="4.894"><properties><property name="score" value="0.5310911163687706" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Interviews\n\nInterview with Ariel Efraim (Soilwork)\n\nInterview with Steven Rosen (August Burns Red)\n\nInterview with Brent Hinds (Mastodon)\n\nInterview with Paul Kuhr (I See Stars)\n\nInterview with Christian Antidormi (Hatebreed)\n\nInterview with Jack Syperek (Structures)\n\nInterview with Ben Gordon (The Treatment)\n\nInterview with Doc Coyle (God Forbid)\n\nInterview with Chris Barnes (Six Feet Under)\n\nInterview with Mike D (The Sword)\n\nInterview with Russ Rankin (The Slackers)\n\nInterview with Claudio Sanchez (Coheed and Cambria)\n\nInterview with Udo Dirkschneider (Accept)\n\nInterview with Ben Weinman (Dillinger Escape Plan)\n\nInterview with Matt Pike (High On Fire)\n\nInterview with Joe Duplantier (Gojira)\n\nInterview with Johannes Eckerstr\xf6m (Evergrey)\n\nInterview with Pete Loose (Reel Big Fish)\n\nInterview with Pete Wentz (Fall Out Boy)\n\nInterview with Daniel Svensson (Raubtier)\n\nInterview with Chino Moreno (Deftones)\n\nInterview with Pierre Bouvier (Simple Plan)\n\nInterview with Jon Egan (The Ghost Of A Thousand)\n\nInterview with Mike D (Lack Of Afro)\n\nInterview with Michael Crafter (Sevendust)\n\nInterview with Charlie Benante (Anthrax)\n\nInterview with Rob Caggiano (Volbeat)\n\nInterview with Ronnie Atkins (Pretty Maids)\n\nInterview with Gee Anzalone (Exhorder)\n\nInterview with Micky Shirley (Five Finger Death Punch)\n\nInterview with Oderus Urungus (GWAR)\n\nInterview with Ryan Downey (Texas Hippie Coalition)\n\nInterview with Chris Chasse (Unearth)\n\nInterview with Lee Altus (Suffocation)\n\nInterview with Roy Mayorga (Soulfly)\n\nInterview with Gus G (Firewind)\n\nInterview with Paul Allender (In Flames)\n\nInterview with Francesco Artusato (All Shall Perish)\n\nInterview with Rody Walker (Pyogenesis)\n\nInterview with Mike Dean (Slipknot)\n\nInterview with Angela Gossow (Arch Enemy)\n\nInterview with Trevor Strnad (The Black Dahlia Murder)\n\nInterview with Mike Peters (The Alarm)\n\nInterview with Jonas Ekdahl (Freak Kitchen)\n\nInterview with Dave Ellefson (Megadeth)\n\nInterview with Corey Taylor (Slipknot)\n\nInterview with Keith Buckley (Every Time I Die)\n\nInterview with Jack Frost (Kill Devil Hill)\n\nInterview with Josh Rand (Temple of Brutality)\n\nInterview with Michael Amott (Arch Enemy)\n\nInterview with Scott Gorham (Thin Lizzy)\n\nInterview with Matt Heafy (Trivium)\n\nInterview with Joe Kresge (Unearth)\n\nInterview with Randy Blythe (Lamb of God)\n\nInterview with Benji Webbe (Dimmu Borgir)\n\nInterview with Max Cavalera (Soulfly)\n\nInterview with John Bush (Armored Saint)\n\nInterview with Davey Muise (Exhorder)\n\nInterview with Stephen Carpenter (Deftones)\n\nInterview with Dani Filth (Cradle Of Filth)\n\nInterview with Leif Cuzner (Darkane)\n\nInterview with George \u201cCorspegrinder\u201d Fisher (Cannibal Corpse)\n\nInterview with Steve Asheim (Hatebreed)\n\nInterview with Bobby Blitz Ellsworth (Overkill)\n\nInterview with Adam Dutkiewicz (Killswitch Engage)\n\nInterview with Mick Murphy (Exhumed)\n\nInterview with Doug Pinnick (King\u2019s X)\n\nInterview with Alexi Laiho (Children Of Bodom)\n\nInterview with M. Shadows (Avenged Sevenfold)\n\nInterview with Alex Holzwarth (Rhapsody Of Fire)\n\nInterview with Carlos Cruz (Soulfly)\n\nInterview with Glenn Hughes (Deep Purple/Black Country Communion)\n\nInterview with Oliver Palotai (Kamelot)\n\nInterview with Meegs Rascon (The Mars Volta)\n\nInterview with Joakim Brod\xe9n (Sabaton)\n\nInterview with Tim \u201cRipper\u201d Owens (Iced Earth/Yngwie Malmsteen)\n\nInterview with Matt Heafy (Trivium)\n\nInterview with Warrel Dane (Nevermore)\n\nInterview with Zach Myers (Shinedown)\n\nInterview with Michael Poulsen (Volbeat)\n\nInterview with Michael Amott (Arch Enemy)\n\nInterview with Mat Sinner (Primal Fear)\n\nInterview with Bj\xf6rn \u201cSpeed\u201d Strid (Soilwork)\n\nInterview with Matt Heafy (Trivium)\n\nInterview with Rob Flynn (Machine Head)\n\nInterview with Serj Tankian (System Of A Down)\n\nInterview with Ville Valo (HIM)\n\nInterview with Jimmy DeGrasso (Megadeth)\n\nInterview with Joey Jordison (Slipknot)\n\nInterview with Chris Jericho (Fozzy/Fozzy Osbourne)\n\nInterview with Jon Donais (Shadows Fall)\n\nInterview with Jason Hook (Five Finger Death Punch)\n\nInterview with Steve Mazur (Grave Digger)\n\nInterview with Bj\xf6rn \u201cSpeed\u201d Strid (Soilwork)\n\nInterview with Sonny Sandoval (P.O.D.)\n\nInterview with Serj Tankian (System Of A Down)\n\nInterview with Fredrik Thordendal (Meshuggah)\n\nInterview with Wes Borland (Limp Bizkit)\n\nInterview with Alexi Laiho (Children Of Bodom)\n\nInterview with Blaze Bayley (Iron Maiden)\n\nInterview with Mark Hunter (Chimaira)\n\nInterview with Max Cavalera (Soulfly/Sepultura)\n\nInterview with Dave Lombardo (Slayer)\n\nInterview with Bruce Dickinson (Iron Maiden)\n\nInterview with Morgan Lander (Fear Factory)\n\nInterview with Oliver Palotai (Kamelot)\n\nInterview with James Hetfield (Metallica)\n\nInterview with Al Jourgensen (Ministry)\n\nInterview with Corey Taylor (Stone Sour/Slipknot)\n\nInterview with Alex Skolnick (Testament)\n\nInterview with Charlie Benante (Anthrax)\n\nInterview with Chris Reifert (Autopsy)\n\nInterview with Lajon Witherspoon (Sevendust)\n\nInterview with Jason McMaster (Grave Robber)\n\nInterview with Jon Howard (Light The Torch)\n\nInterview with Neil Fallon (Clutch)\n\nInterview with Dino Cazares (Fear Factory)\n\nInterview with Oderus Urungus (GWAR)\n\nInterview with Rex Brown (Pantera)\n\nInterview with Dave Lombardo (Slayer)\n\nInterview with Serj Tankian (System Of A Down)\n\nInterview with John Moyer (Disturbed)\n\nInterview with Mike Portnoy (Adrenaline Mob)\n\nInterview with King Diamond\n\nInterview with Trevor Strnad (The Black Dahlia Murder)\n\nInterview with Mikael \xc5kerfeldt (Opeth)\n\nInterview with Scott Ian (Anthrax)\n\nInterview with Paul Gray (Slipknot)\n\nInterview with Casey Chaos (Amen)\n\nInterview with Opeth\u2019s Mikael \xc5kerfeldt\n\nInterview with Mark Morton (Lamb Of God)\n\nInterview with Shawn Drover (Megadeth)\n\nInterview with Chino Moreno (Deftones)\n\nInterview with Serj Tankian (System Of A Down)\n\nInterview with Bruce Corbitt (Corrosion Of Conformity)\n\nInterview with Ray Alder (Fates Warning)\n\nInterview with Rob \u201cBlasko\u201d Nicholson (Ozzy Osbourne)\n\nInterview with Tom Araya (Slayer)\n\nInterview with Max Cavalera (Soulfly)\n\nInterview with Ryan Martinie (Cynic)\n\nInterview with Max Cavalera (Soulfly)\n\nInterview with Jason Newsted (Voivod)\n\nInterview with Mick Murphy (Exhumed)\n\nInterview with Lars Frederiksen (Rancid)\n\nInterview with Mike Muir (Suicidal Tendencies)\n\nInterview with Dave Mustaine (Megadeth)\n\nInterview with Neil Fallon (Clutch)\n\nInterview with Bruce Dickinson (Iron Maiden)\n\nInterview with Maynard James Keenan (Tool)\n\nInterview with Mike \u201cRat\u201d Ratner (S.O.D.)\n\nInterview with Max Cavalera (Soulfly)\n\nInterview with Tom Araya (Slayer)\n\nInterview with Chuck Billy (Testament)\n\nInterview with Tom Araya (Slayer)\n\nInterview with Phil Anselmo (Down)\n\nInterview with King Diamond\n\nInterview with Corey Taylor (Stone Sour/Slipknot)\n\nInterview with Dave Mustaine (Megadeth)\n\nInterview with Lars Ulrich (Metallica)\n\nInterview with Jerry Cantrell]" time="4.704"><properties><property name="score" value="0.09577819279261998" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Apple\u2019s ResearchKit is off to a good start, but there\u2019s plenty of work to be done.\n\n\n\nResearchKit was launched in March as a way for Apple\u2019s developers to easily tap into medical data from the iPhone to help researchers track and collect information about a variety of health conditions. Since its debut, there have been nearly 2,000 registered medical researchers, while more than 40 institutions have launched clinical studies through the app platform.\n\n\n\nBut there are several hurdles for Apple and its partners. On Wednesday, Dr. Harvey Fineberg, president of the Institute of Medicine, will join at a panel discussion at Mount Sinai Hospital in New York to examine what works and what\u2019s missing in the fast-evolving field of mobile health.\n\n\n\nThe session, titled \u201cWhat Are the Future Directions for ResearchKit?\u201d comes after a first wave of successful projects on ResearchKit, including apps that help people manage diabetes, heart disease and asthma.\n\n\n\nThere are also some examples of the problems Apple is facing as it tries to get more partners to adopt the tool. Earlier this month, a major player in the diabetes community said it would not be taking part in the app\u2019s data collection. The Juvenile Diabetes Research Foundation said in a statement that its members were not interested in the iPhone platform and \u201cwould prefer to conduct research studies using our own patient data that is gathered on an Android platform.\u201d\n\n\n\nThat\u2019s likely an expensive prospect for the organization, given the popularity of iPhones in the United States. As of last year, Android devices accounted for just over half of all U.S. smartphone subscribers, while iOS devices made up around 41 percent. That\u2019s according to the Pew Research Center, which estimated last year that there were 121 million iPhone users in the United States.\n\n\n\nThe success of Apple\u2019s platform will hinge in part on how it handles data privacy and security concerns. Apple had to clear some hurdles in the early days of ResearchKit because many people were concerned about privacy when they heard it would rely on data from Apple\u2019s servers. Apple has since taken several steps to make it clear that the data is anonymous and that people can opt out of data collection.\n\n\n\nThe app program could also be a major part of the iPhone\u2019s future if it takes off. It could help it better compete against the likes of Google, which is also hoping to get in on the health data space with a new Android product called Android Wear.\n\n\n\nThe first public version of ResearchKit will be unveiled on Wednesday, the panel will also include representatives from Mount Sinai and Pfizer.\n\n\n\nThat event will be at 11:30 a.m. EDT in New York City, and it will be livestreamed on Fortune\u2019s website.]" time="1.542"><properties><property name="score" value="0.4640413721402486" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[\n\nSend this thread via email by filling out the fields below\n\n\n\nThread: Lost 10lbs in the last 3 weeks Send To:\n\nUse semi-colon to separate multiple addresses Your Name: Your Email: Add a personal message: Send this thread via email by filling out the fields below\n\n\n\n]" time="0.200"><properties><property name="score" value="0.8082296252250671" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[I don't think it is your attitude. I think it is just the way you are.\n\nIt isn't a criticism, just a statement of fact. It's like saying I don't think you are stupid.\n\nJust that you are naive, sweet, guileless, open. You have so much going for you, you don't have to be cynical.\n\nBut it isn't fair to the rest of us, not fair to expect us to feel your innocence and not to respond to it.\n\nYou need to toughen up, otherwise you will just be hurt. But don't ever be cynical.]" time="0.323"><properties><property name="score" value="0.9995108842849731" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Get the biggest daily news stories by email Subscribe Thank you for subscribing We have more newsletters Show me See our privacy notice Could not subscribe, try again later Invalid Email\n\nAs the Queen's cousin, Lady Amelia Windsor is more used to socialising with the royal family than you or I.\n\nBut even for her, it seems the sight of seeing Kate Middleton and Prince William in attendance at Pippa Middleton's wedding - as well as the newlyweds' own big day - was enough to make her feel &quot;uncomfortable&quot;.\n\nSpeaking at the wedding of one of her friend's brothers in Kent, the 28-year-old spoke about how it made her feel to have them there, reports the Daily Mail .\n\nShe said: &quot;Well I think it is a beautiful wedding, obviously, because it's Pippa and James.\n\n&quot;I've never been to a wedding before, so I don't know really what it should be like. But I do find it quite uncomfortable seeing them.&quot;\n\nThe Royal family came together to celebrate the wedding of Kate's sister to hedge fund manager James Matthews in the Berkshire village of Englefield.\n\n(Image: Getty Images Europe)\n\n(Image: PA)\n\n(Image: PA)\n\nAnd while Pippa and James' wedding was a private affair with just 200 guests, they were still surrounded by some royals as well as other notable names.\n\nLady Amelia's father is the Queen's first cousin, so it was no surprise to see the Queen herself attend the wedding, alongside Prince Harry and Duchess Meghan.\n\nThe Princesses Beatrice and Eugenie were also there, as well as Kate's sister, Pippa's brother James Middleton, David and Victoria Beckham and Sir Elton John.\n\nLady Amelia revealed that she even had a few words with Harry and Meghan, with the former asking her how she felt about her cousin's big day.\n\nShe said: &quot;I had a little conversation with Harry and Meghan, and they said how lovely the wedding was. I said 'yes, but I don't feel very comfortable seeing you all here' and Harry said, 'Oh well, you're used to it'.\n\n(Image: PA)\n\n(Image: PA)\n\n&quot;I said, 'Yes, but it is still very strange to see you here'. I suppose you can't blame me, I am a bit naive. But the whole thing was really lovely.&quot;\n\nShe added: &quot;I mean, James is my cousin, I grew up with him and I love him dearly, so it was really lovely to see him. It was a lovely wedding, they are such a lovely couple. I am so pleased they are so happy.&quot;\n\n(Image: Getty Images Europe)\n\n(Image: Getty Images Europe)\n\nAnd as for what she thought of her father's outfit - which caused much comment among fans - she admitted she wasn't sure.\n\nShe said: &quot;Well, he always wears an odd choice of clothes, but that is what he likes. It wasn't bad, I mean he didn't look like a big prawn.&quot;\n\nShe added: &quot;It was lovely to see him happy.&quot;]" time="1.609"><properties><property name="score" value="0.9994182189305624" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Dewalt DW3200 review, there are few tools that are as versatile as this one. You may have a thousand bucks of cordless tools, but there\u2019s nothing that will do as much as this one will do. As you might have noticed, the construction of this unit is incredibly sturdy, so that you don\u2019t have to worry about it breaking when you use it. Also, it is a bit heavy, and it will take a while to get used to using it, especially if you have not used any other heavy drills.\n\nPower\n\nWe also tested the power of this Dewalt DW3200 and we must say that it is very impressive. It delivers 3,300 RPM, which is more than enough to bore holes and drill in all kinds of materials. It can easily penetrate into wood, brick, or concrete and it can bore holes up to two inches. The trigger lock mechanism also works well, which makes this drill safer to use.\n\nCordless or corded?\n\nIt is important to remember that this drill can work either as a corded or a cordless unit, and this is very important for you to know. This Dewalt DW3200 uses the 20 volt lithium-ion batteries and they can last longer than you might expect. They are also very light and durable, so you will be able to use them for a long time. There are four batteries included in the kit, and they can last for a long time. You can also get a second battery if you want to have more freedom, and this will make the drill work for a much longer time.\n\nExtra Features\n\nYou may also like the extra features included in this Dewalt DW3200. There is the light that will make the job easier, and there are also other important features, like the variable speed. This is also a very durable drill, so it will last for a very long time. Also, the build quality is fantastic, so you don\u2019t have to worry about it breaking easily.\n\nOverall\n\nIf you have a lot of different drilling needs, this is the drill for you. This drill is perfect for the casual user, but it can also be used by professionals. It is also very durable, so you can use it for a very long time, and you will get the most for your money. You should also remember that this drill is very powerful, and it can work for a long time, without breaking down.\n\nIf you want to see what more we have, you can visit Amazon.com and check all of the available drills. You can also check our article on the best reciprocating saw, to see what the best product in this category is. Also, don\u2019t forget to like and share this article, if you liked it.]" time="1.439"><properties><property name="score" value="0.9771516521771749" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Source: Food\n\nAs I\u2019m sure you\u2019re well aware by now, it\u2019s officially National Grilled Cheese Sandwich Month. With that in mind, this week\u2019s Food Week: Hors D\u2019Oeuvre features a variety of hors d\u2019oeuvres you can make in your own home. After all, a good grilled cheese is just like a fine wine: it gets better with age.\n\nContinue reading]" time="0.250"><properties><property name="score" value="0.8901404738426208" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Jack and Justin are two of the hottest and most popular guys on the block. They make the perfect couple and know exactly what each other likes. But when they meet Tyler, things get complicated. This sizzling trio enjoys their time together and both Tyler and Jack crave more. But Justin is not so easily convinced. How will this all work out?\n\nWarning: This story contains three scorching hot guys and lots of hot and hard action.\n\n\n\n\xa92015 Scarlet Rose (P)2017 Scarlet Rose]" time="0.240"><properties><property name="score" value="0.9998045563697815" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[LAS VEGAS, NV - JANUARY 07: Recording artist Demi Lovato performs onstage during the 2017 DIRECTV NOW Super Saturday Night Concert at Park Theater at Park MGM on January 7, 2017 in Las Vegas, Nevada. (Photo by Kevin Winter/Getty Images for DIRECTV) (Photo: Kevin Winter, 2017 Getty Images)\n\nLAS VEGAS - Former Disney star Demi Lovato will release a new single from her sixth studio album on June 15th.\n\nThat's according to a video of Lovato, 26, posted on the YouTube channel of Las Vegas radio station KVEG.\n\nThe new album is the follow-up to 2015's &quot;Confident,&quot; which features the top 10 hits &quot;Cool for the Summer&quot; and &quot;Confident.&quot;\n\nMore: Everything you need to know about this year's 2017 Route 91 Harvest festival\n\nMore: Demi Lovato set for national tour with DJ Khaled\n\nMore: Demi Lovato re-engages with ex-boyfriend Wilmer Valderrama, so could there be more?\n\nThe pop singer also revealed she is launching a partnership with American Airlines to celebrate Hispanic Heritage Month, and she plans to go on a national tour with DJ Khaled.\n\nDemi Lovato to Release New Single, Announce Album Title on June 15th https://t.co/OBkKbXWySZ \u2014 Pop Crave (@PopCrave) June 1, 2017\n\nRead or Share this story: https://usat.ly/2rVY4mA]" time="0.750"><properties><property name="score" value="0.8391260802745819" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[RACISM IN BRITAIN HAS BECOME SO entrenched that it is becoming a bigger problem than it was at the time of the Birmingham pub bombings, when the threat of violent nationalism in the late 1970s prompted the first official inquiry into racial attacks, according to a report published today by the Equality and Human Rights Commission (EHRC).\n\nThe report, We\u2019re still struggling: the legacy of the Stephen Lawrence inquiry, says that despite years of measures to increase the representation of ethnic minorities in public life and protect them from discrimination, they are still \u201con the receiving end of racism\u201d, with black people twice as likely to suffer police stop and search tactics as whites.\n\nFigures from the Crown Prosecution Service (CPS) show that more than 40 per cent of cases involving charges of racially aggravated harassment are dropped, with charges reduced to more minor ones, or not pursued. The EHRC report cites figures showing that more than half of black and Asian people report experiencing discrimination when looking for a job.\n\nIn October, the Telegraph reported how]" time="0.580"><properties><property name="score" value="0.8731753528118134" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[# This Source Code Form is subject to the terms of the Mozilla Public\n\n# License, v. 2.0. If a copy of the MPL was not distributed with this\n\n# file, You can obtain one at http://mozilla.org/MPL/2.0/.\n\n# The following comment must be at the start of a new line.\n\n#\n\n#\n\n# The contents of this file are subject to the Mozilla Public\n\n# License Version 1.1 (the &quot;License&quot;); you may not use this file\n\n# except in compliance with the License. You may obtain a copy of\n\n# the License at http://www.mozilla.org/MPL/\n\n# Software distributed under the License is distributed on an &quot;AS\n\n# IS&quot; basis, WITHOUT WARRANTY OF ANY KIND, either express or implied.\n\n# See the License for the specific language governing rights and\n\n# limitations under the License.\n\n# The Original Code is Mozilla Communicator client code, released\n\n# March 31, 1998.\n\n# The Initial Developer of the Original Code is\n\n# Netscape Communications Corporation.\n\n# Portions created by the Initial Developer are Copyright (C) 1998\n\n# the Initial Developer. All Rights Reserved.\n\n# Contributor(s):\n\n#\n\n# Alternatively, the contents of this file may be used under the terms of\n\n# either of the GNU General Public License Version 2 or later (the &quot;GPL&quot;),\n\n# or the GNU Lesser General Public License Version 2.1 or later (the &quot;LGPL&quot;),\n\n# in which case the provisions of the GPL or the LGPL are applicable instead\n\n# of those above. If you wish to allow use of your version of this file only\n\n# under the terms of either the GPL or the LGPL, and not to allow others to\n\n# use your version of this file under the terms of the MPL, indicate your\n\n# decision by deleting the provisions above and replace them with the notice\n\n# and other provisions required by the GPL or the LGPL. If you do not delete\n\n# the provisions above, a recipient may use your version of this file under\n\n# the terms of any one of the MPL, the GPL or the LGPL.\n\n#\n\n# ***** END LICENSE BLOCK *****\n\nMOCHITEST_MANIFESTS += [\n\n' xpcshell.ini ' ,\n\n' mochitest-glue.ini ' ,\n\n' mochitest-plain.ini ' ,\n\n]\n\nDIST_DIRS += [\n\n' external/ ' ,\n\n' ipc/ ' ,\n\n' modules/ ' ,\n\n]\n\nEXTRA_COMPONENTS += [\n\n' mozLoop.js ' ,\n\n' webextensions.js ' ,\n\n]\n\ndeps = [\n\n'../base ' ,\n\n'../browser/base ' ,\n\n'../browser/components/migration ' ,\n\n'../browser/components/places ' ,\n\n'../browser/themes/shared ' ,\n\n'../content/base ' ,\n\n'../content/html ' ,\n\n'../content/xul ' ,\n\n'../embedding/browser ' ,\n\n'../embedding/components/browser ' ,\n\n'../embedding/components/toolkit ' ,\n\n'../extensions/browser/api ' ,\n\n'../extensions/common/moz.build ' ,\n\n'../gfx/2d ' ,\n\n'../gfx/layers ' ,\n\n'../intl ' ,\n\n'../netwerk ' ,\n\n'../printing ' ,\n\n'../services/sync ' ,\n\n'../toolkit/components/global ' ,\n\n'../toolkit/components/printing ' ,\n\n'../toolkit/components/satchel ' ,\n\n'../toolkit/crashreporter ' ,\n\n'../toolkit/library ' ,\n\n'../toolkit/mozapps/extensions ' ,\n\n'../toolkit/mozapps/extensions/extensions-types.js ' ,\n\n'../xpcom/base ' ,\n\n'../xpcom/build ' ,\n\n'../xpcom/build/shared-nspr ' ,\n\n'../xpcom/build/shared-nspr/src ' ,\n\n'../xpcom/ds ' ,\n\n'../xpcom/ds/io ' ,\n\n'../xpcom/ds/xpcom ' ,\n\n'../xpcom/glue ' ,\n\n'../xpcom/io ' ,\n\n'../xpcom/threads ' ,\n\n'../xpcom/threads/src ' ,\n\n'../xpcom/threads/xpcom ' ,\n\n' ../xpfe/appshell/nsAppShell.cpp ' ,\n\n' ../xpfe/appshell/nsAppShell.h ' ,\n\n' ../xpfe/appshell/nsAppsShellService.cpp ' ,\n\n' ../xpfe/appshell/nsAppsShellService.h ' ,\n\n' ../xpfe/appshell/nsNativeAppSupportWin.cpp ' ,\n\n' ../xpfe/appshell/nsNativeAppSupportWin.h ' ,\n\n' ../xpfe/appshell/nsNativeAppSupportXPConnect.cpp ' ,\n\n' ../xpfe/appshell/nsNativeAppSupportXPConnect.h ' ,\n\n' ../xpfe/appshell/nsWindowRoot.cpp ' ,\n\n' ../xpfe/appshell/nsWindowRoot.h ' ,\n\n' ../xpfe/appshell/nsXULAppAPI.h ' ,]" time="3.154"><properties><property name="score" value="0.03840629756450653" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The world of professional wrestling is always changing, but the superstars themselves can sometimes get lost in the shuffle. Sometimes, the changes within the industry are so drastic that a superstar's popularity can get left in the past. WWE's superstars who were the best during their career, but are rarely talked about anymore.\n\nWWE is still the leader of the wrestling world. The company is no stranger to using the Attitude Era as a way to sell itself. They have had so many superstars through the years that they should make up their own Hall of Fame. Some of those superstars are the ones who are mentioned the most.\n\nThe casual fan knows about Hulk Hogan, Stone Cold, and The Rock. That is what makes this so fun. Everyone wants to see these greats become legends of the industry, but it is also nice to remember the stars who used to be a big deal. Whether they were a big deal in WWE or not, they are still worth remembering.\n\nContinue scrolling to keep reading Click the button below to start this article in quick view Share Tweet Email Copy Link Copied\n\n10 Sin Cara\n\nvia trbimg.com\n\nSin Cara has become a joke of a superstar. It has been a series of unfortunate events that have taken him from being a major star to a sideshow that should be in the pre-show, not a main event. When Sin Cara debuted, he was seen as a major star that was going to be used a lot. It took some time, but it finally happened.\n\nThe problem was that he never delivered.\n\nHe was so unimpressive that he was booted from WWE in 2016. He is making an appearance at WrestleMania 33. For his sake, it would be best if he kept quiet and didn't interact with the crowd.\n\n9 Kelly Kelly\n\nvia tumblr.com\n\nKelly Kelly was a fan favorite during her time in WWE. She made her debut as a popular model who was wrestling on ECW. From there, she rose to the top as a valuable part of the roster. Kelly became a popular figure who had her own dedicated fans. She had the right amount of talent to be a decent worker, but it was her good looks that gave her the extra push.\n\nAfter leaving WWE in 2012, Kelly kept wrestling. She is back on the roster with TNA, but she has not appeared on a television episode of Impact Wrestling since December of 2015.\n\n8 Alberto Del Rio\n\nvia wrestlingrumors.net\n\nIt is a sad truth that Alberto Del Rio has fallen from his former position in the industry. He was a very big part of WWE. He was not only the United States Champion, but also a World Champion. Alberto had a unique presence as a superstar that could do things no one else could do. He also had the microphone skills to back up his moves.\n\nHis tenure in WWE was marred with incidents that made it clear he was not going to be there much longer. His time in Lucha Underground was more memorable, but it was a step down from what he used to be.\n\n7 The Great Khali\n\nvia cagesideseats.com\n\nThe Great Khali is not really a superstar that anyone really misses. His presence was laughable at times. The thing about Khali is that he had a lot of potential. He was very big and very strong, but he never seemed to have any clue about how to use those traits to his advantage.\n\nKhali is now an actor and a former wrestler. He is part of the reality show Khali Kisses, which involves him doing a lot of kisses on people. It is a far cry from his WWE days.\n\n6 Kelly Kelly's Boyfriend\n\nvia youtube.com\n\nWrestlers who are lucky enough to find a wife have a much easier time in the wrestling industry. The same could not be said for Mike Knox. He was only a superstar for a very short time, and he was a part of one of the most bizarre moments in wrestling history when he got the Attitude Adjustment from Kane. Knox is not talked about much these days, but there was a time when people were curious about him.\n\nHe married the popular WWE superstar Kelly Kelly. The couple got married in 2014. It was a marriage that was not meant to last, as the couple divorced in 2016.\n\n5 Santino Marella\n\nvia ringsidenews.com\n\nSantino Marella will forever be a part of WWE. The WWE Hall of Fame is home to some great wrestlers and some very bad wrestlers. Santino is among those who have made an impression for one reason or another. His career with the company was not as successful as he had hoped, but he made a huge impact. He was the guy who was hilarious because he was so bad.\n\nSantino's last match was in 2014. He has done some backstage work, but the star has pretty much disappeared from WWE.\n\n4 Tyson Kidd\n\nvia heavy.com\n\nTyson Kidd was a true performer in WWE. He had a unique look, and he could really do it all. The man was skilled on the mic, in the ring, and he had a good personality. Everything came together for Tyson, but that would all change during an unfortunate injury. Tyson was a mid-carder who could have risen to a top level, but that was not meant to be.\n\nKidd is now recovering from his neck injury. He is still hoping to get into the ring again, but it is looking unlikely.\n\n3 Christian\n\nvia bodybuilding.com\n\nChristian is not a name that is being mentioned much these days. It is a shame, as he is one of the best performers of his time. Christian has made a career of being known for his impressive in-ring work and his ability to entertain fans. Christian was a popular superstar during his time in WWE and he did some good things for TNA.\n\nChristian has not been in the ring since December of 2015. He is trying to get back into the ring, but there is a possibility that his neck injury may have ended his career.\n\n2 Dolph Ziggler\n\nvia wrestlingnewssource.com\n\nDolph Ziggler had the chance to become a very big star in WWE. Unfortunately, things have not been going well for Ziggler. He was once a main eventer, but he has fallen to the bottom of the card. He is one of the many guys who will be pushed as the underdog who is going to have the moment of their life.\n\nZiggler has not been the focus of WWE. It is likely that he will be entering a new chapter of his career in the near future. The most likely place would be in Ring of Honor, but WWE has shown a lot of interest in Ziggler.\n\n1 Wade Barrett\n\nvia wrestlingrumors.net\n\nIt is going to be very hard for the casual fan to remember Wade Barrett. He had a short run in WWE, but he did his part as a mid-carder. Barrett is a former Intercontinental Champion. He did well in WWE as a member of The Nexus, but his star power never quite took off.\n\nBarrett has been out of WWE since 2014. He has had some ups and downs in the independent circuit, but his career has largely gone downhill. It is possible that he could get another chance with WWE. That would probably be the only way that fans would get to see him again.\n\nBarrett has been trying to make a name for himself in the business world. He launched a website last year. It is not likely that he will become a success, but it is nice to see him try to make something of his life.]" time="3.927"><properties><property name="score" value="0.9997806463922773" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[University: The primary mission of the University is to provide education. For that purpose, the University has established different faculties and departments, of which the oldest and most important is the faculty of theology. The latter is a privileged entity, by virtue of its academic reputation, both in the local and the global sphere. The mission of the University is to educate the professors and the students in the science of theology, which means, it also needs to equip them with the specific professional skills, which make them able to exercise the profession of the ministry.]" time="0.277"><properties><property name="score" value="0.9997656941413879" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The brother-in-law of the Palestinian-American teenager who was filmed being brutally beaten by Israeli police at a West Bank checkpoint says the incident \u201chas put a wedge\u201d between the Palestinians and Israelis.\n\nYoussef Almogy said he has had more than 50 phone calls from news media outlets following the brutal arrest of 16-year-old Tariq Abu Khdeir, the Associated Press reported.\n\n\u201cWhen I saw it, I cried,\u201d Almogy said. \u201cI was so shocked, but after that, I started thinking about my brother, who was jailed for 11 years, and then released for nothing. And I said, \u2018If they can do that to him, then imagine what they can do to us.\u2019\u201d\n\n\u201cAnd this has put a wedge between Palestinians and Israelis,\u201d he added.\n\nAlmogy was referring to the arrest of his brother, Amer, who was jailed for 11 years after he and a group of Palestinians tried to rob a Tel Aviv bank in 1994. Almogy said Amer was never convicted of a crime and that he believes the former prisoner is being held because of his brother\u2019s connection to the teen, the AP said.\n\n&quot;We're used to seeing this sort of thing,&quot; he said.\n\nAbu Khdeir, who was born in Florida but living with his family in East Jerusalem, was arrested Thursday after he was identified as being involved in a clash between]" time="0.775"><properties><property name="score" value="0.9790650010108948" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[\n\nToday we\u2019re excited to announce that Bethesda Game Studios has joined Oculus as a developer for the Oculus Rift. The partnership will see the creation of two exclusive, made-for-VR games from Bethesda Game Studios; the first is currently being designed for virtual reality and the second is being built for traditional monitors.\n\n\n\n\u201cAt Oculus, we\u2019re dedicated to helping developers and partners to grow, and the support and dedication we\u2019ve received from Bethesda Game Studios has been outstanding,\u201d said Brendan Iribe, CEO of Oculus. \u201cJohn Carmack, founder of id Software and now CTO at Oculus, believes in our vision and has created a custom version of Doom 3 BFG Edition for Oculus. We look forward to continuing to work with Bethesda as we develop our first generation of virtual reality games and experiences.\u201d\n\n\n\n\u201cI have a lot of personal interest in making sure that the VR gaming space is successful because I really believe in the technology and its potential to change gaming,\u201d said Carmack. \u201cAs a developer, the Rift is something that I\u2019ve wanted to work with for a long time, but the existing technology wasn\u2019t quite there yet. I\u2019m happy to be able to help to make the Rift a success and contribute to the rebirth of VR gaming.\u201d\n\n\n\nAt Bethesda Game Studios, developers are working on a number of unannounced projects including new entries in major franchises. Those who have seen this year\u2019s E3 show floor, Gamescom or Quakecon presentations have seen teases of what the studio is up to and a few of the titles we have in development are also being designed for virtual reality. The two games announced today are entirely new IP and we can\u2019t wait to show them to you when they\u2019re ready for primetime.\n\n\n\nFor those of you unfamiliar with Bethesda Game Studios, the acclaimed developer is responsible for creating many of the industry\u2019s most popular games including the \u201cElder Scrolls\u201d series, the single-player add-on to the original \u201cDoom\u201d and the \u201cFallout\u201d series. In fact, \u201cFallout 3,\u201d \u201cFallout: New Vegas\u201d and \u201cThe Elder Scrolls V: Skyrim\u201d were each awarded game of the year at the 2008, 2010 and 2011 BAFTA awards, respectively.\n\n\n\n\u201cWe have made some long-term technology investments with VR in mind,\u201d said Todd Howard, Game Director and Executive Producer at Bethesda Game Studios. \u201cHardware is finally catching up to the visions developers have had for virtual reality. Now when you put on a VR headset, you\u2019re transported to new and amazing worlds.\u201d\n\n\n\nThis year, Bethesda Game Studios showcased \u201cDoom 3 BFG Edition\u201d for Oculus. \u201cDoom 3 BFG Edition\u201d is a complete overhaul of \u201cDoom 3\u201d that brings modern technology and a classic, storied franchise into the world of virtual reality. This is the first title designed for virtual reality.\n\n\n\n\u201cJohn Carmack\u2019s work in the early 90s helped usher in the era of modern 3D gaming, and his genius continues to inspire developers to innovate,\u201d said Brendan Iribe, CEO of Oculus. \u201cThe team at Bethesda Game Studios is made up of the most talented developers in the industry and we can\u2019t wait to bring their newest games, made for virtual reality, to life.\u201d\n\n\n\nOver the past 20 years, Bethesda Game Studios has earned its reputation as one of the industry\u2019s most respected and accomplished game development studios. Creators of the 2006 and 2008 Game of the Year, \u201cThe Elder Scrolls IV: Oblivion\u201d and 2010 Game of the Year, \u201cThe Elder Scrolls V: Skyrim,\u201d Bethesda Game Studios has earned a reputation for excellence and quality with each of its five releases.\n\n\n\nThis partnership is another step toward realizing the future of virtual reality. It\u2019s an exciting time to be a gamer with the release of these groundbreaking VR titles, the arrival of next generation consoles and the continued evolution of the PC.\n\n\n\nStay tuned for more news on Bethesda Game Studios\u2019 titles for Oculus.\n\n\n\n\u2013 The Oculus Team]" time="2.125"><properties><property name="score" value="0.9480832517147064" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[\n\nHi~ B een a long time no see. Thank you for the continued support on the blog. So, how have you all been doing?\n\nToday, I have a special post. I know most of you do not know but it is the 'World Animal Day'. The day is celebrated worldwide on October 4 to create awareness of animal rights and welfare. The day is a very special one for me because I am a vegetarian, animal lover and a dog-person.\n\n\n\n\n\nSo for the last one year, I have been trying to spread the awareness among the people. I have been trying to encourage people to stop using animals as fur coats, cosmetics and for food. I want people to know the feelings of these animals, the pains they suffer and what they go through while they are alive. Also, the environmental damage they cause to the planet. It is very important to spread this message because animals are treated as if they are some sort of toys that we can play with.\n\n\n\n\n\nToday, I would like to share with you some tips on how we can help the animals and make the planet a better place to live. I also have a small personal pledge to make. Also, if you have not signed the pledge yet, you can also make your own pledge to protect the animals and the planet.\n\n\n\n\n\n\n\n\n\nTips and pledges to help the animals\n\n(made by my friend!)\n\nOne way we can make a change is by making an effort to not use cosmetics that are tested on animals. We can choose to use all cruelty free products and make a conscious effort not to use anything that has been tested on animals.\n\nWays to reduce our carbon footprints and help protect the environment:\n\nAvoid eating meat, or even better, go vegetarian (dairy, eggs are not vegan but I still encourage you to stop using dairy products as well)\n\nUse eco-friendly products - Not just cosmetics but also in our everyday lives.\n\nGo cruelty free - Choose cruelty free products and wear clothes that have not been made out of leather, fur or wool.\n\nStop buying those animals that we do not need. Also, choose to buy second hand or recycled clothes. Avoid buying pets if you cannot provide the proper care for them.\n\nAdopt, don't buy! You can adopt a dog or cat from an animal shelter. Do not buy animals from pet stores or breeders. They are sold to you as &quot;breed-free&quot; animals but most of them have been bred in horrible conditions. Also, make sure you know the sex of the animal you are getting, because they usually are sold as 'un-neutered' or 'un-spayed' animals.\n\nMake a conscious effort to not use plastic bags. Use reusable bags or better still, try to avoid using bags at all.\n\nTry to use products that do not harm the environment and are biodegradable.\n\nWays to help the animals:\n\nSign the pledge - To show that you support the cause to protect the animals.\n\nShare the post - Let others know of the cruelty that the animals go through. Let them know how they can help the animals too.\n\nJoin or start a protest or campaign to help the animals.\n\nThe small pledge I am making is to help the animals. Every month, I will be donating some money to an animal shelter in my town to help feed the dogs and cats there.\n\nThese are just some of the ways we can help the animals and the planet. We can all make a change, and make this planet a better place to live.\n\n\n\n\n\nIf you want to know how to become a vegan, and which cosmetics are tested on animals, you can read this post: How to be a vegan?\n\nRemember, we are what we eat, so be what you want to be. Be a good person, who loves animals and helps to save the planet.\n\nWith lots of love,\n\nPriya]" time="1.964"><properties><property name="score" value="0.9960466921329498" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Top Customer Reviews\n\nWhere reviews refer to foods or cosmetic products, results may vary from person to person. Customer reviews are independent and do not represent the views of The Hut Group.\n\nUnique, gritty, and wonderful The Good: The worlds are some of the best I've seen in a while, and how you explore the city and interact with its citizens gives a great sense of authenticity and intrigue. As a huge comic fan, the references are abound. The story and narrative is one of the best I've ever experienced. The Bad: The game is very linear and the combat while satisfying can get old at times. Overall: This is one of the best games I've ever played, period. Everything about it is gritty, dark, atmospheric, and enticing. While you have a very linear path to follow in the main story, the side quests and interactions you can have with the citizens is one of the best I've seen. You are truly the villain in this game, so if you want to be good and helpful to everyone, this is not for you. But for the fans of dark, gritty stories, this is one of the best. The story and narrative are top notch. The villains are some of the most compelling I've ever seen in a video game, with even the main villains being so much more compelling than the good guys. I loved this game from start to finish, and can't wait to see where the series goes from here. The ending while great is somewhat confusing, so if you aren't]" time="0.774"><properties><property name="score" value="0.9971693456172943" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Mrs. Mahilani Konczey\n\n&quot;The tutors at HW have made a huge difference in my child's academic performance, as well as in her confidence. The tutors are friendly, professional and punctual. My child enjoys her sessions and she has learned a lot from them. HW has a good location and easy parking. I highly recommend them.&quot;]" time="0.203"><properties><property name="score" value="0.9992594122886658" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99925941&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99925941
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Flavor Pairs\n\n(Home)\n\nPickled Meat\n\nBacon, Flank Steak, Pork Butt, Bacon, Hams, Sausage, Canned Ham\n\nFor Cooking/Grilling\n\nPeppers, Onions, and other Vegetables\n\nPotatoes, Onion, Green Peppers, Jalapenos, Garlic, and other Vegetables\n\nPastas\n\nPasta, Macaroni, Shells, Ramen Noodles, Ziti, Gnocchi, Fusilli, and other Pastas\n\nSoups and Stews\n\nSoups, Stews, Chili, and other One-Pot Meals\n\nCorn, Peas, and other Grains\n\nCorn, Peas, Risotto, Polenta, Grits, Quinoa, Rice, and other Grains\n\nBeverages\n\nBeer, Alcohol, Soda, Juices, Tea, Coffee, Soda, and other Beverages]" time="0.492"><properties><property name="score" value="0.9997889399528503" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[And now, ladies and gentlemen, I give you the breathless high-speed ranting I did when I saw the (potentially) wackadoo conclusion to this storyline.\n\nOh, my God, you guys, this episode has a freaking atomic bomb on the side of it.\n\nOkay, cool. It\u2019s the mid-season finale. I know there\u2019s going to be some serious plotlines.\n\nAnd seriously, what was up with the weird science in this episode?\n\nNo. No, I did not just watch an episode with the headline: \u201cOuch, My Butt!\u201d\n\nThat plotline was weird as hell.\n\nBut I can understand why they did it. He\u2019s trying to be a \u201csophisticated ladies man\u201d and not the boy they\u2019ve been dealing with all these years.\n\nOh, now this is hilarious. But it\u2019s also dangerous, if you\u2019ve watched any superhero movie ever.\n\nIn the brief moment that I was allowed to breathe, I actually enjoyed this part.\n\nNot only is this scientifically improbable, but also\u2026 how does a security camera shoot a laser?\n\nI know this is only in fiction, but I do believe that they\u2019re committing suicide.\n\nThis is also part of the reason why I wanted to talk about this episode. I know it\u2019s not a good look. It\u2019s not exactly smart of him to do this.\n\nI don\u2019t even know what to say here.\n\nWHY?\n\nHe\u2019s not afraid of bugs, but he\u2019s scared of that snake?\n\nI mean, yeah, I get that he\u2019s doing it for a different reason, but\u2026\n\nIf it weren\u2019t for that, then I would not have laughed so hard at this.\n\nOkay, look, I was done with this season for a while. I thought that it was becoming a boring clich\xe9d plot that they were sticking to, so I was ready to say goodbye to it. But then I watched this, and I just felt like I was watching a new show. There was so much going on, and they were throwing so much at us at once, and it was just such a good use of time. I\u2019m really happy that they had this show, because it was seriously pretty funny and it was also action-packed. There was a lot going on, but it was very clear, and it was fun to watch. I just loved how much was going on. I loved it, and I really did enjoy it. It was a bit much, and they took a lot of crazy chances, but it was just so fun to watch. That was some crazy stuff, and it was nice to see something so good after so much of the regular garbage we were getting. Overall, I loved it.]" time="1.533"><properties><property name="score" value="0.9058307210604349" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Specification:\n\nProduct Name Bags Model SA01A Material Canvas + Zipper Size 29*15.5*10.5cm Color Red, Green, Black, Blue, Gray, Purple, Brown, Khaki, Yellow, Orange Weight 10g Quantity 1 Piece(s) Pattern Solid Inner Pocket Yes Closure Drawstring\n\nPackage include:\n\n1 X Canvas Bags\n\n\n\n\n\nDescription:\n\nDimensions of the storage bag, cool way to store the supplies for travel and daily life.\n\nLightweight, durable and sturdy canvas construction.\n\nA simple pouch, plenty of storage space to put many small items.\n\nThere is a handle for easy carrying and a good choice for daily life.\n\nUsed for carrying, practical, suitable for outdoor, camping, traveling, adventure, daily life, perfect for storing jewelry, makeup, keys, etc.\n\n\n\n\n\nSize: 29*15.5*10.5cm\n\n\n\nWeight: 10g\n\nMaterial: Canvas + Zipper\n\n\n\nColor: Red, Green, Black, Blue, Gray, Purple, Brown, Khaki, Yellow, Orange\n\nPattern: Solid\n\nInner Pocket: Yes\n\n\n\nClosure: Drawstring\n\n\n\n\n\n\n\nPackage include:\n\n1 X Canvas Bags]" time="0.584"><properties><property name="score" value="0.9935429096221924" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Author:\n\nFollowing a stint as a park ranger in Wyoming, Dan Beal relocated to Seattle and received a degree in science writing from the University of Washington.\n\nFollowing a stint as a park ranger in Wyoming, Dan Beal relocated to Seattle and received a degree in science writing from the University of Washington.]" time="0.208"><properties><property name="score" value="0.9994940757751465" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[(Adapted from the 2014 Urban Tree Canopy Conference)\n\nPotholes are problematic for all of us, and for our city\u2019s trees they represent a real danger. Roots can become entangled in the small cracks between the asphalt and the cement to create the impression that the tree is growing out of the road! And when the roots do become entwined, the pressure can be so great that the roots are destroyed. We all know the danger this creates for the trees.\n\nThe solution to this problem is obvious: pave over the whole road surface. Not so fast! In the City of Chicago, for example, there is a practice that preserves the utility of the tree\u2019s roots and of the road surface itself. The practice is called \u201cmilling\u201d. Milling involves cutting the asphalt out in a shape that will not disturb the roots and adding back in a new surface. While the tree is protected, the roots are not disturbed and the road surface is left intact. When a tree is scheduled for removal, this is often the method of choice. Milling also involves improving the root zone around the tree. The roots of many trees can\u2019t penetrate the compacted soil that is found beneath pavement, so it is necessary to mill the area to promote root growth.\n\nMilling is less expensive than paving over a road surface and can also be less expensive than removal of a tree. Milling should be done whenever possible and certainly before a tree is scheduled for removal. There is no need to sacrifice a mature, healthy tree.\n\nAdvertisements]" time="0.785"><properties><property name="score" value="0.026988595724105835" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[How do i search for files with a specific suffix using Unix?\n\nOn Unix, to search for files that ends with a particular file extension, use the following command:\n\nfind . -name '*.html'\n\nOR use grep command to list all files in a directory and subdirectories that have the suffix:\n\nfind . -type f | xargs grep -l &lt;suffix&gt;\n\nSample outputs:\n\n./Ajax/contact-us/index.html ./test/index.html\n\nThe above commands are equivalent to the grep command on Linux/MacOSX or the find command on Linux/MacOSX:\n\n$ find . -name &quot;*.html&quot; ./test/index.html ./Ajax/contact-us/index.html $ grep -l -Ri -f -E &quot;.*\\.html&quot; ./ test/index.html Ajax/contact-us/index.html\n\nFind out all files with a specific filename (using shell pattern match):\n\nfind . -name '*.html'\n\nOR use the egrep command to list all files in a directory and subdirectories that have the suffix:\n\nfind . -type f | xargs egrep -l &lt;suffix&gt;\n\nSample outputs:\n\n./Ajax/contact-us/index.html ./test/index.html\n\nOR\n\n$ find . -name &quot;*.html&quot; ./test/index.html ./Ajax/contact-us/index.html $ egrep -i &quot;.*\\.html&quot; . Ajax/contact-us/index.html test/index.html\n\nFind files in a directory and subdirectories:\n\nfind . -name &quot;*.html&quot;\n\nOR\n\nfind . -type f | xargs grep -l &lt;suffix&gt;\n\nSample outputs:\n\n./Ajax/contact-us/index.html ./test/index.html\n\nOR\n\n$ find . -name &quot;*.html&quot; ./test/index.html ./Ajax/contact-us/index.html $ egrep -i &quot;.*\\.html&quot; . Ajax/contact-us/index.html test/index.html\n\nFind files in a directory and all]" time="1.194"><properties><property name="score" value="0.005786925554275513" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Russia has expanded its ban on U.S. pork to include all American meat imports as the diplomatic row over alleged election meddling and the situation in Syria rumbles on.\n\nLast week, the Russian government added five new countries to the blacklist: South Korea, Montenegro, Iceland, Liechtenstein and Ukraine. While the list included major agricultural producers such as Brazil and Canada, U.S. meat producers were noticeably absent. However, on Wednesday, the government added the U.S. to the list of banned meat importers.\n\nThe ban was announced by the state consumer watchdog Rospotrebnadzor. A government official told Russian media that American meat imports were now being kept out of Russia on a \u201ctemporary basis\u201d.\n\nRussia is the fourth largest importer of U.S. meat, buying $51 million worth of poultry and $22 million of pork in 2016. According to the National Pork Producers Council, the U.S. shipped 27,600 metric tons of pork to Russia in 2016, which made up 0.3 percent of total U.S. pork exports.\n\nRead more: Polls Say Few Americans Think Trump is Putin's Puppet\n\nThe ban comes at a particularly difficult time for the U.S. meat industry, which is still struggling to recover from a disastrous salmonella outbreak that began in 2015. China, Vietnam and Mexico have all banned U.S. pork imports, while the EU and South Korea are still testing shipments.\n\nThe bans come as U.S. meat exporters have had difficulty convincing importers to trust American meat following the 2015 crisis. The outbreak, which resulted in more than 60 percent of pork being recalled, was linked to at least 184 illnesses in 20 states. While U.S. authorities traced the outbreak back to a farm in Missouri, they were unable to identify the exact source of contamination.\n\n\u201cThis has been a major setback for us and it will take a while for the industry to recover,\u201d said National Pork Producers Council spokesman Dave Warner.\n\nIt's not clear when Russia will lift its ban on U.S. meat, but Russia's state consumer protection watchdog has said it would allow Ukrainian meat imports.]" time="1.160"><properties><property name="score" value="0.8646958470344543" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Is there anything you would change about this book?\n\nIt's too long. I got bored. It's the wrong length for a book about the Crusades. I know you want to make it as comprehensive as possible, but let's face it, the Muslim perspective is the only one that matters. You can understand why people fought for this or that minor piece of territory, but you can't understand why they fought in the first place.\n\nHow would you have changed the story to make it more enjoyable?\n\nI wish they had done a better job of establishing the atmosphere of the time, instead of just going for sensationalism.\n\nHow could the performance have been better?\n\nThe narrator had a good voice for this.\n\nWas Crusade the right choice for you? Why or why not?\n\nNo, I'd never read about the Crusades before.]" time="0.395"><properties><property name="score" value="0.9998273253440857" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Most Helpful Positive Review\n\nI would purchase this product again\n\nI have used a baby wrap before for a previous baby and loved it, but this one is my favorite. The material feels great and is breathable, the design is cute and easy to use, and I love...Read complete review\n\nI have used a baby wrap before for a previous baby and loved it, but this one is my favorite. The material feels great and is breathable, the design is cute and easy to use, and I love how it fits under my shirt to show a little but isn't quite as obvious as a sling. I use this for my 1 month old and we have never had any problems with the carrier not staying tight enough. I would definitely recommend this product!]" time="0.342"><properties><property name="score" value="0.9998278617858887" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[A northwest Houston woman is under arrest, accused of child abuse.\n\nPolice say they were called to the woman's home on Almeda School Road around 10:45 a.m. Thursday. They say her husband had to take their child to the emergency room after the child suffered a burn.\n\nPolice say the child suffered burns to the back of his neck, shoulder and upper back area.\n\nPolice say the wife, 31-year-old Mercedes Victoria Gonzalez, was arrested.]" time="0.304"><properties><property name="score" value="0.9994427561759949" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[(NC)\u2014Despite the odds, there are some consumers who decide to head to the stores this weekend to start Christmas shopping. For some, the reason is because they still have some of their shopping list to complete. Others, however, have an entirely different reason.\n\nAccording to Statistics Canada, the average Canadian adult will spend $473.12 on gifts this holiday season. That means the average consumer will spend almost $35 more this year than last year.\n\n\u201cA lot of consumers have gone overboard,\u201d says Ted Michalos, founder of CMG Financial. \u201cIt\u2019s not necessarily the spending that\u2019s the problem, it\u2019s what they\u2019re spending the money on.\u201d\n\nThat\u2019s because many shoppers are buying gifts for themselves. That doesn\u2019t necessarily mean they\u2019re spending on the latest fashions, but rather toys and electronics. And the majority of these gifts, at 41.8 per cent, are going to be purchased online.\n\nHowever, the reason why so many people are spending online is because they\u2019re trying to avoid the crowds and waiting in lines.\n\n\u201cSome consumers just get too stressed and they go online to get their stuff,\u201d says Michalos. \u201cAnd if they do that, they\u2019re likely going to be paying more for the gifts.\u201d\n\nThat\u2019s because while there are fewer people at the store, they\u2019re paying more per person. For example, if someone is paying $50 for a gift in the store, they\u2019re likely to pay more than $50 online.\n\n\u201cThis is really the time where you\u2019re going to have to do your best to have everything done by the 15th,\u201d says Michalos. \u201cAnd it\u2019s definitely a good idea to sit down with your spouse or significant other and look at your budget. And when you have a child, the best thing to do is to write down what you\u2019re going to spend. That\u2019s a real good thing to do, because if you don\u2019t, you\u2019ll overspend.\u201d\n\nAnd that could come with consequences, because if you\u2019re overspending, you could be limiting your ability to spend on your children or other family members in the future.\n\n\u201cYou can go overboard and give away everything you have to your kids and that could be a problem,\u201d says Michalos. \u201cThis could be an opportunity to teach your kids about moderation, and also the difference between \u2018I want\u2019 and \u2018I need.\u2019\u201d\n\nBecause with Christmas, most people are shopping with the mindset of \u2018I want,\u2019 not \u2018I need.\u2019 And that can be a problem, because it will prevent you from saving or investing.\n\n\u201cIf you\u2019re going to be shopping on a credit card, you\u2019re going to be in big trouble,\u201d says Michalos. \u201cIf you\u2019re not able to pay off that debt by the end of January, you\u2019re going to be in big trouble.\u201d\n\nWant to know more about debt? Try these articles:\n\nNew Christmas ad showcases true meaning of the holidays\n\nTake advantage of these holiday deals\n\nHow to get your kids to be thankful for what they have\n\nWhat do you think? Join the conversation in the comments below or contact us using the \u201cHave Your Say\u201d form on this page.]" time="1.841"><properties><property name="score" value="0.9997683167457581" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Developing in a Virtual Machine on your computer or laptop\n\nHello , Here we are going to learn how to create virtual machine on your computer for testing purpose. Before we start , it is better to know what is virtual machine. A virtual machine is a simulated hardware that allows you to run other operating systems (or other versions of the same OS) at the same time as your current OS, inside a window on your current OS. It can also be thought of as a computer inside a computer. With a virtual machine, you can: Run another operating system (or another version of the same operating system) at the same time as your current operating system. Try out a new operating system without having to install it on your hard disk. Run applications for one operating system on another, without having to reboot your computer.\n\nStep 1: Create a New Virtual Machine\n\nVirtualBox provides a wizard that will assist you in creating a virtual machine. Launch the program and select \u201cNew\u201d.\n\nStep 2: Create a Virtual Machine\n\nThe first window you see is the New Virtual Machine Wizard. Select \u201cCreate a virtual machine\u201d.\n\nStep 3: Create a Virtual Machine\n\nIn the next window, enter the name for your virtual machine. This name will appear in the main VirtualBox window next to the icon for the virtual machine.\n\nStep 4: Create a Virtual Machine\n\nThe next window is where you select the operating system you want to install.\n\nStep 5: Create a Virtual Machine\n\nIn the next window, select the amount of memory you want to allocate to your virtual machine. The more memory you have, the more applications you can run simultaneously. However, keep in mind that the more memory you allocate, the slower your computer will be.\n\nStep 6: Create a Virtual Machine\n\nSelect \u201cCreate a virtual hard disk\ufffd]" time="0.867"><properties><property name="score" value="0.9991336464881897" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The internet is a funny place, where people can get away with saying and doing the weirdest stuff that would otherwise be totally unacceptable.\n\nBut it turns out not even the internet is a free pass to sexually harass people.\n\nThat's what one woman learned when she was sent a totally inappropriate, NSFW photo from a stranger, that happened to be someone she knew in real life.\n\nTwitter user @kendtee, who describes herself as &quot;an arm candy&quot;, tweeted about her encounter with the man she knew.\n\nAccording to @kendtee, the incident occurred after she agreed to take a picture with a guy. She then deleted the picture from her camera roll when it was over, and the man asked if she had. She said yes, and then he said this:\n\n&quot;It's a shame we can't see the photo anymore because it was pretty good&quot;\n\nShe didn't know how to respond. He then tweeted her back, asking her to &quot;leave [the photo] up for a little while&quot;. She declined.\n\nApparently, he then texted her back, asking her to send him a copy of the picture.\n\nShocked by this behaviour, she shared the exchange on Twitter, and said the worst part was how &quot;unbothered&quot; he seemed about it.\n\nEventually she shared a screenshot of the exchange, and people were not impressed.\n\n&quot;Oh wow. 'I'd do ya if I could' and then the photo request. You should probably just block him and be done with it. 'Doing ya' or not,&quot; one person said.\n\nAnother person had an idea on how to deal with it, tweeting, &quot;You know what? Be a bigger person, send it to him. For posterity.&quot;\n\nMany others were in complete agreement with this sentiment.\n\n&quot;I don't know this man and I probably never will. But I 100% understand why he's going to be blocked. A man who says that is definitely not a good person. All women should know their boundaries with men and this seems to me to be clear cut,&quot; one person said.\n\nIn the end, @kendtee went back to him and asked him to explain why he'd sent her that photo.\n\nAccording to @kendtee, he sent her this message in response:\n\n&quot;How do you think I know that you're gonna see this tweet and block me or delete it.\n\n&quot;I'm not trying to be a creep you just said that you'd do me so I figured we had a shot lol\n\n&quot;I like your style and want to get to know you better\n\n&quot;If you send the picture I promise not to be creepy and stop bothering you\n\n&quot;I'm being serious here and I don't think I deserve the reaction I got\n\n&quot;All I wanted to do was take a picture with a nice girl and see if she'd let me\n\n&quot;I don't want to be a bad person or make you feel uncomfortable.\n\n&quot;If you like I'll just go away\n\n&quot;You're gorgeous and I'm sorry I don't deserve to meet you but I thought I'd try my luck\n\n&quot;I won't message you again if you don't want me to\n\n&quot;I'll delete the tweet and block you\n\n&quot;I just wanted to try\n\n&quot;Please&quot;\n\nThe rest of Twitter wasn't impressed, calling him out for still not getting it, and for trying to shift the blame.\n\n&quot;Like you have to explain what 'likes your style' means, like you're trying to shift the blame. But it's not ok. It's not ok for you to message someone for a photo. It's not ok for you to say 'I'd do you' and it's not ok for you to try and justify your actions,&quot; one person said.\n\nAnother pointed out that he was trying to pressure her into sending him the photo, and wasn't taking responsibility for his actions.\n\n&quot;He's a creep, and still wants to force you to send him a picture of your body. He'll delete the tweet if you do, otherwise you can block him. This is on him, not you. You have no responsibility to 'meet' him,&quot; another person said.\n\n&quot;Ok let's stop there with the 'I would've taken a photo with you anyways' thing. That's not what this is about. This is about someone sending unsolicited nudes,&quot; another person pointed out.\n\nIn response, @kendtee wrote that she wasn't bothered, because at the end of the day she was a &quot;pretty woman who didn't need to be respected&quot;.\n\n&quot;I didn't get mad or freak out or feel the need to drag his name through the mud. I just calmly informed him that I have a man in my life and that I'm flattered but I don't share photos like that and I'm taken. He didn't like it so he blocked me and I'm okay with that,&quot; she wrote.\n\nALSO ON HUFFPOST AUSTRALIA]" time="2.607"><properties><property name="score" value="0.9016384959220887" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Perhaps one of the most beautiful areas of Southern California is the Pacific Palisades, a neighborhood filled with high rises, luxury estates, and tons of culture. From the obvious to the under the radar, Pacific Palisades has some of the most fascinating museums, gardens, and activities you can visit. When you need to get out of the house and explore a new part of Los Angeles, the Pacific Palisades is a fantastic choice.\n\nThe Getty Villa\n\nThe Getty Villa is a sight to behold, and is a must see when you are in the Pacific Palisades. You are welcome to explore the museum for free on the first Thursday of every month. They offer a free trolley that will take you to the Getty Villa, so you don\u2019t have to worry about navigating through the Pacific Palisades on a crowded day. This is the perfect time to visit the Pacific Palisades.\n\nThere are two events going on at the Getty Villa on Thursdays: Story Time and Pacific Voices. Story time will give you a better understanding of ancient Greek and Roman culture. It is a good idea to purchase your tickets in advance, so you can get a discount. The Pacific Voices event is a singing performance, and a great way to enjoy your time at the Getty Villa.\n\nSanta Monica Mountains National Recreation Area\n\nThe Santa Monica Mountains are a beautiful way to spend a day in the Pacific Palisades. The National Recreation Area has a fantastic number of trails you can explore, and is a great way to get out of the house and spend time with your family. Take the trail to Leo Carillo State Beach, or go for a hike along Rustic Canyon. Whichever trail you take, you are sure to enjoy your time in the Pacific Palisades.\n\nSanta Monica Baykeeper\n\nThe Santa Monica Baykeeper is a fantastic organization that works hard to clean up the Santa Monica Bay. This is a great way to spend time in the Pacific Palisades and make a real difference. Check their website for upcoming volunteer opportunities. You will learn more about the Santa Monica Bay, and get a chance to spend time with some interesting people.\n\nPacific Palisades Center for the Arts\n\nThe Pacific Palisades Center for the Arts is an excellent way to spend an afternoon. This isn\u2019t a museum, so you don\u2019t need to book in advance. It is free to walk around the grounds, and to watch a performance. They have everything from jazz concerts to visual arts. If you want to learn more about the Pacific Palisades, then this is a great place to start.\n\nPacific Palisades Farmers Market\n\nThe Pacific Palisades Farmers Market is an excellent way to spend your Saturday in the Pacific Palisades. From fruits and vegetables, to homemade items and baked goods, there are plenty of delicious items to choose from. This is a great place to bring the kids, as they have tons of activities.\n\nPacific Palisades Art Walk\n\nThe Pacific Palisades Art Walk is a fantastic opportunity to view some of the beautiful artwork created by locals. It is a fun evening filled with laughter and good food. You are welcome to bring your own picnic basket and enjoy it in the beautiful Pacific Palisades.]" time="1.576"><properties><property name="score" value="0.9997794032096863" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[We have been very pleased with the support, workmanship and professionalism of the team at WeFix. They have been able to provide quick and effective solutions for any problems that have arisen. The attention to detail, willingness to understand and understand the importance of high quality work have been important to us. The flexibility of being able to accommodate and adapt to any situations has been very helpful in terms of our team and their working styles.\n\n- Rick Phillips, Site Manager \u2013 Maersk Oil]" time="0.229"><properties><property name="score" value="0.9998255372047424" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Mastering financial knowledge and planning for a good life is a lifelong process. But when it comes to money, how many people have the knowledge to plan for a good life? And what if you had this information, would you use it? How would you live? Would your life change for the better? Would your relationships improve? Would your life be happier?\n\nIn most cases, we don\u2019t have enough financial knowledge to make the right decisions. We don\u2019t have the knowledge and therefore, we end up making bad decisions. The result of these bad decisions is that we lose more than we gain. We lose money and we also lose our peace of mind.\n\nI have experienced both sides of the coin and I know the emotional, mental and physical impact of being financially unstable. Money makes and breaks people. It has broken me in the past and it took me years to recover. But I am better now.\n\nIf you too want to make peace with your money]" time="0.416"><properties><property name="score" value="0.9997904896736145" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Tom Shanks (writer)\n\nTom Shanks (born 1961) is a Scottish writer, journalist, and musician. He is the author of eight novels, including &quot;The Hidden&quot;, the &quot;Vampyrrhic Times&quot; trilogy, and &quot;Eve&quot;, an &quot;Eve&quot; (comics) spin-off novel.\n\nShanks has been living in France since 1983 and is married to the French writer Marie H\xe9l\xe8ne Poitras.\n\nShanks was born in Edinburgh, Scotland in 1961.\n\nShanks has been writing since the mid-1990s and has published several novels.\n\nShanks has also been writing for the British comic book magazine &quot;2000 AD&quot;. His first script for them was the short story &quot;Trevor's Tale&quot;, drawn by Simon Harrison, in &quot;2000 AD&quot; #1704, published in March 2008.\n\nThe comic book &quot;Eve&quot; was written by Shanks. It was released by Black House Comics in 2011 and the second volume was published in 2012. In 2012, IDW Publishing published a comic book prequel to &quot;Eve&quot;, &quot;Eve: The Awakening&quot;. The book was written by Shanks and Marie H\xe9l\xe8ne Poitras and illustrated by Enrica Angiolini.\n\nShanks has also written the script for &quot;Apocalypse Cow&quot;, which was originally published in French and German as &quot;La Vache Apocalypse&quot;. It was adapted as a comic book by Tony Lee.\n\nIn March 2010, the novel &quot;Vampyrrhic Times&quot; was published by Gollancz in the UK and by Roc in the US. It was illustrated by Tony Harwood. The second volume of the trilogy, &quot;The Wisdom of Dead Men&quot;, was published in the UK in 2011 and the US in 2012. The third volume, &quot;The Trade of Queens&quot;, was published in the UK in 2013 and in the US in 2014.\n\n\n\n\n\n\n\nShanks and Poitras run the Black House Comics publishing house.\n\nShanks is a member of the band The Bassholes, along with co-writer of Eve, Marie H\xe9l\xe8ne Poitras. The band is composed of Shanks and Poitras on vocals and bass, Poitras' husband, French artist St\xe9phane Paitreau, on guitar, and drummer &quot;Chino&quot; Gonzales. They released their first album, &quot;No More Mr Nice Guy&quot;, on May 1, 2010, on Stuck on a Name Records. The album includes covers of classic rock songs, such as &quot;Jailhouse Rock&quot; and &quot;Born to be Wild&quot;. In 2013, the band released a second album, &quot;Sex Lies &amp; Basslines&quot;.\n]" time="1.421"><properties><property name="score" value="0.4642360806465149" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Rick Lake\n\nRichard G. Lake II (born April 7, 1979) is an American mixed martial artist who most recently competed in the Welterweight division of the Ultimate Fighting Championship. A professional competitor since 2004, he has also competed for the Quad Cities Silverbacks of the IFL, and was a competitor on Spike TV's &quot;&quot;.\n\nBorn and raised in the Quad Cities area, Lake attended East Moline High School.\n\nAfter graduating from high school, Lake worked for two years as a full-time welder, before deciding to pursue a career in mixed martial arts.\n\nLake made his professional mixed martial arts debut on October 30, 2004 for the Quad Cities Silverbacks of the International Fight League. He won his first three fights for the promotion, before suffering his first loss to Ian Loveland on April 8, 2005.\n\nAfter picking up three wins in the IFL, Lake was signed by the UFC. He was expected to make his debut at UFC Fight Night 7 against Jason Gilliam. However, Gilliam was forced from the card with an injury, and was replaced by Jorge Rivera. Lake lost to Rivera via first-round KO on August 23, 2007.\n\nFor his second fight in the promotion, Lake dropped down to Lightweight and faced Kyle Bradley at UFC Fight Night 11 on November 17, 2007. He won the fight via unanimous decision (30-27, 30-27, 29-28).\n\nIn his third fight for the promotion, Lake dropped to Featherweight to take on Mike Brown at UFC Fight Night 13 on June 16, 2008. He lost the fight via unanimous decision (29-28, 29-28, 29-28).\n\nFor his fourth fight in the promotion, Lake dropped to Bantamweight and faced Brad Pickett at UFC 95 on February 7, 2009. He lost the fight via TKO in the first round.\n\nFor his fifth fight in the promotion, Lake faced Tommy Hayden at UFC 100 on July 11, 2009. He won the fight via submission in the first round, earning &quot;Submission of the Night&quot; honors.\n\nFor his sixth fight in the promotion, Lake faced George Roop on March 21, 2010 at . He won the fight via unanimous decision.\n\nFor his seventh fight in the promotion, Lake fought Roland Delorme on June 19, 2010 at . He lost the fight via submission in the first round.\n\nFor his eighth fight in the promotion, Lake faced John Gunderson on November 20, 2010 at UFC 122. He won the fight via split decision.\n\nFor his ninth fight in the promotion, Lake fought Jason Reinhardt on February 5, 2011 at UFC 126. He won the fight via TKO in the first round.\n\nFor his tenth fight in the promotion, Lake faced newcomer Darren Uyenoyama on June 11, 2011 at . He won the fight via unanimous decision (30-27, 30-27, 30-27).\n\nFor his eleventh fight in the promotion, Lake faced former WEC Featherweight Champion Mike Brown on October 29, 2011 at UFC on Versus 6. He lost the fight via unanimous decision (29-28, 29-28, 29-28).\n\nFor his twelfth fight in the promotion, Lake faced Nam Phan on May 15, 2012 at UFC on FX 4. He lost the fight via unanimous decision.\n\nFor his thirteenth fight in the promotion, Lake faced Sam Sicilia on July 11, 2012 at UFC 149. He won the fight via unanimous decision.\n\nFor his fourteenth fight in the promotion, Lake faced Ivan Menjivar on December 15, 2012 at UFC 155. He won the fight via split decision.\n\nFor his fifteenth fight in the promotion, Lake faced Jeremy Larsen on June 15, 2013 at UFC 161. He won the fight via TKO in the second round.\n\nFor his sixteenth fight in the promotion, Lake faced promotional newcomer Justin Salas on November 16, 2013 at UFC 167. He lost the fight via split decision.\n\nFor his seventeenth fight in the promotion, Lake faced fellow newcomer Max Holloway on February 15, 2014 at UFC 169. He lost the fight via TKO in the first round.\n\nAfter his release from the UFC, Lake dropped down to Bantamweight and made his debut on April 11, 2015 at IFC: International Fighting Championship 2 in Des Moines, Iowa against Rafael &quot;Morcego&quot; da Silva. He lost the fight via knockout in the first round.\n\nLake faced Gabriel Checco on May 28, 2015 at CXF 5. He lost the fight via split decision.\n\nLake faced Daniel Weichel on August 14, 2015 at Superior Challenge 12. He lost the fight via submission in the first round.\n\nIn his fourth fight since returning from the UFC, Lake faced Ryan Quinn on October 23, 2015 at CXF 8. He won the fight via knockout in the first round.\n\nLake faced Jim Alers on January 8, 2016 at CXF 10. He won the fight via TKO in the second round.\n\nLake faced Martin Stapleton on February 19, 2016 at MFW: A Night of Champions. He lost the fight via split decision.\n\nLake next faced Dylan Logan on September 2, 2016 at Cage Warriors 82. He won the fight via TKO in the second round.\n\nOn February 28, 2017, it was announced that Lake had signed with the UFC and was expected to face Alessio Di Chirico on March 18, 2017 at UFC Fight Night 107. However, on March 7, Di Chirico pulled out from the event, citing injury, and was replaced by promotional newcomer Marcin Prachnio. Lake won the fight via TKO in the second round.\n\nLake faced Gregor Gillespie on June 25, 2017 at UFC Fight Night 112. He lost the fight via split decision.\n\nLake faced Jarred Brooks on December 9, 2017 at UFC Fight Night 123. He lost the fight via submission in the first round.\n\nLake faced Drakkar Klose on February 24, 2018 at UFC on Fox 28. He lost the fight via TKO in the second round.\n\nLake faced Andre Fili on August 25, 2018 at UFC Fight Night 135. He lost the fight via TKO in the first round.\n\nLake faced Desmond Green on February 17, 2019 at UFC on ESPN 2. He lost the fight via technical knock out.\n\nIn July 2007, Lake was diagnosed with Crohn's disease, an inflammatory bowel disease that causes ulcers in the digestive tract. He began experiencing symptoms shortly after moving to California to train full-time. His Crohn's was brought under control through medication and diet, and it is currently in remission.\n\n\n\n]" time="3.395"><properties><property name="score" value="0.9707017838954926" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Fenix Funk\n\nFenix Funk is the debut studio album by German electronic music group Telepopmusik. It was released on 9 October 2000 by Columbia Records. It was a success in Japan and peaked at #50 on the Oricon Albums Chart.\n]" time="0.182"><properties><property name="score" value="0.9423989653587341" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[If you\u2019re working on a SharePoint app, you\u2019ll eventually need to get hold of a SharePoint context. At the time of this writing, there are two options: SharePointContext and SPUtility. Both work fine, and I think it\u2019s really up to your personal preference and comfort level to choose either. But for my purposes, I\u2019m a fan of using SharePointContext because it will use the most current context available, which is what I like to do.\n\nLet\u2019s take a quick look at the two SharePoint context providers that we have at our disposal: SharePointContext and SPUtility. We\u2019ll use these in conjunction with the new SharePoint REST API to get a SharePoint context.\n\nSharePointContext\n\nThe SharePointContext class represents a SharePoint context. The context is the central part of the SharePoint client object model, and it\u2019s available from any client, including SharePoint-hosted, remote, or SharePoint-hosted-remote applications.\n\nTo create a SharePointContext object, you call the SharePointContext.GetContextAsync() method and pass in a SharePointContextOptions object. The SharePointContextOptions object provides the user information and authentication settings. We\u2019ll dive into the details in the \u201cAuthorization\u201d section of this article.\n\nOnce you have your context, you can then call methods on it. The context has a number of methods on it to retrieve different information.\n\nLet\u2019s take a quick look at the available methods:\n\nGetContextAsync() - The method that is used to create the context.\n\n\n\n- The method that is used to create the context. SharePointContextPermission - The property that contains the permission token used to retrieve the current user.\n\n\n\n- The property that contains the permission token used to retrieve the current user. SharePointPermissions - The property that contains the current permission level of the user.\n\n\n\n- The property that contains the current permission level of the user. ListPermissions - The property that contains the permission level for the user to access lists.\n\n\n\n- The property that contains the permission level for the user to access lists. SPHostUrl - The property that contains the URL of the current host application.\n\n\n\n- The property that contains the URL of the current host application. Site - The property that contains the current site of the user.\n\n\n\n- The property that contains the current site of the user. UserPrincipalName - The property that contains the username of the current user.\n\n\n\n- The property that contains the username of the current user. Web - The property that contains the current web of the user.\n\n\n\nSPUtility\n\nThe SPUtility class provides static methods to retrieve a SharePoint context and return it. To get the context, you call the SPUtility.GetSharePointContext() method.\n\nThe methods you have available on the SPUtility.SharePointContext object are:\n\nClientContext\n\n\n\nWeb\n\n\n\nGetClientContext()\n\n\n\nGetWeb()\n\n\n\nSharePointContext Permission\n\nBefore we dive into the SharePoint context, let\u2019s talk about permission. The SharePoint context is going to be pretty useless without a permission token. We\u2019re going to talk about this permission token a lot in this article, so let\u2019s just take a quick look at how it\u2019s generated.\n\nWe\u2019ll need to use the SharePoint permissions provider to generate the token. To do so, you\u2019ll need to call the SharePointPermissionManager.GetPermissionsAsync() method. You\u2019ll need to pass in an instance of the site where you want the permissions for.\n\nIn this example, we\u2019re going to call GetPermissionsAsync() and get the current user\u2019s permissions.\n\nGetPermissionsAsync() returns a SharePointPermissions object, which you can use to set your permissions on lists, folders, and documents. The SharePoint permissions are applied in a single pass. This means that you can\u2019t apply different permissions to different entities. For example, if you want to apply permissions on the web, the site, and the lists and folders, you\u2019ll need to pass in the permission on the web, then on the site, then on the lists and folders. You can, however, pass in different permissions to different lists and folders.\n\nOne other thing to note about SharePoint permissions is that they only grant access to the web, not to the list or to the folder itself.\n\nFor example, if a user doesn\u2019t have the Edit permission to a list, he will still be able to view it. But if the user has the Edit permission to the list, he will be able to edit it.\n\nIf you want to deny access to the whole list, you can set the permissions for the list and all its subfolders to Deny.\n\nIn order to create the context, we\u2019ll use the context, the SharePoint permissions provider, and a helper class, which we\u2019ll talk about later in the article. The full example is shown in Listing 1.\n\nThe context object will now be passed to the constructor of a class called FileRootsProvider. The FileRootsProvider class is part of a helper class that I wrote and will be available to you in this article.\n\nThe provider will be created with the context and then called to generate the access token. Once the access token is generated, it will be returned to us, and we can start to retrieve information about the site and its lists and folders.\n\nAccess Token\n\nWhen you call the SharePoint permissions provider, you\u2019ll receive a token that represents the permissions for the current user. The token is returned as a SharePointPermissions object, which is what we\u2019ll use to apply the permissions.\n\nTo create a SharePointPermissions object, you need to pass in the site URL where the permissions should apply. For our example, we\u2019ll pass in the site URL to the SharePointPermissions constructor.\n\nAfter you\u2019ve created the permissions, you can apply them to the lists and folders. You can do this by passing in an instance of a list or folder to the SharePointPermissions constructor. If you want to apply permissions to all lists and folders, you can simply pass in an empty list or folder.\n\nNow that we\u2019ve created the permissions, we can use the FileRootsProvider class to generate the access token. The access token is returned as a SharePointContextPermissions object. The SharePointContextPermissions object contains the permission token and other details that we\u2019ll use to get information from the SharePoint site.\n\nAs you can see, the code to get the access token is pretty straightforward. The example is shown in Listing 2.\n\nList\n\nGetting the access token isn\u2019t all you need to do. You also need to be able to get a list and get a list item. To do this, you\u2019ll need to call the SharePointContext.GetListAsync() method. The method is easy to use and returns a list object.\n\nIf you\u2019re looking for a specific list, you can pass in the list GUID or the URL of the list. Listing 3 shows an example of both the URL and GUID to get the lists from.\n\nItems\n\nThe code to get an item from a list is just as easy as getting the list. You can pass in the list item GUID to the GetListItemAsync() method. Listing 4 shows an example of how to get the list item with the GUID.\n\nSummary\n\nThis article was a quick look at how to get a SharePoint context. The examples used a helper class that I created to make it a little easier to get access to lists and items from SharePoint.]" time="4.002"><properties><property name="score" value="0.9997959392411369" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[A few months back, I wrote an article about the journey from yearning to prayer. Since I have been praying on and off for many years, I was finally blessed with the chance to journey from yearning to prayer. It took a few weeks but it did happen. In the past, my experience of praying was somewhat a miserable one. It was not that I was not able to pray but the truth was I never really enjoyed the praying process. As a matter of fact, I used to hate praying. When I was younger, I was more interested in action than in words. I just want to know what I have to do and get it done. Thus, I never really understood praying. I just felt I was doing nothing but looking at the ceiling. It was only after I found my faith in Jesus that I have understood praying.\n\nMy first experience of prayer was on the cross. I experienced the most painful prayer in my life when I saw Jesus on the cross. He was in pain and so am I. He was looking at the ceiling of the cross and so did I. In the midst of my pain, I opened my eyes and saw Him. I realized that my prayers were not on my behalf but for Him. I looked at Him and realized that He was the only one who can pray for me. I looked at Him and asked him to pray for me. I have never felt anything like it. I have never felt more connected to my God than that moment. I have never felt more at peace than that moment. I realized that praying is actually for Jesus. I realized that Jesus came to this world for a purpose. The whole point of his coming was to save mankind. I realized that He came to the]" time="0.881"><properties><property name="score" value="0.9997894763946533" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The Outrageous Princess Costume is a fun, plus size costume for Halloween. The costume is made of plush, purple velvet. The bodice is purple with a black bow at the top. Black ruffles hang from the bodice. A gold foil belt is attached to the skirt with gold studs. The skirt has a black tulle underskirt with a gold and silver trim. Two purple tulle bows are attached at the hem of the skirt. The wig is a fun purple, curly hairstyle with a tiara and gold tiara veil. Get some props and have fun with this plus size costume.\n\nKey Features:]" time="0.307"><properties><property name="score" value="0.9997357726097107" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[See What We\u2019ve Been Up to!\n\nDo you want to know what we\u2019ve been doing with the movies? Well, here\u2019s a quick rundown.\n\nCOLD WAR (with Tim Roth) \u2013 Dan is currently in development on a feature-length version of the short with Tin Star TV.\n\nLIFE SUPPORT (with Julie Walters) \u2013 The full feature has just finished post-production and is looking for a distributor. It has had a private screening at this year\u2019s BFI London Film Festival.\n\nDREAMWORLD (with Douglas Henshall) \u2013 Life and times of Alfred Wallace, who came up with the theory of natural selection, but had his ideas pinched by the infamous Charles Darwin. The film is in pre-production with Electric Pictures.\n\nSHELTER (with Adam Rayner and Bernard Hill) \u2013 UK feature about a homeless man and the estranged daughter who cares for him. Dan is currently in pre-production with Lionsgate.\n\nCHEERIOS (with Harry Lloyd) \u2013 Another one of Dan\u2019s shorts, this is a comedy drama about a family with a dying son. The short has been selected for film festivals around the world and won numerous awards.]" time="0.554"><properties><property name="score" value="0.2003767490386963" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[To find the best restaurants near you, just enter your location into the search field and the restaurants will show up in a few seconds.\n\nWhat to eat in Italy?\n\nThe selection of Italian restaurants near me is the best on a restaurant and cafe app. Select your region or type in a town, village, postcode, a street name or a square and you will see a list of restaurants near me. Filter by price, type of cuisine or what's popular.\n\nItalian cuisine is known for it's fresh ingredients and regional specialties. To get a good feel for Italian cooking, you need to try each region's specialities.\n\nItalian Cuisine and Where to Eat it\n\nItalian food is all about local and fresh ingredients. Expect pasta, meat, cheese, fish, wine, and bread to feature heavily on a typical Italian menu. Eating out in Italy is an experience in itself, with food served on real plates and in restaurants with atmosphere to match. You can get a taste for Italian food in Britain, but it won't match the real thing.\n\nItalian food in Britain\n\nCrispy, fried, or grilled? With so many different types of pizza, the Italian restaurant menu can be hard to navigate. Luckily, we're here to help you out. Get ready to tuck into some of the best food on the planet. What are you waiting for? Order now and enjoy the best pizza!\n\nItalian Food Delivery\n\nDo you want to make the perfect meal but you are not sure what to cook? Don\u2019t worry, in the times that you are not sure of what to cook, the Italian restaurants have the solution. Just order what you want and our restaurant is the solution for you. Now you can order online and everything you need is in the palm of your hand.]" time="0.943"><properties><property name="score" value="0.24193137884140015" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Unleash the Power of Your Language\n\n\u201cLearn How To Speak French Fluently In Just a Few Weeks\u201d\n\nFrench For Beginners: Step-By-Step French (5 CD set) \u2013 Learn to speak and understand everyday French words and phrases!\n\nDiscover The Easy Way To Learn How To Speak French\n\nAchieve Your Language Learning Goals \u2013 And Discover What French Is All About!\n\nSee a sample of this course\n\nTOTAL PAGES: 90\n\nGet Started In French Today\n\nMastering the French language may seem challenging, but with this easy-to-follow course, you will quickly become more confident in the French language. This course is designed to help you achieve your goals in speaking, reading, and understanding the language.\n\nLanguage is a fascinating topic, and knowing how to communicate effectively with other people is a key to success in a globalized society. This course is an effective means for you to learn the language of French, and more specifically, the local French spoken in France.\n\nIf you have never learned a foreign language before, this course can get you started right.\n\nIt will give you a clear insight into the basics of the language, and its essential parts. It is also a complete guide to getting to know the local French.\n\nYou will learn how to pronounce French words with clarity, and have a basic understanding of how to converse in the language.\n\nLearn how to hold a conversation in French with confidence!\n\nThis course has been written for people who would like to start learning the French language, as well as those who want to increase their vocabulary, expand their cultural understanding, or need French for their work.\n\nIf you have been studying French for some time and have gotten to a plateau, then this course will help you go beyond that plateau. The course can also be used as a refresher course to give you more confidence in the language.\n\nThere is a no-risk guarantee with this course. You are protected by a 60 day, money-back guarantee, and are fully covered by our no-risk, no-nonsense guarantee.\n\nSome of the topics covered in this course include:\n\nFacts About The French Language\n\nGreetings In French\n\nIdioms And Phrases In French\n\nA Guide To French Pronunciation\n\nAs well as grammar rules, grammar explanations, and hundreds of vocabulary terms.\n\nHere\u2019s What People Are Saying About This Course\n\nGaby Hernandez \u2013 \u201cGreat course! My instructor is patient and very thorough. Thanks for providing such a great resource!\u201d\n\nChloe Evans \u2013 \u201cThis is an excellent course. I learned so much! I would definitely recommend this course to anyone who wants to start learning French.\u201d\n\nPeter Johnson \u2013 \u201cI bought the course to learn a bit more French and for the advanced grammar. The instructor is excellent. The course is concise and well organized.\u201d\n\nSandra Hernandez \u2013 \u201cI am an advanced French student who has taken 4 years of French in high school and 6 years of French in college. I have a French instructor who is native French. I bought the course to brush up on grammar and to work on my listening skills. I thought the course was really good and the instructor is really good. I really enjoyed the course and feel that it is good for anyone, whether they are beginners or advanced. I would recommend this course to anyone who wants to learn French.\u201d\n\nChristina B \u2013 \u201cI bought this course because I am planning a trip to France, and I wanted to brush up on my French. I was pleasantly surprised. I\u2019ve learned so much. The course covers a lot of material in 5 hours. The instruction is very clear. I am very happy with my purchase.\u201d\n\nWhat People Are Saying About LanguageForBeginners.com\n\n\u201cThe lessons are very well laid out and explained.\u201d\n\n\u201cVery thorough, well presented and written.\u201d\n\n\u201cAn excellent course and great teacher.\u201d\n\n\u201cVery clear and to the point.\u201d\n\n\u201cThis course is perfect for beginners.\u201d\n\n\u201cThe course has helped me a lot.\u201d\n\n\u201cExcellent course for beginners.\u201d\n\nThe audio is very good. The teacher is excellent. The course is very well done.\n\nSee what you\u2019ll learn inside this course:\n\nWhat The French Language Is All About\n\nFrench Grammar In Detail\n\nBasic Sentence Structures\n\nGreetings In French\n\nWhat It Is Like To Speak French\n\nA Few Examples Of Conversations In French\n\nHow To Pronounce Words In French\n\nA Few Simple Phrases In French\n\nA Few French Idioms\n\nHow To Learn The Language Of French\n\nAnd much more!\n\nPlus, with this no-risk guarantee, you can\u2019t lose!\n\nIf you want to learn how to speak French, then you are in the right place. Take the first step towards improving your communication skills, confidence, and experience with a new language.\n\nClick the \u201cBuy Now\u201d button now to get started on your journey!\n\nRelated\n\nLanguage For Beginners: How To Speak French]" time="2.630"><properties><property name="score" value="0.9997014045715332" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[BOOM!! 10 years of Brikk Fast! Wow, time sure flies when you are having fun, making things, and working hard. Brikk Fast has changed a lot over the years, and has become much more than we ever thought it would. The staff at Brikk Fast has grown, we've brought on some awesome employees that have become just as much a part of the Brikk Fast family as we have. In 2010, we opened up a storefront, and that has allowed us to expand on all the wonderful things we do, and the way we do them. This year we've grown a lot as a company, in ways both great and small. As a company we've become much more involved in the local community, with events and fundraisers]" time="0.383"><properties><property name="score" value="0.9996287822723389" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Self-Selection and Pupils\u2019 Labour Market Outcomes\n\nYoung adults from low socioeconomic status backgrounds are significantly less likely to go to university than their better-off peers. There is a significant self-selection in the probability of attending university and of high A-level achievement, which are key drivers of university participation. This paper exploits the introduction of minimum attainment targets in the National Curriculum for schools to explore the impact of early education policies on students\u2019 achievement and on their likelihood to study at university. We find that when minimum attainment targets are introduced, pupils from low socioeconomic status backgrounds are more likely to reach higher grade boundaries in the A-level exams, are more likely to go to university, and are more likely to study science or maths degrees.\n\nContents\n\nIntroduction\n\nData and methods\n\nAnalysis\n\nDiscussion\n\nConclusion\n\nIntroduction\n\nUniversity participation in the UK has grown substantially since the early 1990s. There are now 1.4 million undergraduates, which is almost double the number in 1992. However, participation rates have remained relatively low among those from lower socioeconomic groups (Furlong and Millar, 2008). For example, 40 per cent of young adults from high-income backgrounds were participating in higher education in 2011-12, compared to 21 per cent from the poorest fifth of families.\n\nTo some extent, low participation among young people from less advantaged backgrounds is to be expected given the historical association of university education with privileged groups. Participation rates increased significantly in the 1960s, after state support for university was extended to less affluent groups. Participation rose further during the 1960s, 1970s and 1980s, but then stagnated in the 1990s and declined slightly after 2002. Only in the past decade has participation among the less well-off again risen to the levels recorded in the early 1990s (see Figure 1).\n\nFigure 1. University participation rates by parental income, 1962-2012\n\nNote: Figures are cohort entry rates for those who went to university in each year. Sources: Department for Business, Innovation and Skills (2014); Sutton Trust/Ipsos Mori (2012); Education Policy Institute (2013).\n\nDespite increased participation, the gap in participation rates between young people from more and less advantaged backgrounds is still large, even after controlling for qualifications. On average, young people from the most advantaged fifth of families are about twice as likely to go to university as those from the least advantaged fifth (see Figure 2).\n\nFigure 2. Proportion of young people who go to university by family income, 1996-2012\n\nSource: Participation rates (log scale) are from the UCAS End of Cycle Reports, 1996-2012. The relative income quintiles are from the Family Resources Survey, 2008-2012. The quintiles have been adjusted to allow for comparison across years by using a fixed poverty line.\n\nAt a recent seminar organised by the ESRC Centre for the Microeconomic Analysis of Public Policy, Professor Martin Brown (2012) suggested that it is \u2018a great mystery why this gap is not widening even more\u2019. He also pointed out that there is a significant self-selection in the probability of attending university and of high A-level achievement, which are key drivers of university participation. It is well-known that, in addition to their socio-economic background, non-cognitive skills, such as motivation, are important predictors of academic achievement. Since the increase in university participation has been driven by the growth in participation of less advantaged groups, it is not surprising that it has also been associated with a fall in the average grades of university entrants.\n\nThis paper is motivated by this observation, that early education policies have played an important role in the increase in participation among less advantaged groups. We focus on two such policies: the requirement to meet minimum attainment targets in English, mathematics and science, and the national literacy and numeracy strategies. We explore the impact of these policies on students\u2019 achievement and on their likelihood to study at university. We find that when minimum attainment targets are introduced, pupils from low socioeconomic status backgrounds are more likely to reach higher grade boundaries in the A-level exams, are more likely to go to university, and are more likely to study science or maths degrees. We also find that the increase in the A-level scores and the higher participation rate associated with the introduction of minimum attainment targets cannot be explained by the improvement in the overall level of academic performance in schools or by the lower socioeconomic status of those schools that introduced minimum attainment targets.\n\nData and methods\n\nTo analyse the impact of minimum attainment targets, we use the school-level dataset that the Department for Education makes available annually for the National Pupil Database (NPD). The NPD contains rich information about individual pupils and their schools. We first use the NPD to examine the evolution of achievement at the end of compulsory education by using cohort-level information on final examinations. We then explore the impact of minimum attainment targets using changes in the composition of achievement. The results we present in this paper are robust to the alternative methods of analysis of achievement levels and changes in the composition of achievement.\n\nIn the first stage of our analysis, we explore the evolution of the A-level examination results by using the cohort-level information on final examinations. To study the evolution of achievement at the end of compulsory education, we use the data on A-level grades in GCSE and A-level examinations from 1996 to 2012. In this period, the examinations have changed significantly, from the General Certificate of Secondary Education (GCSE) to the General Certificate of Secondary Education (GCSE) and the A-level examinations. The A-level examination was first introduced in the early 1950s as an academically rigorous alternative to the less selective B-level examinations (Wynne, 2010). Until 1988, A-level grades (i.e. O/A levels) were assigned by teachers, not by the students. From 1988, the responsibility for grading moved to schools. The examination was revised in the early 1990s, after which it was separated into two components: A-level and AS-level. The A-level examinations were re-structured in 2000. Most importantly, the exams have become linear rather than modular. The exams in each subject are now taken at the end of the two-year course. The exams are now externally assessed, and the grade is assigned according to the nationally defined A-level grade boundaries. The percentage of the cohort that sat the A-level examinations increased from about two-thirds in the early 1990s to more than 90 per cent in the mid-2000s. This reflects a policy shift in the assessment of vocational qualifications, which reduced the proportion of vocational qualifications that counted towards a student\u2019s A-level grade.\n\nWe use a slightly modified version of the grading system developed by Hayward, Parry and Stewart (2009) and Bias et al. (2011). We use a scale of 5 grades to assess the performance of students in different subjects: 5 is an \u2018A*\u2019 grade, 4 is an \u2018A\u2019 grade, 3 is a \u2018B\u2019 grade, 2 is a \u2018C\u2019 grade and 1 is a \u2018D\u2019 grade. The best possible grade in each subject is an \u2018A*\u2019. This grading system differs from the one used in most of the existing analyses of A-level results, which use an A-E grading system. The difference between the two grading systems is illustrated in Figure 3, which shows that about 60 per cent of the students in the upper grade boundary of \u2018A*\u2019 under the old system would be below the upper boundary of \u2018A*\u2019 under the new system. Similarly, about 45 per cent of the students in the lower boundary of \u2018C\u2019 would be in the upper boundary of \u2018C\u2019 under the old system, and about 25 per cent of students in the lower boundary of \u2018C\u2019 would be in the upper boundary of \u2018C\u2019 under the new system. We expect that the increase in the A-level scores associated with the change in the grading system will be smaller for subjects that were not directly affected by the reform, such as English, maths and science. We do not analyse students\u2019 performance in the core and non-core subjects, because data on the subject of A-level examination was not available for every pupil.\n\nFigure 3. Grading scales of GCSE and A-level examinations\n\nIn Figure 4, we present the cumulative distribution of A-level scores for each cohort, by the subject studied. Figure 5 shows that the average A-level score increased from a little less than \u2018C\u2019 in 1996 to more than \u2018C\u2019 in 2012, which is consistent with the improvement in the average GCSE results. The gap in the average grades between different subjects remained stable. The increase in the average A-level grades was particularly large in science subjects and in humanities subjects. We do not find evidence of a widening gap in A-level scores between the richest and the poorest students. In the humanities subjects, there is some evidence of an increase in the A-level scores of the less well-off. However, the increase in the average A-level scores was so large that the difference in the average scores between the richest and the poorest students decreased even though it was not eliminated.\n\nFigure 4. Cumulative distribution of A-level scores in different subjects, 1996-2012\n\nSource: National Pupil Database, Department for Education.\n\nFigure 5. Average A-level scores, by subject, 1996-2012\n\nSource: National Pupil Database, Department for Education.\n\nA key issue in the analyses of trends in A-level achievement is the impact of the increase in the proportion of students taking A-level examinations on the average score]" time="5.032"><properties><property name="score" value="0.6157663047313691" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[East Face\n\n\nSkye\n\nThe Complete Guide\n\n\nSummit view from the top of the Cuillin Ridge\n\nTrevor Nace\n\n\nFirst published in Great Britain in 2015 by\n\nThe Crowood Press Ltd\n\nRamsbury, Marlborough\n\nWiltshire SN8 2HR\n\nwww.crowood.com\n\nThis e-book first published in 2015\n\n\xa9 Trevor Nace 2015\n\nAll rights reserved. No part of this publication may be reproduced or transmitted in any form or by any means, electronic or mechanical, including photocopy, recording, or any information storage and retrieval system, without permission in writing from the publishers.\n\nBritish Library Cataloguing-in-Publication Data\n\nA catalogue record for this book is available from the British Library.\n\nISBN 978 1 84797 713 2\n\nFrontispiece: Climber on the summit of the Cuillin Ridge in front of\n\nMuckross Ridge and Beinn Eighe\n\n\nTo my partner Sally, whose support and love made it possible for me to fulfil my dream of a lifetime on Skye.\n\n\nContents\n\n\nTitle Page\n\nDedication\n\nCopyright\n\nAcknowledgements\n\nList of Plates\n\n1 Introduction\n\n2 An Introduction to the Cuillin\n\n3 Access\n\n4 West of Glen Brittle\n\n5 The Coire Gabhail / North-West Ridge\n\n6 The Black Cuillin Ridge\n\n7 The Central Cuillin Ridge\n\n8 The South-East Ridge\n\n9 The Ar\xeate of the Great Tower and the Clach Glas Ar\xeate\n\n10 An Dubh Sgeir\n\n11 The North-East Ridge\n\n12 The Minnoch Burn\n\n13 Selected Bibliography\n\n\nList of Plates\n\n\nPlate 1: Spade Level, Coire an Lochain, Glen Brittle, in winter\n\nPlate 2: View of the Beinn Eighe massif from the east\n\nPlate 3: Snow-filled Coire na Banachdich, Coire an Lochain\n\nPlate 4: Black Cuillin ridge from the Spade Level\n\nPlate 5: Near the end of the Black Cuillin ridge\n\nPlate 6: Climbers on the summit of Tower Ridge\n\nPlate 7: Skye seen from the summit of the Cuillin ridge\n\nPlate 8: Climbers near the summit of Clach Glas Ar\xeate\n\nPlate 9: Snow on the summit of An Dubh Sgeir\n\nPlate 10: Landscape of the Minnoch Burn\n\n\nAcknowledgements\n\n\nFirst of all, I would like to thank John Brailsford, who accompanied me on my very first walk in the Cuillin in 1996 and whose generous and knowledgeable company has been a constant source of inspiration to me. A particular thanks also to Donnie Hall for all his practical support and guidance during my exploration of the Cuillin. I would also like to thank the helpful staff of the Skye Library in Kyle of Lochalsh, and Alison Connell for the excellent maps. I am grateful to the Climbers\u2019 Club for publishing my article on the Cuillin.\n\nFinally, a big thank you to Sally, who has been a constant source of encouragement, inspiration and love, and to my family and friends who have given me endless support.\n\n\nPlate 1: Spade Level, Coire an Lochain, Glen Brittle, in winter\n\n\nPlate 2: View of the Beinn Eighe massif from the east\n\n\nPlate 3: Snow-filled Coire na Banachdich, Coire an Lochain\n\n\nPlate 4: Black Cuillin ridge from the Spade Level\n\n\nPlate 5: Near the end of the Black Cuillin ridge\n\n\nPlate 6: Climbers on the summit of Tower Ridge\n\n\nPlate 7: Skye seen from the summit of the Cuillin ridge\n\n\nPlate 8: Climbers near the summit of Clach Glas Ar\xeate\n\n\nPlate 9: Snow on the summit of An Dubh Sgeir\n\n\nPlate 10: Landscape of the Minnoch Burn\n\n\n1\n\n\nIntroduction\n\n\nThe spectacular granite peaks of the Cuillin on the Isle of Skye have captured the imagination of hillwalkers and climbers for over a hundred years. When George Henry Alcock (1864\u20131941), a climber and mountaineer, visited the Cuillin in 1893, he wrote in his book Through the Highlands and Islands of Scotland:\n\nNo one who had not seen the Black Cuillin and the peculiar forms which the rocks assume would believe that nature could assume such strange shapes as are seen here. The first view of the Cuillin from the south makes them look like a huge fortress of rock, with fantastic turrets and chimneys, and battlemented walls.\n\n\nToday these classic features, viewed from the slopes of the nearby mountains of the mainland, such as the Beinn Eighe and M\xf2ine Mh\xf2r, have hardly changed. But the Cuillin are not just aesthetically pleasing. To their first British explorers in the mid-nineteenth century, the Black Cuillin, which form a massive ridge running from Coire an Lochain in Glen Brittle to An Dubh Sgeir, seemed a forbidding mountain barrier, \u2018a fortress of rock\u2019. Their modern reputation, and that of the smaller range of the Red Cuillin to the south-east, as a major challenge to both rock climbers and mountaineers, was established during the second half of the twentieth century. However, despite the presence of the two main ridges on the Black Cuillin (North-West Ridge and Central Ridge), many individual peaks, and the climbing routes up them, were relatively unexplored. This was to change in the 1970s and 1980s when many of the hard, mixed and winter routes, most of which were first climbed in the Cuillin, were recorded by enthusiastic pioneers. Today, climbing and scrambling in the Cuillin can be combined with hillwalking, and the varied routes, combined with the varied weather of Skye, offer an extremely satisfying and testing experience for a mountaineer.\n\nIn writing this guidebook, I have had to make many decisions regarding what routes to describe, and which to omit. Many climbers may not agree with my choices, and it is inevitable that a guidebook such as this will not be comprehensive. As well as describing my chosen routes, I have tried to include a range of other interesting aspects of the Cuillin, in order to make the book relevant to anyone interested in a mountaineering trip to Skye.\n\nGeology\n\nThe geology of Skye, and especially the Cuillin, is fascinating, and as well as a number of peaks and ridges, it has given rise to a series of dramatic glens, the upper parts of which are in the Inaccessible Pinnacle (Coire nam Bian) to the north of the main ridge. The Black Cuillin consist of metamorphic rocks which were formed as a result of a major mountain-building event in the Caledonian Orogeny, the mountain-building event which took place between 500 and 400 million years ago. At this time, a number of folds (anticlines and synclines) were formed, some of which can be seen today. The Blue Stack of the Old Man of Storr, for example, was the result of the raising up of a syncline (Fig. 1). This event also resulted in the granite being exposed to the surface, but this did not happen everywhere. To the east of the island, most of Skye is underlain by metamorphic rocks, but to the west, particularly on the western coast of the Trotternish peninsula, granite is exposed. The Blue Stack is also part of a major fold which has resulted in the formation of the Trotternish Ridge.\n\nFigure 1. The Blue Stack of the Old Man of Storr.\n\n\nThese major events have also resulted in a relatively complex geological map, with three main rock types (Fig. 2). A general layer of granite, called the Inaccessible Pinnacle Quartz Diorite, extends from the Blue Stack north-eastwards to Beinn Eighe. This granite has been intruded by the Red Cuillin Granite, which has a different composition and thus has different characteristics to the underlying rock. In the central and eastern parts of the main ridge of the Black Cuillin, the rocks are composed of the Papcastle Gneiss, which is a granitic metamorphic rock. However, to the west, around Sgurr nan Gillean and Glas Bheinn, the Papcastle Gneiss has been intruded by the Kintail Granite, giving rise to a more mottled appearance in this part of the Cuillin. A further complex of rocks is exposed on the north side of Glen Brittle, which includes the Torridonian Sandstone, Lewisian Gneiss, and Dalradian schists.\n\nFigure 2. Geology of Skye showing the main rock types.\n\n\nPeaks\n\nAs well as providing interest to geologists, the geomorphology of the Cuillin is also fascinating for climbers and walkers. With so many ridges and peaks, there is much to explore, and each summit offers its own rewards. Some of the individual peaks are not very high, but offer serious scrambles and rock climbs, while others, such as the highest summit of Sgurr nan Gillean (1,018 m) and the first major peak on the Central Cuillin Ridge, Sgurr Dearg (974 m), offer impressive views. Sgurr nan Gillean]" time="4.845"><properties><property name="score" value="0.2270495370030403" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.22704954&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.22704954
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The Orangeville Banner\n\nOrangeville, Ontario, Canada\n\nApril 9, 1905\n\nPAPER HUNT CONTINUES\n\nINSTEAD OF CASHING $5.00 BILL\n\nA message has been received at the Herald office from Orangeville to the effect that a lady, who has had $5.00 stolen from her purse, did not give the money to an expressman, as stated in last issue, but to another man whom she has not been able to identify, and that the messenger from St. Catherine will not get the money unless he turns it over to her. We have been searching for this lady for several days, and when the Herald office was last searched, we overlooked this message.\n\nAnother story of the stolen money says the woman, whose name is not given, gave the money to a young man who said he was a messenger from St. Catherine, and promised to deliver the money to the party it was addressed to at Kincardine.\n\nTHE MAYOR DISQUALIFIED\n\nA meeting of the citizens of Orangeville was held last Saturday evening at the Orangeville Town Hall, with Dr. A. S. Gibson, M.P., in the chair, to take into consideration the statement made by W. A. MacLean that Mayor D. M. McKenzie was disqualified from holding office under the Municipal Act of Ontario. It was resolved that the matter be placed before the proper authorities for them to decide. The mayor then withdrew from the hall.\n\nMayor McKenzie\u2019s statement in relation to his withdrawal from office is as follows:\n\n\u201cIn this connection I have no desire to act in any unbecoming manner, and I shall resign from the position of mayor. However, I do not desire to leave town. I will reside here and attend to my business as usual, and my many friends who know me and who are not residents of the municipality will be glad to meet me. I want to get at the bottom of this business and find out who it was that drew up the notice of my being disqualified. I was the first one to speak in this town against the big sewer, but I am satisfied now that my opinions are shared by the majority. I do not think that the matter will go much further than the city hall, and as soon as the legal steps are taken the city will be rid of this question.\u201d\n\nSILVER ORGANIZATION\n\nWe are glad to learn that a Silver organization has been formed in this town, and that a meeting is to be held at the Orangeville Opera House next Saturday evening. This is an organization for the purpose of working for free silver, and they are working under the leadership of W. H. Phillips, M.P.\n\nThe representatives of the Anti-poverty society who are here this week will probably have an engagement to-morrow evening. They are endeavoring to make Orangeville the headquarters of the work in this vicinity, and several of the agents will be here for some time. The meeting will be held at the town hall.\n\nA prominent Orangeville man, who is in the wholesale hardware trade in this city, returned this week from Chicago.]" time="1.752"><properties><property name="score" value="0.27541611591974896" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[&quot;Outstanding! Don't miss this book. No matter what is going on in your life right now, this book will speak to you and help you move beyond whatever has you stuck or hurting. Buy it, read it, and be better for it.&quot;\n\n&quot;I am so very grateful that Kari Byron chose to share her story with the world. As a writer, I love to read memoirs, but this one was different than others I have read. The author shared her story in such a way that it was as if I was hearing it from a dear friend. In this era of time, it is refreshing to read something from someone who has the courage to share the good and the bad from their lives. I laughed, I cried, and I grew to appreciate the value of living. I hope this book will find its way into the hands of many who are struggling with life in a way that brings freedom to their hearts.&quot;\n\n&quot;Kari Byron's book is a refreshing look into a woman's journey to finding joy after hardship. She offers hope and practical tips that can be used by women of all ages. This book is a must read.&quot;\n\n&quot;I couldn't put it down! I'm a mom of 3 young kids and struggle with guilt daily, but Kari Byron's words of encouragement and her willingness to share her personal struggles make me feel like I can face my own obstacles and not feel alone.&quot;\n\n&quot;Kari Byron shares her heart in this honest memoir and gives you the tools to follow your heart, your passion and live your life to the fullest! You are sure to find her journey inspiring!&quot;\n\n&quot;I loved this book. From the moment I picked it up, I couldn't put it down. It was a fun, easy read, but the lessons and life advice Kari shared had me reflecting and taking notes. I laughed and cried, and I'll be reading this book over and over again to encourage myself and remind myself of the lessons it teaches.&quot;]" time="1.124"><properties><property name="score" value="0.8066774904727936" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Since the beginning of time, humans have been fascinated by the mysterious. The sky has always been full of questions. Our ability to dream big and speculate has enabled us to discover, theorize, and theorize even more. But where do we get the answers? These are the top 10 theories on the unsolved mysteries of the universe.\n\n10. Extraterrestrial Origin\n\nThe human race began to emerge from the Stone Age. The caves and forests were just getting familiar with the new inhabitants. There was a long time of fear and ignorance. Before we started to understand the world, they began to imagine what is outside the world. Some thought of creatures so powerful that they could destroy the world. Other speculated that we are a higher level of evolution and are the result of genetic experimentation of extraterrestrial beings.\n\nLater, these speculations were transformed into mythology and the imagination of the humans. In our modern world, it has turned into science. Many scientists still believe that we are not the only life in the universe and that at some point in history there was a connection with extraterrestrials. Some of them even believe that the first stages of our life were carried out with the help of aliens.\n\n9. Time Travel\n\nSince the beginning of history, people have always been fascinated with time. To travel back in time would allow us to explore the unknown, but would it be worth it? We live in a world where time is constant. Change is one of the most important principles of human life.\n\nNevertheless, many scientists still consider the possibility of time travel. The hypothetical method would be to build a giant machine that would make a hole in time. But how can this happen? It would involve breaking the speed of light and space.\n\nThe machine would have to be so big that it would create a massive black hole. Then the time machine would be moved into the center of the black hole, which would allow it to break the laws of time. So far, scientists have not been able to prove the validity of this theory.\n\n8. Time Travelers\n\nDespite the fact that time travel is still a hypothesis, there are people who believe that it is not. They claim to have traveled back in time to the present or to a previous generation. For this reason, there are also numerous stories about people who claim to have returned from the future to save the world.\n\nMost people who claim to have traveled back in time are just crazy. However, many scientists consider it possible that in the future, when we understand the laws of physics better, we may be able to travel through time. This possibility has been a mystery for a long time. So, this is a list of top 10 theories on the unsolved mysteries of the universe.\n\n7. Deja Vu\n\nThe theory of Deja Vu is so controversial that many scientists do not even dare to think about it. The definition of deja vu is, when a person experiences something, although he has never done it before. Many people think of it as a kind of visual hallucination. The most popular theory about the phenomenon is that it is a past memory that has not been triggered by the usual way.\n\nAnother possible explanation is that the brain creates a fake memory to deceive the person who feels it. The theory is that in reality the person does not experience a deja vu, but it is an internal response to the same scenario. In short, the brain recognizes the environment and recognizes the story that it represents.\n\nIn this case, the illusion is that the person has lived through this scenario. Nevertheless, there are many people who do not think so. The reason is that deja vu is a phenomenon that occurs in any person. So, we have 10 theories on the unsolved mysteries of the universe.\n\n6. Water World\n\nIt is one of the most popular theories in the world of science. It is believed that the moon is a result of a collision between the Earth and a body much larger than it. As a result of the collision, our planet was literally ripped. But what happened to the mass that fell on the surface? According to scientists, this part of the Earth remained liquid.\n\nThe evidence is so compelling that the theory of the water world has become a reality. For example, there is a large number of scientists who believe that the moon was the result of the Earth\u2019s rotation and gravitational collapse.\n\nMoreover, there are evidences that this theory may be true. For example, when scientists tried to drill in the Antarctic, they found water underneath. Also, according to many scientists, the planet Mercury, which has no magnetic field, may be the result of the remains of the water world.\n\n5]" time="2.430"><properties><property name="score" value="0.9997625708580017" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[There are no spoilers in this review\n\nFrom the off, the third episode in the new Star Trek Discovery series has a slow pace. It is the most quiet of the three episodes and is more of a character study in the struggles faced by all of the Discovery crew and how that manifests in different ways. With that in mind, I find it strange that this is not getting the universal praise that the first two episodes were. But that\u2019s the power of spoilers I suppose, no one knew what was coming in the first two episodes.\n\nThe episode itself, despite being more quiet than the first two, still manages to deliver plenty of story. You do have to take it with a grain of salt because some of the \u2018bad guy\u2019 stuff is almost unbelievably naive but that\u2019s the Star Trek universe for you. Despite the low stakes, I was still hooked on every word and scene.\n\nThere are plenty of hints that we can be certain that there is a big bad out there that we have not seen yet. Perhaps that is what made the episode more interesting. There is something looming over them and we don\u2019t know what it is yet.\n\nFor some reason, people seem to have a problem with the fact that the crew spend time getting to know one another, with only a couple of scenes of violence and action. Personally, I found that the social dynamics between the characters to be very interesting and engaging. Yes, this is probably true to life and does have some real life parallels. I am glad they did not waste time with all of that mindless action that I\u2019ve become so accustomed to.\n\nThere are no huge shocks in this episode and the pay offs to the mini story arcs were a bit weak. But when the pay off came, I was still glad that they took the time to introduce us to the characters in this way.\n\nIf you have been put off watching Star Trek Discovery because of the fear of spoilers, then I would suggest you don\u2019t worry too much and go ahead and start the show. If you can manage to resist the temptation of the spoilers for the next two weeks, then I think you will be very happy with the third episode and the direction the show is heading in.\n\nLet us know what you thought of the episode in the comments below and if you have seen the first two episodes.]" time="1.200"><properties><property name="score" value="0.9963693420092264" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Price includes 15% HST, 10% goes to the developer (Living Balance)\n\nWaterfront construction, 11 homes, classic tudor style, quality &amp; contemporary style homes to suit all tastes. Located in Waterfront Developments development at Schooner Bay, golf and public beaches nearby. Great potential with 3 floor plans offered: MLS # V3643336 Single Family Detached, MLS # V3643329 Single Family Detached, MLS # V3643330 Single Family Detached.\n\nThe trademarks MLS\xae, Multiple Listing Service\xae and the associated logos are owned by The Canadian Real Estate Association (CREA) and identify the quality of services provided by real estate professionals who are members of CREA. Used under license. The trademarks REALTOR\xae, REALTORS\xae, and the REALTOR\xae logo are controlled by The Canadian Real Estate Association (CREA) and identify real estate professionals who are members of CREA. Licensed REALTOR\xae. Used under license. Data last updated: 2019-09-21T20:58:44.137.]" time="0.535"><properties><property name="score" value="0.5617370009422302" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[You might know of me. I am one of the project managers for Windows 7. Or Windows 7 team is one of the teams that I work in. As a lot of you know, my favorite and almost only thing to do when I am not working is to read a good book. Well, I have done it again. I just finished the book I am about to recommend to you. This book is called \u201cReady Player One\u201d by Ernest Cline. I can\u2019t recommend this book enough. So what\u2019s the big deal? You are about to find out.\n\nSo what is it about?\n\nThis book is a story about Wade, a teenager, living in the year 2045. The world is in a bad state.\n\nTechnology has progressed very fast, but a lot of humans have been left behind. Wade\u2019s life sucks. He lives with his aunt in the slums of Oklahoma. All he does is play video games.\n\nThen, one day, he hears of a contest to win a fortune by finding the first Easter egg in the virtual reality game OASIS. The creator of OASIS has died and has left behind a secret quest. A quest that has the prize to the first person to complete the quest. But there is a catch: only the creator of the game knows how to find the Easter egg. And he has hidden the clues in the game.\n\nIt is up to Wade to try to find the Easter egg first. But not alone. He quickly assembles a group of fellow OASIS players. This group consists of his best friends and fellow gamers. This quest will soon turn into a journey. And I don\u2019t want to tell you more, because I don\u2019t want to spoil the story for you.\n\nA bit about the Author:\n\nErnest Cline is the author of the book Ready Player One. He is from the state of Texas. He has lived in New York City for a few years, but has now moved back to Texas. He is 40 years old and has been writing books for a long time. His first book was published in 2001.\n\nErnest Cline is very much into old science fiction books, and especially the 1980\u2019s. He loves the music, TV shows, video games and movies of that decade.\n\nMy thoughts:\n\nWhen I read this book, I really couldn\u2019t put it down. I just wanted to know how the story would end. I wanted to know if the good guys would win. I wanted to know how they would beat the bad guys.\n\nI am sure you will love this book, just like I did. It is filled with great references to the 1980\u2019s. If you are born after that time, you will probably have to read the story first to get the references, but I am sure you will enjoy it.\n\nI will also tell you, that this book is a great story, and I can easily recommend this book to any reader. It is not just a story about technology. It is also a story about friendship and courage. And in the end it is a story about hope and dream.\n\nIf you want to read this book, you can find it on Amazon. Or your favorite bookstore. I am not going to tell you how much it costs. I am just going to tell you that it is well worth the price.]" time="1.714"><properties><property name="score" value="0.9998165567715963" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Additional information\n\nApplicants must be 18 years of age or older and must complete and submit the National Labor Relations Board (NLRB) #6 or #7 form, indicating that they will not engage in secondary boycotts or strikes.\n\nThis position is in the bargaining unit.\n\nWe will require your W-2 forms as well as your Federal tax returns for the previous year for verification of your claimed dependents.\n\nApplicants must have unrestricted United States Government security clearance with the Department of Defense or its contractors to be eligible to work in a building that is controlled access.\n\nThis position is located in a National Defense Center. Individuals must be U.S. Citizens or U.S. Nationals (no exceptions) or eligible to be a U.S. Citizen or U.S. National to be hired in this position.\n\nApplicants who have not yet completed a Security Clearance must be able to obtain one in a reasonable amount of time.\n\nApplicants must have or be able to obtain a high school diploma, GED, or other equivalent qualification, and two years of job related experience to be hired.]" time="0.618"><properties><property name="score" value="0.8322146534919739" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Exterior painting\n\nBrian's painting was done in a timely manner, kept the site clean and did a nice job with all the details. We will definitely use him again. I would recommend him to all my family and friends.]" time="0.188"><properties><property name="score" value="0.9933651089668274" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Corporate financial managers are responsible for financial reporting, financial planning and analysis, cash management and budgeting, debt management, financial analysis, and financial system development.\n\nWe can help you on the following course:\n\nCorporate Financial Management (DFM)\n\nFurther information about Corporate Financial Management (DFM)]" time="0.210"><properties><property name="score" value="0.8227195143699646" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[True, she is not the most charismatic of the four ladies, but her choices are very strategic and her lines usually have the ring of truth. If there were ever a reason to cast someone as a drag queen, it would be La Chatelaine. Her mannerisms, her expressions, and the way she carries herself scream Drag Queen. My best guess is that she has a very deep understanding of the difference between a woman and a drag queen and she knew how to use that to her advantage.\n\nHollywood]" time="0.251"><properties><property name="score" value="0.9994833469390869" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[This is the terrifying moment a deadly great white shark was caught on camera devouring a huge tuna fish near a group of stunned fishermen.\n\nThe fisherman from Esperance, Australia, couldn't believe their eyes when they spotted the enormous shark leap out of the water with a tuna fish in its mouth, as they were fishing for yellowfin tuna at Blue Hole, off the coast of Esperance.\n\nFisherman from Esperance, Australia, spotted a great white shark jumping out of the water with a tuna fish in its mouth\n\nThe footage was captured by amateur fisher Domenic Porcelli and posted to Facebook by fishing website Shark Alerts South Australia and South West Australia.\n\n'Watch in amazement as this huge great white shark leaps out of the water with a massive tuna fish in its mouth,' the]" time="0.400"><properties><property name="score" value="0.9997475743293762" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[HARRY POTTER and the Order of the Phoenix is a 2007 British-American fantasy adventure film directed by David Yates and distributed by Warner Bros. Pictures. It is based on the novel of the same name by J. K. Rowling. The film, which is the fifth instalment in the Harry Potter film series, was written by Michael Goldenberg and produced by David Heyman and David Barron.\n\nThe film stars Daniel Radcliffe as Harry Potter, alongside Rupert Grint and Emma Watson as Harry\u2019s best friends Ron Weasley and Hermione Granger.\n\nThe story follows Harry\u2019s fifth year at Hogwarts as the Ministry of Magic is in denial of Lord Voldemort\u2019s return. Harry must uncover the truth, with the help of his friends, and confront Voldemort.\n\nThe film began production in early 2007 and was released in theatres in the United Kingdom and United States on 12 July 2007, grossing a total of $292 million at the worldwide box office.\n\nThe film was nominated for two BAFTA Film Awards in 2008.\n\nFollowing a Harry Potter fan\u2019s dream that Harry\u2019s late headmaster Albus Dumbledore is alive, and in a critical condition at the Ministry of Magic, Harry Potter and his friends Ron Weasley and Hermione Granger, decide to rescue him, as the school year comes to a close.\n\nOn the night of their attempt to break into the Ministry, Ministry of Magic employee Delores Umbridge slashes Rubeus Hagrid\u2019s hand with a knife, accusing him of stealing her kitten.\n\nHarry, Ron, and Hermione fight the Death Eaters. Lucius Malfoy and his son Draco both try to attack Harry, but he manages to save himself. Harry, Ron and Hermione get separated, and in the chaos, Ron manages to grab the prophecy from the Ministry\u2019s Hall of Prophecies.\n\nHarry and Hermione escape the Ministry, and Ron awakes to see Harry and Hermione\u2019s signatures on the prophecy, but does not know that Harry had ripped the prophecy out and left it in the hands of his godfather, Sirius Black, before he was killed.\n\nSirius, Harry, and Hermione head to the headquarters of the Order of the Phoenix, where they meet with fellow members Remus Lupin and Nymphadora Tonks, and Sirius\u2019s brother, Regulus.\n\nThe Order is joined by the regular adult wizards and witches in the battle, as Voldemort and his Death Eaters take over the Ministry of Magic and declare martial law on the entire Wizarding community.\n\nAlbus Dumbledore appears to die in battle, but this is revealed to be a ruse, as he and Severus Snape attack Voldemort and Lucius Malfoy, and attempt to take the prophecy from Ron. Lucius disarms Dumbledore, and an enraged Bellatrix kills him.\n\nBellatrix and Narcissa Malfoy stand by their husbands\u2019 sides. Albus Dumbledore\u2019s portrait is replaced with one of Snape. Harry, Ron, and Hermione are rescued by Order members. Voldemort and the Death Eaters leave the Ministry.\n\nHarry, Ron, and Hermione are told by Remus Lupin that all of the other Order members have been killed. Harry attempts to repair his damaged wand. He is visited by Snape, who tells Harry that the boy\u2019s wand is irreparably damaged and needs to be destroyed, as Voldemort has become aware that the wand is of a \u201cunique\u201d nature.\n\nThe two engage in a fierce duel in which Snape calls on his master to save him. Harry is unaffected by the curse due to his ability to cast a shield charm. He manages to shield himself and fight back, and in his distraction, Snape accidentally breaks his neck and dies.\n\nHarry meets with Dumbledore\u2019s portrait, who reveals to Harry that the boy\u2019s mother died to save him, and Harry is filled with his mother\u2019s love. Harry reveals that he feels angry and confused at this revelation.\n\nHarry also learns that the reason that Voldemort has gone to such great lengths to kill him is that Harry is a Horcrux, a piece of Voldemort\u2019s soul that resides in Harry\u2019s body.\n\nHarry then leaves Dumbledore\u2019s portrait and meets with Ron and Hermione, who have just been told that the students and staff are to evacuate the school, as the Death Eaters have declared that Hogwarts is no longer safe.\n\nHarry, Ron, and Hermione are joined by Remus Lupin, Nymphadora Tonks, and several other members of the Order of the Phoenix, and they apparate to 12 Grimmauld Place, the headquarters of the Order.\n\nThe Battle of the Department of Mysteries is followed by the Death Eaters taking over the Ministry of Magic, as they believe the prophecy to be a key to Voldemort\u2019s victory. Harry returns to Hogwarts to retrieve the prophecy from Voldemort.\n\nHe, Ron, and Hermione get ambushed by Death Eaters, and when Harry goes after the prophecies, he ends up dueling with Voldemort.\n\nHarry calls upon the remaining strength of his mother\u2019s love, and as Voldemort\u2019s Killing Curse rebounds upon him and hits Harry, the battle is ended, and the school is cleared of the Death Eaters.\n\nHarry wakes up to find Dumbledore, Sirius, and Remus in the hospital wing, as well as his friends and the rest of the school, and he realizes that he is safe.\n\nDumbledore then proceeds to give Harry his restored invisibility cloak, explaining to Harry that the Invisibility Cloak is one of the Deathly Hallows. Harry then says goodbye to his friends, as he leaves with Dumbledore to destroy the seven horcruxes that contain fragments of Voldemort\u2019s soul, which will make him immortal.\n\nThe film ends with Harry and Dumbledore leaving the school in a car, accompanied by Professor Minerva McGonagall.\n\nI liked this movie as it is full of action and adventure. The plot is great as well as the dialogue. It is a well made movie and it is very entertaining.\n\nThis movie is definitely a must-watch, as it has plenty of action as well as being very humorous.\n\nAdvertisements]" time="3.306"><properties><property name="score" value="0.9997747043768564" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Hush, little baby, don't you cry\n\nI know a thing or two that will make you happy,\n\nIf you will only come with me.\n\n\n\nHush, little baby, don't you cry\n\nI know a thing or two that will make you happy,\n\nIf you will only come with me.\n\n\n\nIn my bag there's a beautiful surprise,\n\nIf you will only come with me.\n\n\n\nOn my back I'll show you the sights\n\nThere's a rainbow in your eyes,\n\nThere's a sweet, sweet lullaby\n\nIn your mother's smile.\n\n\n\nOn my back I'll show you the sights\n\nThere's a rainbow in your eyes,\n\nThere's a sweet, sweet lullaby\n\nIn your mother's smile.\n\n\n\nHush, little baby, don't you cry\n\nI know a thing or two that will make you happy,\n\nIf you will only come with me.\n\n\n\nIn my bag there's a beautiful surprise,\n\nIf you will only come with me.\n\n\n\nOn my back I'll show you the sights\n\nThere's a rainbow in your eyes,\n\nThere's a sweet, sweet lullaby\n\nIn your mother's smile.\n\n\n\nIn my bag there's a beautiful surprise,\n\nIf you will only come with me.\n\n\n\nIf you will only come with me.\n\n\n\nOn my back I'll show you the sights\n\nThere's a rainbow in your eyes,\n\nThere's a sweet, sweet lullaby\n\nIn your mother's smile.\n\n\n\nIf you will only come with me.]" time="0.793"><properties><property name="score" value="0.8184911608695984" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Mileage : 19,038 (77,223)\n\n: 19,038 (77,223) Transmission : 4-Speed Automatic\n\n: 4-Speed Automatic Exterior Color : Green\n\n: Green Interior Color : Light Blue\n\n: Light Blue Engine : 351\n\n: 351 VIN : #B9QWL214742\n\n: #B9QWL214742 Stock Number : 96988\n\n: 96988 Fuel: Gasoline\n\n1954 Ford Crown Victoria\n\nThis Ford is a very nice car, we are selling it at a fraction of the cost of building it! The motor is a very desirable 351 V8 and is equipped with a nice sounding Holley 650 carburetor. It runs well and has plenty of power. The trans is an automatic C4 and shifts smooth. This Ford is in great shape, the only exterior rust that I could see was on the door bottoms and they are easy to replace. Interior is in very nice shape, new headliner, new seats, new carpet and a nice newer steering wheel. It has power brakes, power steering, air conditioning and power top. These cars are extremely rare to find in this condition and price. If your looking for a car that is cool and nice to drive then this is it.]" time="0.634"><properties><property name="score" value="0.9882292747497559" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[I had a miserable time, once, when I was hospitalized. In the state of euphoria that was my hospital experience, the only one I could really think of to complain about was that I missed my window of opportunity to go on that uncharacteristic mission to munch fresh snow.\n\n\u201cYou mean\u2026EVERYONE doesn\u2019t get a chance to go on a snow mission,\u201d my mother replied. \u201cYou have to have the proper mindset for a snow mission. You have to be excited about it and you have to want to do it. If you]" time="0.319"><properties><property name="score" value="0.8289560079574585" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Create our team or join our team.\n\nIf you are an organization, brand, or personality and you want to take your show to the next level then get in touch! Email us at team@geekarmy.com and we will get right back to you.\n\nCreate a new Team\n\nSimply sign in with your username and password. If you do not have an account you can create one here.\n\nAfter signing in you will be able to create a new Team. You will be able to add a short name, a full name, a URL and a description. You will also be able to upload a logo for your team and a picture for the team members.\n\nEdit an existing Team\n\nSimply sign in with your username and password. If you do not have an account you can create one here.\n\nSelect the team you wish to edit from the My Teams page. From there you will be able to edit the team name, upload a new logo, a picture for the team members, and a URL and description. You will also be able to add members to your team.\n\nSearch for a team\n\nTo find a team click on the Teams Tab and you will be able to search by a number of different criteria including team name, brand name, member count, and popularity.\n\nAfter you find a team you are interested in, simply click the button at the top of the page to see more information about the team.\n\nJoin a team\n\nOnce you find a team you are interested in, simply click on the Join button and you will be taken to the join page. Here you will be able to apply to join the team.\n\nAfter you apply you will be sent a confirmation email. The email will contain a link for you to confirm your join request.\n\nAll of our applications are manually reviewed and you will be sent an email to notify you whether you have been approved or not.]" time="0.966"><properties><property name="score" value="0.9998253583908081" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[&quot;While some species are recognized as a single unit, they often consist of several distinct groups that have a common characteristic,&quot; Davis said. &quot;In this case, the characteristic is the presence of this particular steroid in all the species we looked at. If this steroid is a unique identifier for the group, we wanted to know if the species that contain it are closely related. We discovered that some species that produce the steroid are more closely related than others, and this pattern fits with our prediction that the ecological function of the steroid might be to keep different species from competing with one another for food.&quot;]" time="0.396"><properties><property name="score" value="0.9997981190681458" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[SEOUL, Oct. 30 (Korea Bizwire) \u2014 A shopping mall in southern Seoul opened on Wednesday with facial recognition technology installed in the washrooms for toilet paper dispensers.\n\nShoppers at the VOGUE shopping mall on Gwanghwamun in Seoul can use facial recognition to automatically dispense toilet paper, according to the mall\u2019s management company, which boasts that it\u2019s the first time that facial recognition technology has been applied to a toilet paper dispenser.\n\nThe facial recognition software automatically senses the number of sheets dispensed. It also allows the user to set an alarm to sound after a preset number of sheets has been dispensed.\n\nThe software is set to go on trial until November, when the management company will decide whether to expand its application to other toilet paper dispensers.\n\nFacial recognition technology is already being used in other places in Seoul, such as apartment complexes and an upscale golf course.\n\nBy Joseph Shin (jss539@koreabizwire.com)]" time="0.504"><properties><property name="score" value="0.9136998653411865" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.91369987&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.91369987
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[\n\n15% OFF ANY SINGLE ON EBAY WITH COUPON CODE &quot;PLAYTIME15&quot;\n\n(For US and Canada users only. Offer expires December 13th at 11:59 PM PT)\n\nError: You must have Javascript enabled in order to use this website.\n\nSkyscraper Property Text This card cannot be Special Summoned. Your opponent cannot select a Level 5 or higher monster on their side of the field as an attack target.\n\nTournament Eligibility TCG Advanced TCG Traditional OCG Unlimited Unlimited Unlimited\n\nPurchase Statistics (Last 30 Days) See More Number Purchased Money Spent 0 $0.00\n\nTournament Stats | Yugioh Wiki | Expand/Collapse All\n\nChecking Card Prices Please Wait\n\n(this might take a minute)]" time="0.389"><properties><property name="score" value="0.7567939162254333" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.75679392&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.75679392
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[John Mayer Just Made A Fool Of Himself In An Interview With The New York Times\n\nBy James Wilson-Taylor\n\n&quot;We love to focus on my flaws and make them bigger than they are.&quot;\n\nJohn Mayer has come out fighting in a new interview with the New York Times, discussing a range of topics including the ghostwriting controversy, his upcoming tour with the Dead &amp; Company and the end of the line for the John Mayer Trio.\n\nThe article details a frank conversation with the singer, in which he shares his honest thoughts on everything from his troubled past, to his use of social media, to the over-analytical nature of modern day music criticism.\n\nThe interview was conducted as a two-day workshop with NYU's music journalism department, which of course meant that the whole thing was promptly picked over by the internet on Tuesday night.\n\nHere's a selection of the choicest cuts:\n\nHis Vocal Fry Controversy:\n\n&quot;The most controversial thing I\u2019ve done recently is affect a slightly looser, lower register on some songs, which I think makes the songs more interesting.&quot;\n\nOn Taking Himself Less Seriously:\n\n&quot;I remember thinking I have a platform and I have a microphone and I have a position of power, so let me try to change some things, maybe for the better. And people did not like that. I remember having this conversation where somebody said to me, \u2018When you are perceived as yourself, people have a hard time relating to that.\u2019 And I said, \u2018I think you are misperceiving yourself to be someone who has a choice.\u2019 I\u2019m not saying this like, \u2018Hey, world, my message is this, and you\u2019re going to like it.\u2019 That\u2019s not what it\u2019s about. It\u2019s about the moment in which you are caught off guard, and you\u2019re on tour and you\u2019re not doing what you\u2019re supposed to be doing, but you\u2019re still getting a check.&quot;\n\nOn The Benefits Of Ghostwriting:\n\n&quot;I don\u2019t have any beef with the world. I\u2019ve got a bunch of kids and a beautiful girlfriend and a house that I love and that\u2019s not an accident. And I\u2019ve worked for that and I\u2019ve gotten that because I\u2019m an easygoing guy who really enjoys life. And I\u2019ve always been super-vigilant about my ability to access joy.&quot;\n\nOn Being A 'Bubblegum Pop Star':\n\n&quot;I was, for a very brief period of time, a bubblegum pop star. I was in the No. 1 pop band of the late-\u201990s, early 2000s. I\u2019ve had 18 or 19 top 10 songs. I have the right to call myself a singer, I guess. But I would never want to turn my back on being a guitar player and a songwriter. And the reason why is because I have a fan base that loves me for being a guitar player and a songwriter, and it\u2019s very separate from the people who like my pop songs.&quot;\n\nHis Concert Review Regrets:\n\n&quot;I was like, \u2018Look, we\u2019re going to be under these bright lights, and it\u2019s going to be this one moment in time where we\u2019re going to play it for you. And we want to give you the most you can get for your dollar.\u2019 And then I think I had a bad couple of years. And I lost myself and tried to get myself back, and now I\u2019m playing shows where, like, I\u2019m not having a good time.&quot;\n\nOn Live Music Criticism:\n\n&quot;I think we\u2019re overintellectualizing music. The discussion about music is so myopic and minuscule. \u2018He did this, and it\u2019s a disservice to this particular song.\u2019 And I\u2019m like, well, maybe you don\u2019t like this song. If you don\u2019t like it, just don\u2019t buy it. But that\u2019s not the way it works anymore. So I think we\u2019ve been eating the meat and spitting out the bones. You just want to get the marrow.&quot;\n\nThe New York Times concludes by highlighting the vocal coaching he undertook in 2015 in order to overcome his vocal issues:\n\n&quot;I went and saw a vocal coach who I had seen when I was on my 'John Mayer's Journey Continues' world tour. She was a big hero of mine. She\u2019s the one who said to me, 'John, you just don\u2019t like how you sound. Why don\u2019t you do something about it?' I said, 'I can\u2019t afford to do anything about it.\u2019 She said, 'That\u2019s not true.\u2019 &quot;\n\nWatch our video of John's recent live cover of 'A Hard Day's Night' here:]" time="2.737"><properties><property name="score" value="0.9781781435012817" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[From the outside, the Taurus is a boring, bland looking sedan that's unapologetically designed to appeal to people who are price- and value-conscious, but still want to be able to put some stuff in the trunk. But it has a lot going on under the hood, and it's a great way]" time="0.167"><properties><property name="score" value="0.8674129247665405" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[COLD SPRING, Ky. (WKYT) - The Kentucky Department of Fish and Wildlife Resources says an investigation into the shooting of a horse at the Country Pleasures horse farm has been completed.\n\nAccording to KDFWR, a citation has been issued to a person involved in the incident. The citation will be mailed to the person's last known address. The citation is for the willful waste of wildlife.\n\nThe investigation by KDFWR has revealed that on June 11th, 2013, three horses were shot on a farm off of Logan Road in Bath County. One horse, a white and gray Paint Mare was found dead. The mare, named &quot;Sissy,&quot; had a history of being aggressive towards other horses. A chest wound indicated the horse had been shot at close range.\n\nThe second horse, a dark brown and white mare named &quot;Bonnie,&quot; had been shot in the rump. Bonnie was treated by a veterinarian for the wound, and released. The third horse, a dark brown and white gelding named &quot;Ghost,&quot; had a wound on his shoulder. He was treated by a veterinarian and released.\n\nKDFWR says the investigation concluded that the shooting of the horses was accidental.\n\nKDFWR says no additional charges will be filed in this case.]" time="0.693"><properties><property name="score" value="0.9643329083919525" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Please enable Javascript to watch this video\n\nST. CHARLES, MO (KPLR) - The now former police chief of St. Charles has been arrested for drunk driving. A spokesperson for the St. Charles Police Department says Officer Robert Hertzell was on duty and was driving a marked police vehicle.\n\nSt. Charles Police Chief David Kinkead has put Hertzell on unpaid administrative leave and has initiated an internal investigation.\n\nSt. Charles Mayor Sally Faith says Hertzell will remain on unpaid leave until the investigation is complete.\n\nChief Kinkead released a statement saying, \u201cI was very disappointed to learn about this incident. As a police chief, I hold myself and my command staff to the highest standard. I expect my officers to be role models and to demonstrate professionalism and respect. We have a lot of hard working men and women in the St. Charles Police Department and they are to be commended for their work on a daily basis. The actions of one officer, however, is not reflective of the entire command staff or the great work that our officers perform each day. We must continue to move forward with the work that we have been doing to create a safe and vibrant community for all of our residents and visitors.\u201d\n\nHertzell has been with the St. Charles Police Department since 1995.]" time="0.617"><properties><property name="score" value="0.998702198266983" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[What happened\n\nShares of Applied Optoelectronics Inc (NASDAQ:AAOI) were getting hammered on Thursday after the networking specialist reported a quarterly loss and revenue that missed estimates. The news sent the stock down 15% at the time of this writing, putting it on pace to be the stock's worst day since late 2016.\n\nSo what\n\nInvestors were expecting Applied Optoelectronics to lose money in the fourth quarter, as it has lost money in each of the past two quarters. But the loss this time around was bigger than expected, with Applied Optoelectronics reporting a non-GAAP net loss of $0.40 per share on a 14% decline in revenue to $93.9 million. Analysts, on average, were only looking for a loss of $0.33 per share on $92.1 million in revenue.\n\nNow what\n\nCEO Paul E. Jacobs called the results a &quot;solid performance&quot; and said the company's overall sales were &quot;driven by market growth&quot; and its plan to &quot;increase long-term shareholder value.&quot; Applied Optoelectronics also noted that its total customer backlog rose 20% year over year to $347 million, which is more than analysts were looking for.\n\nWith the company's stock now trading down 45% over the past year, however, it will need to show much more than &quot;solid performance&quot; to prove that its turnaround is working. To start with, it needs to stop the losses and start producing profits again. If it can do that, then the rise in the backlog will be a positive sign, and investors may start to take a fresh look at the company again.]" time="0.905"><properties><property name="score" value="0.984209269285202" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[These two teams have met eight times before with the Houston Dynamo winning five, the most recent meeting coming in last season\u2019s knockout round where Houston won 1-0 in Dallas.\n\nPROBABLE LINEUPS\n\nFC Dallas (4-4-2): Hartman; Loyd, John, Benitez, Morrow; Castillo, Jackson, Michel, Shea; Perez, Cooper.\n\n\n\nHouston Dynamo (4-3-3): Hall; Ashe, Taylor, Horst, Chabala; Moffat, Boswell, Creavalle; Bruin, Weaver, Davis.\n\n\n\nDID YOU KNOW?\n\nThis is the fourth time Dallas has made it to the semifinals. They've reached the final twice, losing to the New England Revolution in 2005 and D.C. United in 2007.\n\nDallas has won the U.S. Open Cup five times, the last coming in 1997. They are the only team in MLS with five Open Cup championships.\n\nDallas has reached the quarterfinals in three of the last four seasons.\n\nOnly one other team, the New England Revolution, has more postseason appearances (8) in the last eight seasons than the Dynamo.\n\nThe Dynamo\u2019s last win in the quarterfinals was over FC Dallas in 2008 when the Dynamo won 2-0 at Robertson Stadium.\n\nThe Dynamo\u2019s last two wins in the quarterfinals came on the road. They defeated Seattle Sounders FC 2-0 at Starfire Sports Complex in 2012, and the Portland Timbers 1-0 at PGE Park in 2011.\n\nForward Will Bruin has scored 12 goals in his last 18 matches in all competitions, including six in his last eight matches in MLS play.\n\nSuspended : None\n\n: None Suspended after next caution : None\n\n: None International duty : None\n\n: None Injury Report: None\n\nProjected Starting XI: Kevin Hartman (GK), Zach Loyd, Matt Hedges, George John, Michel, Jackson, Andrew Jacobson, Fabian Castillo, David Ferreira, Brek Shea, Kenny Cooper\n\nNotes: FC Dallas are undefeated in eight games at home this season. The last time they were shutout at home was against Real Salt Lake in May.]" time="1.156"><properties><property name="score" value="0.9969680607318878" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Sorcerous Origins\n\nUnbound Dancer\n\nUnbound Dancer\n\nSometimes called veils, a particular style of dancer that have sprung up in different cultures across the planes.\n\nCeaseless Dance\n\nAt 1st level you learn a ritual that allows you to spend 8 hours dancing and stamping out a particular pattern. This can be done during a short rest and while you do this you can spend up to half your sorcerer level (rounded up) in sorcery points.\n\nYou must finish a long rest before you can perform this dance again.\n\nWhen you complete the dance you gain the following benefits:\n\nYou have a base movement speed of 15 feet and your maximum movement speed increases by 5 feet.\n\nYou can add half your proficiency bonus (rounded up) to all dexterity ability checks.\n\nYou gain resistance to poison damage.\n\nYour skin becomes extremely smooth and soft and you have advantage on all charisma ability checks.\n\nYou have a +1 bonus to AC.\n\nThe following additional effects apply based on the number of sorcery points you spent on this ability:\n\n4-7 sorcery points: At the end of each of your turns you gain 5 feet of movement.\n\n8-11 sorcery points: You gain resistance to fire and cold damage.\n\n12-15 sorcery points: You gain a flying speed of 30 feet.\n\n16-19 sorcery points: You gain a second resistance to fire and cold damage.\n\n20 or more sorcery points: At the end of each of your turns you gain 10 feet of movement.\n\nPerform the Dance\n\nYou can only perform this dance once per long rest.\n\nStarting at 6th level you gain the ability to copy other creatures movements. You can use your action to study a creature within 5 feet of you for up to one minute. You can use your bonus action to then attempt to imitate that creature for one minute, using one of your known dance forms.\n\nYou must complete a long rest before you can use this ability again.\n\nExpanded Dance Forms\n\nAt 14th level you gain two additional dance forms. These can be from the expanded list on the next page. You can only perform each of these dances once per long rest.\n\nDance Forms\n\nDancing Viper: While dancing you can spend up to 4 sorcery points on the following effects:\n\n2 sorcery points: Gain a burrow speed of 20 feet.\n\n4 sorcery points: Gain a climbing speed of 30 feet.\n\n6 sorcery points: You gain resistance to piercing and poison damage.\n\n8 sorcery points: Gain a fly speed of 20 feet.\n\nDancing Tiger: While dancing you can spend up to 6 sorcery points on the following effects:\n\n4 sorcery points: You gain a climb speed of 30 feet.\n\n6 sorcery points: Your unarmed strikes are considered magical and gain the light property.\n\n8 sorcery points: You gain advantage on all dexterity ability checks.\n\n10 sorcery points: You gain resistance to slashing damage.\n\n12 sorcery points: Your movement speed increases by 10 feet.\n\nDancing Monkey: While dancing you can spend up to 6 sorcery points on the following effects:\n\n4 sorcery points: You gain a climb]" time="1.656"><properties><property name="score" value="0.9458777904510498" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[It's hard to imagine a wittier book than one subtitled A Primer on the Tactics of Scientific Research. But, then, David M. Kaplan and Robert M. May are not your ordinary scientists. At one time or another, they've all but single-handedly invented chaos theory, championed the concept of selfish genes, coined the phrase &quot;keystone species,&quot; and discovered the homeotic genes responsible for the formation of the vertebrate body plan. For sheer force of intellect, they're among the world's most eminent biologists.\n\nYou might, then, think that there would be little for these geniuses to learn from a primer in scientific methodology. But, as it turns out, that's just not so. In the same style of mordant humor that makes &quot;The Ant and the Peacock&quot; one of the best-selling books on evolutionary biology, Kaplan and May delightfully explain why it's difficult for a scientist to break away from deeply ingrained habits of thought.\n\nThey also introduce the problem of citation errors--a problem that has become all too apparent in the aftermath of The Bell Curve, whose authors committed a blizzard of them, citing everything from studies that were never performed to nonexistent journals. Kaplan and May suggest ways to prevent citation errors and to identify them when they do occur. They reveal that it's possible to fabricate a scientific journal, and to fool people into believing it's real.\n\nBut they don't confine their attention to egregious examples of scientific malfeasance. On the contrary, &quot;The Art of Scientific Research&quot; is written as a series of serious and not-so-serious instructions on how to conduct research, how to get it published, and how to cite it accurately. There's even a section on the art of collaboration, complete with sage advice on how to make it work. &quot;The Art of Scientific Research&quot; is a book for anyone who does science, or wants to.]" time="0.930"><properties><property name="score" value="0.10066980123519897" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[What I love about travelling is you get a chance to taste local food, pick up a language and meet new people.\n\nTravel is a great way to expand your cultural awareness and try new things.\n\nBe open-minded to the experiences.\n\nYou\u2019ll see a lot more on a tour, than just your average solo tourist.\n\nI love travelling as much as I love food, so naturally I\u2019m going to blog about them both.\n\nI\u2019ve created this blog to share some of my experiences, places, and reviews.\n\nHere you can find my best recommendations for travel destinations and hotels.\n\nI want you to enjoy your holiday as much as I did.\n\nSo, keep an eye out for my next travel blog.]" time="0.374"><properties><property name="score" value="0.925939679145813" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Today, in the form of yet another unfortunate casualty of homelessness, a man\u2019s dead body was found on the street just outside of a McDonalds. Just a few hundred meters from the student dorms, where a guy was recently stabbed to death.\n\nI\u2019m really getting sick and tired of living in this city, and I\u2019m beginning to seriously think about moving somewhere else. A friend of mine has been living in Boston for the past 2 years, and he\u2019s constantly telling me how amazing Boston is, and how beautiful it is there.\n\nI\u2019m going to head to Boston for the summer, just so I can get a better idea of what it\u2019s like living there. I\u2019m really starting to lose interest in living here, and I want to try out living somewhere else before I settle on doing so. If Boston is not the place for me, I want to at least have an idea of where I should move.\n\nI do think, however, that Boston might be the place for me, and I\u2019m going to be there for the summer, and maybe even beyond. That is what I\u2019m hoping for at least, but we\u2019ll see how it goes. I think I\u2019m going to take a long trip back to India during this time, as well.\n\nI\u2019m going to try to take a lot of pictures, and post them up here on the blog. I will be a bit busy with work and school and my internship, so I don\u2019t know how much time I will have to do that, but I will do my best.]" time="0.805"><properties><property name="score" value="0.999800443649292" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Our friends over at Road and Track have the scoop on what will be the new Mustang. You can read the entire article here, but it goes something like this.\n\nFord Mustang will have a larger, more powerful engine with a V8. A V6 will be offered, but as a two-seater.\n\nThe V8 will be a 5.0 liter, not a 3.7, and the V6 will have the same displacement as the current 3.7, but with the variable valve timing engine from the Ecoboost engine.\n\nIt will be called the Mustang GT, and is the car most of us are interested in.\n\nThe rear-drive GT model will be gone, replaced by an all-wheel drive GT.\n\nFord\u2019s CVT will be the only transmission.\n\nTo be honest, I\u2019m disappointed. The Mustang has always been about high performance.\n\nWhat is the reason for the V8 to lose power? Will we even feel the change? It is possible the V6 will make the most horsepower of all Mustang models, while still being smaller than the 5.0.\n\nDoes anyone want an all-wheel drive Mustang? Sure, you could argue that AWD will help with acceleration in the snow, but it\u2019s pretty rare for the Northeast to get that much snow, and I don\u2019t need all-wheel drive for an off-road trip to the Catskills.\n\nThe only reason I would consider buying an all-wheel drive Mustang is for the style. I think it will look very good.\n\nSo what do you guys think? Are you upset like I am, or are you okay with the changes?]" time="0.887"><properties><property name="score" value="0.9814767837524414" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[La Asamblea, organo pol\xedtico m\xe1ximo de la Administraci\xf3n Municipal, ejerce su funcionamiento en dos periodos:\n\na) Asamblea Ordinaria.\n\nb) Asamblea Extraordinaria.\n\nSe encarga de:\n\na) Intervenir en la conformaci\xf3n de las Autoridades de los Organismos Locales, ejerciendo la tutela de la Ley Org\xe1nica del R\xe9gimen Electoral y del Sufragio, Ley Org\xe1nica de los Partidos Pol\xedticos y dem\xe1s leyes y disposiciones aplicables.\n\nb) Establecer los niveles de las remuneraciones mensuales y de prestaciones de los servidores p\xfablicos municipales, por cualquier concepto.\n\nc) Aprobar el Plan Anual de Gobierno Municipal, el Presupuesto Municipal y su Reglamento.\n\nd) Resolver las quejas, reclamos y peticiones de los habitantes de la Municipalidad de Hualqui.\n\ne) Dictar las normas a que debe sujetarse la actividad administrativa del Municipio, en su ordenamiento jur\xeddico interno y en sus relaciones con otras Administraciones P\xfablicas.\n\nf) Elegir a sus integrantes y reglamentar su funcionamiento interno.\n\ng) Autorizar la emisi\xf3n de los t\xedtulos representativos de la deuda p\xfablica municipal.\n\nh) Dictar resoluciones administrativas que no impliquen la creaci\xf3n, modificaci\xf3n o supresi\xf3n de tributos municipales.\n\ni) Establecer el r\xe9gimen general para la designaci\xf3n de personas naturales o jur\xeddicas, como encargadas de la prestaci\xf3n de servicios p\xfablicos municipales, as\xed como el r\xe9gimen de contrataci\xf3n y supervisi\xf3n de dichas personas.\n\nj) Dictar resoluci\xf3n de acuerdo con el contenido de los proyectos de acuerdo presentados por la Contralor\xeda General del Municipio, as\xed como en el caso de la resoluci\xf3n de la auditor\xeda de las cuentas municipales.\n\nk) Dictar resoluciones que contengan las decisiones y pol\xedticas municipales, las cuales sean de trascendencia y sean definitivas, y que deban ser obedecidas por todos los habitantes del Municipio.\n\nl) Autorizar el otorgamiento de exenciones tributarias, por resoluci\xf3n motivada de la misma Asamblea.\n\nm) Dictar las normas sobre las sanciones a los concejales o autoridades de los organismos y empresas municipales que cometan irregularidades en la gesti\xf3n de la Municipalidad.\n\nn) Dictar resoluci\xf3n sobre las sanciones a los servidores p\xfablicos de la Municipalidad.\n\no) Elaborar el proyecto de Ordenanzas Municipales y otros reglamentos administrativos]" time="1.943"><properties><property name="score" value="0.9990601340929667" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99906013&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99906013
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Australian Energy Regulator highlights South Australian electricity network overspend\n\nThe Australian Energy Regulator has released a draft report highlighting the network cost of electricity distribution in South Australia. The report contains the results of an investigation into South Australian electricity distribution costs undertaken by the Australian Energy Regulator (AER).\n\nThe AER\u2019s report found that the electricity distribution networks in South Australia, namely ElectraNet and SA Power Networks, are materially overspending compared to the amount of revenue they receive from consumers.\n\nThe report states that network costs in South Australia have been growing rapidly, due to a range of factors, including but not limited to:\n\nLow and decreasing real term retail revenue\n\nPoor use of time based network costs\n\nTime based revenue regulation\n\nChanging end-user demand for electricity\n\nInvestments in the network that are not supported by regulation.\n\nThe report further states that distribution charges represent around a quarter of the total cost of electricity, and are ultimately passed on to consumers through higher electricity bills. The report notes that the three largest state owned electricity distribution companies (including ElectraNet and SA Power Networks) have all seen rapid increases in their distribution costs since 2013.\n\nThe report recommends that the South Australian Government take action to:\n\nUndertake an urgent review of electricity retail regulation\n\nUndertake an urgent review of the existing regulation of distribution charges\n\nExamine mechanisms to enable local distribution networks to invest in assets that improve reliability for customers, and facilitate the integration of renewable energy into the network\n\nThe report also recommends that the Australian Energy Regulator should seek to work with industry to seek additional reforms to be considered by the Australian Energy Market Commission (AEMC).\n\nThe report is open for submissions and can be found here.\n\nMore information on how energy networks operate and cost can be found here.\n\nSee our website for further information on energy and power regulation.]" time="1.046"><properties><property name="score" value="0.9114988644917806" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[A special thanks to all the people who have read this story and have offered me constructive criticism.\n\nA Big Thanks to The Prankster and Sage, also to Herr and themuseforlove for providing such helpful and encouraging critiques.\n\nEnjoy.\n\nHis life had been pretty good lately. The old man had finally started letting him do more things on his own, such as make deals with the city of Chicago.\n\nAt the moment he was in his office getting a briefing from Mr. Fredrickson, his assistant.\n\n&quot;The latest bounty is on something that is causing a lot of damage to the people in the city. It's probably best to explain the damage to you in person,&quot; Fredrickson suggested.\n\nThey both headed to the police station. Fredrickson had a theory as to what was causing the damage and had contacted the police.\n\nFredrickson's theory turned out to be correct. It was a monster, and it had been causing a lot of damage and injuries, but they were on the right track to capturing the monster.\n\n&quot;You mean that thing is going to be mine,&quot; the old man said when he got word that the monster had been captured.\n\n&quot;Yes sir, we are on our way to get it now,&quot; the officer informed him.\n\n&quot;I'll meet you at the stadium,&quot; the old man said.\n\n&quot;You want to come to the stadium,&quot; Fredrickson asked.\n\n&quot;I do, I want to see this thing,&quot; the old man said with a grin.\n\n&quot;Why,&quot; Fredrickson asked.\n\n&quot;I just want to see it,&quot; the old man answered.\n\nAs the two walked back to the old man's car Fredrickson suggested he do something a little different with the bounty this time.\n\n&quot;This thing is more dangerous than the last two that were captured, why do you want to go see it,&quot; Fredrickson inquired.\n\n&quot;Let's just say that I want to make sure that the right person is getting the bounty,&quot; the old man answered.\n\n&quot;Fine, I'll call them and tell them to make sure to bring the creature in the arena,&quot; Fredrickson said.\n\n&quot;Good, call them now,&quot; the old man said.\n\nFredrickson called and told the team that they were bringing in the monster and it was going to be brought into the arena and they had to get the area ready.\n\nAfter the call was made, the old man and Fredrickson drove to the arena.\n\nWhen they got there the old man went to the back area and took off his suit and put on a set of referee's gear.\n\n&quot;You sure you want to do this,&quot; Fredrickson asked.\n\n&quot;I'm sure, now go call them and tell them we're ready,&quot; the old man ordered.\n\nFredrickson called and then the two waited for the team to show up.\n\nAfter a few minutes the team showed up with the monster in tow.\n\n&quot;It looks mean,&quot; Fredrickson said as the team pulled the creature into the arena.\n\n&quot;It does look mean,&quot; the old man replied.\n\nAs the two watched the creature in the arena the old man asked a few questions of the men.\n\n&quot;When did it show up here,&quot; he asked.\n\n&quot;It came in about six months ago,&quot; the leader of the team answered.\n\n&quot;How did you capture it,&quot; the old man asked.\n\n&quot;We don't know, we just heard a scream and when we went to check on it, it was already bound up,&quot; the leader answered.\n\n&quot;Do you have a picture of what the thing looks like,&quot; the old man inquired.\n\n&quot;Yes, we have a picture of it,&quot; the leader answered.\n\n&quot;May I see it,&quot; the old man asked.\n\nThe leader handed the old man a picture of the creature.\n\n&quot;This doesn't look like the same one,&quot; the old man said as he looked at the picture.\n\n&quot;We don't know, maybe this is a different one,&quot; the leader suggested.\n\n&quot;This looks more like the picture of the monster I fought a long time ago,&quot; the old man said as he handed the picture back.\n\n&quot;It's not the same thing, the one you fought was green and this one is yellow,&quot; the leader said.\n\n&quot;Yes, I know what I'm talking about,&quot; the old man said.\n\n&quot;If it's not the same thing then why are you concerned,&quot; the leader asked.\n\n&quot;This monster could be very dangerous, why haven't you done more research,&quot; the old man inquired.\n\n&quot;I don't know,&quot; the leader admitted.\n\n&quot;You need to study this thing, do some research and if you can't find anything, then maybe you should move it to another city,&quot; the old man suggested.\n\n&quot;I'll talk to my superiors,&quot; the leader said.\n\n&quot;Good,&quot; the old man replied.\n\n&quot;I'm sure that I can find more information if I can get this one back to my lab,&quot; Fredrickson said.\n\n&quot;I'm sure you can,&quot; the old man replied.\n\nThe team was about to move the creature into the arena to meet with their local hero when a man walked up and started talking to the leader.\n\nThe old man looked at the man and recognized him as a former wrestler, and a very dangerous one at that.\n\n&quot;Hey, that guy is the one who can beat this monster,&quot; the former wrestler said.\n\n&quot;How can you be sure,&quot; the leader asked.\n\n&quot;I can beat this monster because I once fought a monster like this one and he defeated it,&quot; the former wrestler explained.\n\n&quot;Why don't you do it,&quot; the leader suggested.\n\n&quot;Fine, but I want the same money that you were going to give the other guy,&quot; the former wrestler requested.\n\n&quot;Okay, let's get this over with,&quot; the leader said.\n\n&quot;Yes sir,&quot; the former wrestler said.\n\nThe former wrestler walked over to the monster and started talking to it. The monster got angry and threw a punch at him.\n\n&quot;It's angry,&quot; Fredrickson said.\n\n&quot;I can see that,&quot; the old man replied.\n\n&quot;If he can't beat it, we can always use it,&quot; Fredrickson said.\n\n&quot;Let's see what happens,&quot; the old man said.\n\nThe former wrestler dodged the punch and was able to get his arms around the creature's neck and was able to pin it to the ground.\n\n&quot;This could be bad,&quot; Fredrickson said.\n\n&quot;It's not going to be bad,&quot; the old man replied.\n\nThe former wrestler held the creature down and then twisted its head around until it snapped off.\n\n&quot;That's one dangerous man,&quot; Fredrickson said.\n\n&quot;He is, let's go talk to him,&quot; the old man said.\n\n&quot;Okay,&quot; Fredrickson said.\n\nThe two walked over to the former wrestler and talked to him.\n\n&quot;You're good, I want to talk to you,&quot; the old man said.\n\n&quot;I'm listening,&quot; the former wrestler said.\n\n&quot;What are you doing now,&quot; the old man asked.\n\n&quot;I'm working on the docks,&quot; the former wrestler answered.\n\n&quot;I have a friend that owns a gym, and I'm sure that I could get him to help you out,&quot; the old man suggested.\n\n&quot;If I go to this gym, who will I be fighting,&quot; the former wrestler asked.\n\n&quot;That depends on who you think is the right person,&quot; the old man answered.\n\n&quot;I know, but how do]" time="3.923"><properties><property name="score" value="0.9640186514173236" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[\u201cJurassic World\u201d is going bigger, with bigger stars and a bigger budget.\n\nJ.A. Bayona (\u201cA Monster Calls\u201d) is in final negotiations to direct the sequel to \u201cJurassic World,\u201d the hit dino tentpole that earned $1.67 billion worldwide last summer. Colin Trevorrow and Derek Connolly, who penned the script for \u201cJurassic World,\u201d penned the script for the sequel.\n\nChris Pratt and Bryce Dallas Howard are both expected to return, but not necessarily as the leads. Frank Marshall and Pat Crowley are producing the pic.\n\nWhile Bayona and Trevorrow are considered to be a package, the studio is trying to decide whether to retain Trevorrow as a producer or have Bayona work with an experienced producer. That will likely be decided in the next week, with a new writer possibly coming aboard, but for now Bayona is poised to make his English-language feature directing debut.\n\nPlot details are vague, but sources say the film will follow in the footsteps of 2015\u2019s \u201cJurassic World\u201d in focusing on the next generation of the dino franchise. The follow-up will be produced by Frank Marshall and Patrick Crowley. The movie is set for release on June 22, 2018.\n\nBayona broke out in Hollywood with his 2007 thriller \u201cThe Orphanage,\u201d and directed the ghost story \u201cThe Impossible,\u201d which earned star Naomi Watts an Oscar nomination.\n\nBayona is repped by WME, Anonymous Content and attorney Marielle Tenghe.\n\nDeadline Hollywood first reported the news.]" time="0.838"><properties><property name="score" value="0.7845679521560669" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Malcolm Butler played a game with a fractured left fibula in Super Bowl LII, according to a league source.\n\nAfter being carted off the field in the fourth quarter with the injury, the cornerback stayed in the game to finish the Patriots\u2019 win over the Eagles. Butler recorded a crucial interception of Eagles quarterback Nick Foles late in the fourth quarter, which helped the Patriots run out the final 3:21 to seal the victory.\n\nButler had been in noticeable pain after being injured while defending a throw to Eagles receiver Alshon Jeffery.\n\nHe missed practice Wednesday and Thursday last week because of an illness. Butler was not listed on the injury report this week. He practiced Friday, but sat out Sunday\u2019s walkthrough.\n\nThe cornerback was not made available for comment after the game.]" time="0.410"><properties><property name="score" value="0.6974356174468994" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.69743562&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.69743562
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[An 18-year-old man was killed in a crash at Interstate 4 and Kirkman Road in Orlando on Monday morning, officials said.\n\nNo other injuries were reported in the crash, which was reported at 9:24 a.m., said Lt. Robert Terrell of the Florida Highway Patrol.\n\nHe said Juan Bautista-Martinez of Orlando was driving a 1997 Nissan Maxima west on Kirkman and ran off the road into the median.\n\nThe car, which was heavily damaged, then went over the barrier and landed on its roof in the eastbound lanes of I-4, Terrell said.\n\nAll eastbound lanes were closed for a short time as the vehicle was removed.\n\nThis is a breaking news story. Check back for more information.\n\nglotan@orlandosentinel.com or 407-420-5774]" time="0.379"><properties><property name="score" value="0.9998243451118469" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Jules &quot;GhostOf&quot; Wang is the jungler for Team Dragon Knights.\n\nBiography\n\nJules \u201cGhostOf\u201d Wang hails from Nanjing, China. He was introduced to League of Legends by a friend in his last year of high school. From the moment he started playing he was hooked on the game and quickly moved into playing professionally.\n\n2015 Preseason\n\nOn November 24, 2014 it was announced that along with Cloud9 Tempest's LCS roster and Nientonsoh, they would be the main team at the new Cloud9 Challenger roster.[1] After that he played in some amateur tournaments along with the rest of the Cloud9 Challenger roster.\n\n2015 Season\n\nAfter playing for Cloud9 Challenger in the NACS Spring Qualifier, GhostOf joined Team Dragon Knights in February, when the team was invited to the Spring Season of NACS. They finished in third place in the regular season, with a 6-4 record. In the playoffs, they lost in the semifinals to eventual tournament winner Renegades, but won the third place match against Cloud9 Tempest to secure a spot in the summer promotion tournament. However, they were unable to advance to the summer split, as they lost to Team 8 in the first round of their promotion match-up.\n\nTeam Dragon Knights competed in the Summer Promotion Tournament, where they lost 3-0 to Team Fusion in the semifinals. The team also lost their placement matches against Team Coast, meaning that they would be playing in the 2016 NACS Spring Season.\n\n2016 Season\n\nAfter winning the NACS Summer Qualifier, Team Dragon Knights qualified for the 2016 NA LCS Spring Season. In the spring season, they posted a 6-12 record in the round robin and did not qualify for playoffs. However, with the disqualification of Renegades (due to visa issues), TDK received the top seed into the Summer Promotion Tournament, where they were eliminated in the first round by Team Dignitas.\n\nTrivia\n\nLoves to watch anime.\n\nTournament Results\n\nInterviews\n\nRedirects\n\nThe following pages redirect here. There are three types of pages that may appear on this list:\n\nNames formerly used by the player in competition\n\nNicknames or alternative spellings or capitalizations of the player's name\n\nCommon typos that are frequently searched for\n\nThe list is generated automatically. To request an addition to the list, you may use this form.\n\n\n\nNo results found]" time="1.325"><properties><property name="score" value="0.8996343612670898" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[To compare the electrochemistry of a one-step and a two-step biosynthetic pathway, the co-culture of methanogenic archaea and the acetogenic bacterium Butyrivibrio fibrisolvens was examined. Acetate production from glucose in batch culture was investigated, comparing cells from co-culture with pure B. fibrisolvens and pure Methanosarcina barkeri cells. After 22 h, glucose was depleted in pure M. barkeri cultures but acetate production was not complete. After 22 h, glucose was depleted and glucose uptake and acetate production were completely ceased in pure B. fibrisolvens cultures. When cells from the two pure cultures were co-cultured, the cells from M. barkeri appeared dead in the co-culture after approximately 10 h. Cell growth and acetate production occurred in the co-culture, but was lower than the growth rate and the acetate yield in pure B. fibrisolvens cultures. Pure M. barkeri cells grown with pure B. fibrisolvens cells exhibited no growth after 8 h, but some cells were still viable. However, after 18 h all M. barkeri cells were dead and the co-culture was dominated by pure B. fibrisolvens cells. The results indicated that co-culturing of M. barkeri cells with B. fibrisolvens was not an efficient way of acetate production. Acetate production was possible when pure B. fibrisolvens cells were first grown to a certain population and then M. barkeri cells were added. However, complete utilization of glucose was not possible in such co-culture. In this study, the electrode setup was not optimized to detect early anaerobic cells.]" time="0.793"><properties><property name="score" value="0.9945335686206818" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99453357&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99453357
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[If your personal philosophy is a little bit Taoist and a little bit Marxist, there's a new place in LA to scratch that particular philosophical itch. Tao Mao is the only East-meets-West neo-communist tiki bar in LA, a three-story Chinatown bordello of poi dogs and palm trees. To get there, you enter from a rear alley and walk past the washboard abs of dancers in the Bumpin' Uglies Go-Go Bar to the karaoke bar called Red 7, and from there climb a red-lit staircase to the restaurant.\n\nTao Mao, &quot;the great origin,&quot; is the Chinese name for the Milky Way galaxy, as seen from the Earth. But it's also the name of a star in that galaxy. And in Mandarin, Tao Mao is pronounced &quot;Dow Mao,&quot; which sounds like &quot;D.M.A.,&quot; or &quot;doctor of philosophy.&quot; And that is exactly what Tao Mao is\u2014a tiki bar where all the drinks have the names of philosophers.\n\nYou may be wondering: Is Tao Mao a tiki bar for philosophers? Or a bar for philosophers that serves tiki drinks? Or is it both? Does it matter?\n\nActually, no, it doesn't. Tao Mao is a philosophical tiki bar, that's all you need to know. The more you think about it, the less sense it makes.\n\nBefore I try to explain Tao Mao, I should mention that the bartender told me the bar's name is pronounced &quot;Dow Mao.&quot; And he is a white dude. And yet... all the cocktail names are in Mandarin. And he only speaks English. So what the fuck do I know? Maybe I'm the crazy one.\n\nI can tell you that the people who seem most at home at Tao Mao are little clusters of friends\u2014mainly white dudes who look like they'd be hanging out at some over-designed tiki bar that was actually inspired by Tao Mao. For example, I met two TV writers who were there for the first time. The owner was there with his mom. The first time I visited, the woman in the next booth to me was there with her sister, and]" time="1.083"><properties><property name="score" value="0.0013606846332550049" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Considered a New York classic, the Fourth Avenue Deli has been a staple of the city\u2019s Lower East Side for decades. Deli aficionados and tourists alike have been lining up outside the restaurant\u2019s doors for more than 50 years, eagerly awaiting their first taste of Katz\u2019s famed pastrami sandwich. Katz\u2019s special blend of spices, smoked over pinecones in traditional ovens, and the deli\u2019s commitment to quality meat make the perfect recipe for mouthwatering pastrami. Every bite of the succulent, slow-cooked meat is an explosion of flavors.\n\nKatz\u2019s pastrami, made with fatty beef from the belly, is a little softer than that of the more common brisket, but has a much richer, beefier flavor. The deli also offers hot pastrami, which is much more spicy than its cold counterpart. But it\u2019s Katz\u2019s hearty slices of hand-carved corned beef, a delicacy with a history dating back to the 1700s, that has helped make the deli a household name.]" time="0.507"><properties><property name="score" value="0.9989047050476074" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[IceBerg Originally Posted by No she was not a &quot;sexy voice&quot; in any way, shape or form. I've heard many, many callers on various airwaves that were/are far more seductive and far more sexy sounding than this woman ever will be.\n\n\n\nWhat did you ever think was going to happen when you allowed &quot;a show host&quot; to be able to call into the show and be able to communicate with the host and be able to flirt with the host? I mean, that's pretty much all that happened. It didn't take a rocket scientist to see what was going to happen.\n\n\n\nAnyhow, whatever the circumstances were, I think it's pretty cool that she came back to apologize. It's pretty easy to judge when you're not in that situation.\n\n\n\nAnd to the person that said &quot;You have to look for and find bad things that can be found in anything to be unhappy&quot;, you're absolutely right, but it doesn't take a rocket scientist to look for the bad and find it.]" time="0.604"><properties><property name="score" value="0.9998156428337097" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Boeing to Cease Orders for Canadian 737 Max 8\n\nThe sudden grounding of Boeing\u2019s 737 Max aircraft is causing a major disruption to the global airline industry.\n\nThis morning, Canadian plane manufacturer Bombardier announced that it was ending its contract to sell 75 of its CSeries aircraft to the lessor, and parent company of Norwegian Air Shuttle.\n\nAnd now, it has been revealed that U.S. aerospace giant Boeing has cancelled its orders for the Canadian plane, putting a hold on any future deliveries.\n\nThe grounding of the Max model of the aircraft has thrown the airline industry into turmoil. While Boeing does not have a huge presence in Canada, it is estimated that the company has invested $6.4 billion into the country in the last two years alone.\n\nIn a statement, Boeing said it \u201ccancelled a previously announced purchase order for 16 Next-Generation 737-800 passenger aircraft from a U.S. customer and added a new order for 16 Next-Generation 737-900ER passenger aircraft from the same customer.\u201d\n\nIt added, \u201cthis change resulted in a reduction of eight 737-8 MAX deliveries from the previously announced 64 deliveries through 2023.\u201d\n\nWith the order having been cancelled, it is unclear if it will be restored in the future. Boeing said the cancellation of the order was part of its routine actions \u201cin response to market opportunities, customer preferences, and our results in a competitive environment.\u201d\n\nBoeing says it remains committed to its relationship with Canada, but the news comes amid concern over the safety of the aircraft. The model was involved in two deadly crashes in Indonesia and Ethiopia, where 346 people died. The cause of the crashes remains unclear, but both involved issues with the new anti-stall system.\n\nAfter the Ethiopian crash, U.S. President Donald Trump said the plane was \u201ccertainly\u201d unsafe to fly.\n\nFollowing the grounding of the aircraft, Boeing has grounded its 737 Max planes across the globe. In addition, countries and airlines have pulled their planes from the skies.\n\nCanadian airline WestJet has also removed its Max models from its fleet, as did Air Canada and Lufthansa.\n\n\u201cThe decision to temporarily remove our Max 8 fleet from service was made as a precaution and is consistent with actions by regulators around the world, including the FAA,\u201d said WestJet in a statement.\n\n\u201cWhile we have faith in the safety and airworthiness of our Max 8 aircraft, this temporary action is as a precaution while we await the findings from both the FAA and Boeing\u2019s ongoing investigations into the cause of the Ethiopian Airlines Flight 302 crash.\u201d\n\nFeatured Image: Pixabay]" time="1.372"><properties><property name="score" value="0.3384327491124471" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Image Credit: Angie B., Albuquerque, NM\n\nPeople feel alone when they don't have a purpose. They feel like they are nothing when they don't have something to do. They feel insecure because they feel like their life doesn't have a meaning. Without these feelings people will feel like they are a nobody. They will feel like their life is worthless. Without a purpose to live for it will be hard for people to be happy and be themselves. When people have no purpose to life they can be so unhappy. These feelings come up in people when they are very upset. People are less likely to be happy when they are alone. With being alone it makes it harder to have fun. Without fun in your life it makes it harder to have a positive attitude. With a positive attitude people will be able to live without fear. They will be able to live like they are in a good mood. With having a good mood people will not be sad and they will feel good about themselves. People will also feel happy when they have a good life. When people feel good about their life it will help people have a positive attitude. They will also be able to have a better life. It will be easier to have a better life without bad feelings. When people have a bad life it will make it harder for them to be happy. This is because their life is not the way they want it. It is very hard for people to be happy when they don't have the life they want. Without having the life they want they will not be happy. People who have a positive attitude will not feel so sad when their life isn't the way they want it. With having a positive attitude people will be able to feel happy. When they are happy it will help them to live a better life. With a better life they will be happier. It will be hard for people to be happy when their life isn't the way they want it. This is because people feel like their life is going wrong. When people feel like their life is going wrong they are less likely to be happy. People who are happy are more likely to be happy. When people are happy they will not be sad. People will not be sad when they are happy. With having a good attitude it will help people be happy. It will also help people to have a good life. People who have a good attitude will feel good about themselves. They will feel good about their life and will have a good life. With a good life they will feel good about themselves. People who feel good about themselves will not feel so sad. This is because they have a good life. They also have good friends. Having good friends will help people be happy. With having good friends people will be happier. People will be happier with good friends. They will also be happier with a good life. With a good life and good friends people will feel happy. It will be hard for people to be happy without a good life. With a good life people will be happier. It will also be hard for people to be happy without good friends. Without good friends people will feel less happy. This is because they have no one to talk to. People with a good life and good friends will feel more happy. It will be easy for people to be happy with a good life and good friends.]" time="1.785"><properties><property name="score" value="0.9865332543849945" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Court rules out taking blood samples for cell phone surveillance\n\nIn the United States, the Supreme Court ruled out taking blood samples for cell phone surveillance, the second verdict in just a few days.\n\nThe Supreme Court in Washington (USA) ruled that the police is not allowed to search a person's body for a cell phone's location and information.\n\nIt was decided in a 4-to-3 vote. This is the second verdict in a week in which the Supreme Court ruled that blood samples must not be taken from a person arrested for driving under the influence of alcohol.\n\nAccording to the relevant law, the police can take a blood sample for alcohol from an arrested person if there are grounds for suspicion. It also specifies that if a person refuses, they can be forcibly taken.\n\nA man was arrested for driving under the influence of alcohol in 2011. He was ordered to give blood samples. He refused, but the police took the samples forcibly. The court found out that the defendant was under the influence of alcohol. He appealed, but his appeal was rejected in 2012. He then appealed to the Supreme Court.\n\nIn a 4-to-3 vote, the Supreme Court overturned the verdict and ordered the verdict to be re-tried.\n\nIn the meantime, the verdicts are difficult for police. They can't take blood samples from suspects in alcohol-related accidents, nor take blood samples of suspected drug addicts, nor use DNA to find out the suspects in rape cases. They have to be very careful about searching people on the street.\n\nOther websites also commented on the news.]" time="0.825"><properties><property name="score" value="0.99982088804245" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Another Apple Exec Talks About Bringing Down The Cost Of Macs\n\nOn the heels of some recent remarks by Apple\u2019s Phil Schiller about the company\u2019s view on entry-level pricing, the current vice president of worldwide marketing, Phil Schiller, was quoted today in a Wall Street Journal article about PC prices.\n\nSchiller says that for the sake of keeping PC prices down, the platform needs to avoid fragmentation by maintaining one version of Windows that\u2019s supported by all vendors. \u201cThat helps the consumer know that if I buy a PC in the next six months, it\u2019s going to run that software. That\u2019s good for consumers, and it helps keep prices down,\u201d he told the newspaper.\n\nThose remarks are similar to what Schiller had to say about Apple\u2019s Mac sales earlier this month, when he told analysts during a conference call that the company\u2019s philosophy is to not \u201cgive customers a lot of choices.\u201d\n\nHe also said that Apple has taken a different approach to pricing compared to its competitors. \u201cIt\u2019s like when you go into a restaurant, you know what the food is going to cost. You don\u2019t have to go through five steps of saying I want the steak for this much money, the chicken for this much money, the pasta for this much money.\u201d\n\n\u201cWe\u2019re not trying to segment or isolate ourselves from anybody,\u201d Schiller said. \u201cIt\u2019s more like, if you want a Mac, this is the price. If you want a PC, go over here.\u201d\n\n[Source]]" time="0.807"><properties><property name="score" value="0.8805960416793823" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[OK, this is all you need to know about this whole &quot;Ryan Fitzpatrick's against the Browns&quot; narrative that is now in full force, after his dismal performance against the Browns last year, and has been growing in fervor ever since news of his re-signing broke last week:\n\nIt's not going to happen again.\n\nYou may not believe it now, and it may not be true on Sunday when the Browns take on the Bucs in their Week 1 matchup. But just like the Browns have a way of winning the games they're supposed to, Fitzpatrick has a way of having performances like he did against the Browns last year.\n\nIt just happens in reverse.\n\nThe defense was ranked 30th in the league in pass defense and would give up more yards in the first quarter than in the entire game the year before, yet it's Fitzpatrick who throws for 5,367 yards and 40 touchdowns that year. That's not a typo. And it's not because he was the quarterback of the future and Josh Gordon was the No. 1 receiver of the future, either. (We'll ignore the fact that the Browns traded for Jason Campbell at the end of the season to play quarterback and traded away Trent Richardson for basically nothing.)\n\nFitzpatrick and the Browns were matched up against each other in the third week of the 2012 season, and by that point in the season Fitzpatrick had thrown for 4,008 yards and 27 touchdowns in just 12 games. He finished the season with an astronomical 8.18 yards per attempt, which was the highest in NFL history. (Since 2001, at least.) He averaged two touchdowns per game for the first three weeks of that season before throwing just five more in the final nine games.\n\nAnd that was in his first season with the Bucs. They acquired him on the last day of August, after he'd spent the last few months being cut and re-signed by the Bills, which was the latest example of the Bills treating him like a starting quarterback without actually starting him. When the Bucs traded for him, it was widely considered a sign that they'd be moving on from Josh Freeman, but Fitzpatrick's job was only to hold the fort. He didn't even start his first game until the Bucs' eighth game of the season.\n\nThis time around, Fitzpatrick is more than holding the fort. The Bucs have put him on the short list of best quarterbacks in the NFL, with his name appearing on every single MVP ballot last year, and after a few years of analysts treating him like a highly replaceable backup, he has emerged as a bonafide starter. And like most No. 1 quarterbacks, he also seems to thrive against certain teams.\n\nIt seems like a good rule of thumb to get out of the way when he's playing his best.\n\nThe Bucs will almost definitely start out slow against the Browns on Sunday, but they're not going to go through another slow start and blow a lead against the Browns for a second time. It just won't happen.\n\nThis is a different year, a different Browns team, and a different quarterback. It's also a different Ryan Fitzpatrick.]" time="1.481"><properties><property name="score" value="0.9998149275779724" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Get daily updates directly to your inbox Subscribe Thank you for subscribing We have more newsletters Show me See our privacy notice Could not subscribe, try again later Invalid Email\n\nSunderland AFC striker Steven Fletcher has had his \xa320,000 Bentley vandalised and stolen, sparking a massive police hunt.\n\nThe \xa345,000 luxury motor, which was parked up outside his home, was spotted being driven away by an offender, believed to be male, wearing a black crash helmet and driving a silver Ford Mondeo.\n\nPolice say it was not a random attack and is being treated as a targeted theft.\n\nThe car has the registration number CN55 ANT and Sunderland AFC can confirm the vehicle belongs to Fletcher.\n\nThe vehicle is a black Bentley Continental GT Speed with blacked out windows.\n\nA spokeswoman for Northumbria Police said: \u201cWe received a report of a car theft in Sunderland on Monday night.\n\n\u201cThe vehicle was parked outside a house and was seen being driven away by a man, who was wearing a black crash helmet.\n\n\u201cThe vehicle was a Bentley Continental GT Speed with blacked out windows.\n\n\u201cThe car had the registration number CN55 ANT.\n\n\u201cThis was not a random incident. It was a targeted theft.\n\n\u201cWe are keen to hear from anyone who has any information about this incident, or who has been offered a similar vehicle for sale.\n\n\u201cWe are particularly keen to hear from anyone who saw the vehicle between 9.30pm and 10pm on Monday night.\u201d\n\nThe theft of Fletcher\u2019s Bentley follows the theft of Sunderland\u2019s team bus in 2008.\n\nThe club\u2019s coach was stolen from outside the Stadium of Light on the evening of the final game of the season against Fulham.\n\nA club spokesman said at the time: \u201cWe\u2019re concerned and disappointed that someone could stoop so low as to steal the team bus and all of the club\u2019s transport plans.\n\n\u201cThis is an act that should shame football fans.\n\n\u201cWe would appeal to whoever stole the bus to think again and drop it back to us.\u201d\n\nAnyone with information is asked to contact Northumbria Police on 101 ext 69191.]" time="1.073"><properties><property name="score" value="0.999811589717865" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Another episode of Tetsuo: The Iron Man! It\u2019s getting there. I will be enjoying this episode a bit less than the previous few because the creators seem to be in a big rush to make the series come to an end. Well, at least they will finish the story. At any rate, enjoy the recap!\n\nRead the rest of this entry \u2192]" time="0.224"><properties><property name="score" value="0.9997960925102234" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Cross-posted at The Huffington Post.\n\nThe recent announcement that, for the first time in three decades, a new strain of avian influenza is spreading from one species to another is concerning but not, as some news outlets have claimed, evidence of a \u201cfatal pandemic in the making.\u201d A strain of H5N1 avian influenza (called H5N1-V) was detected in two apparently healthy wild ducks found in Germany. These ducks, according to the World Health Organization, most likely caught the virus from chickens, because H5N1-V has not previously been detected in wild birds, and H5N1 has not previously been found circulating among wild birds anywhere in the world. The two ducks apparently had not recently come into contact with any chickens. The H5N1-V strain found in the ducks is one of at least 14 genetically distinct H5N1 subtypes. While the H5N1-V strain can infect chickens, it cannot efficiently spread among them or to other species.\n\nH5N1 is an avian influenza virus, meaning that it naturally infects birds. Over the past 14 years, H5N1 has caused at least 555 infections and 282 deaths in humans. While those numbers are higher than any other human infection since influenza was first identified in humans in 1918, they are still a small fraction of the estimated 5 billion to 8 billion birds that have been infected with H5N1 over the past decade. The World Health Organization says that \u201cit is difficult to estimate accurately the number of cases that have gone undetected.\u201d Since the virus has spread to so many different countries, it is likely that many cases have gone undetected. Nevertheless, the case-fatality rate (defined as the percentage of people who die from a given disease) from H5N1 in humans is estimated at 60 percent. It is difficult to compare H5N1\u2019s case-fatality rate to those of other viruses. The case-fatality rate from the H1N1 influenza virus that caused the 1918 influenza pandemic, which killed an estimated 50 million to 100 million people worldwide, was only 2.5 percent, yet that virus killed about 2 percent of the world\u2019s population.\n\nGiven the high death toll, it is important to find the best ways to limit the spread of H5N]" time="1.276"><properties><property name="score" value="0.8305776715278625" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[As word spread of the new church\u2019s intentions, many Muslims came to have a look. \u201cThey were surprised to see Christians.\u201d\n\nKoberg said he was surprised to see the modest construction was built as quickly as it was.\n\n\u201cWe have enough of everything we need. We have enough food, we have shelter, and there is security. We\u2019ve built what we could,\u201d said Mohammed. \u201cThey are very good, very nice. We\u2019re working together.\u201d\n\nBut Muslim leaders in the area are concerned that the church is being built on a spot where a mosque stood for about 80 years.\n\nEarlier this year, tensions between Christians and Muslims flared up when a Catholic hospital in the nearby city of Bauchi shut down its maternity ward, arguing that the government did not give it enough money to operate. As Muslims felt that the ward\u2019s closing was unfair, they attacked the hospital, which is run by the Italian-based congregation of Daughters of Mary Immaculate.\n\nKoberg said he was surprised that no one has reacted to the construction of the church so far.\n\n\u201cNobody has said anything.\u201d\n\nThe Protestants in Maiduguri had been looking for a new place to worship for months. They are not the only Christians in town, but they are the only ones without a church. Now they can freely worship their God.]" time="0.819"><properties><property name="score" value="0.9700043499469757" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Bungie has confirmed a new competitive mode for Destiny 2, with a new ranking system and playlists.\n\nFollowing the release of the Curse of Osiris expansion, Destiny 2 will receive some new updates to the Crucible, as well as the addition of private matches for the first time. In a recent blog post on Bungie's website, it was confirmed that Crucible rank is returning to Destiny 2, and that it'll be more straightforward than it was before.\n\n&quot;We believe that to earn your highest rewards, you should be playing the most challenging activities possible for your skill level,&quot; the post reads. &quot;With that in mind, we\u2019re making the following changes to the player investment required to reach your highest rank:\n\n- Strike-specific rewards (Medallions, Boots, Class Items, etc.) will no longer be tied to your highest achieved rank\n\n- You will be required to reach a specific rank (HM5) to earn the Nightfall-specific rewards\n\n- You will be required to reach a specific rank (HM5) to earn the Trials of the Nine reward weapons\n\n&quot;In addition to the above changes, we\u2019re also increasing the overall player investment in the Crucible by requiring a minimum Light of 251 to participate in the most challenging endgame activities.&quot;\n\nIn the same post, Bungie has confirmed a new ranking system and matchmaking for the Crucible, which will include new playlists.\n\n&quot;To celebrate our launch of private matches and the new matchmaking system, we\u2019re implementing a new system for ranking up in the Crucible. The highest rank you can achieve in Competitive play is now \u201cLegend,\u201d and your Glory will be available to view on our new Competitive playlist HUD. Glory earned in the Competitive playlist will be used to rank up within this new system. We\u2019ll continue to expand on these improvements in the coming months.&quot;\n\nIt's interesting that the new rank system is being implemented for just the Competitive playlist, as well as the fact that your Glory is limited to a specific playlist. Hopefully, this doesn't mean there will be no more way to rank up in the Crucible outside of Competitive.\n\nAs part of the changes to the Crucible, Bungie is adding Private Matches to Destiny 2, which will be coming with the Curse of Osiris expansion in December.\n\n&quot;If you\u2019ve played our April Update, you\u2019ve had a taste of what a competitive, matchmade experience can be like in Destiny 2. Now, we\u2019re ready to give you the power to host your own private Crucible matches. This new feature will become available in the next update.&quot;\n\nPrivate matches can be played with two to twelve players, and will have matchmaking, as well as private lobbies. There will be a wide range of options available to players, with &quot;full playlist and customization support&quot;. Private matches will also come with their own rewards.\n\nPrivate matches will be available to all players who own Destiny 2 and have access to Curse of Osiris. You'll be able to make them as accessible as you like, whether it's making it a full-scale party affair, or something more intimate. Whatever your needs are, we\u2019re looking forward to you showing us what you\u2019re made of in your own custom arenas.&quot;\n\nDestiny 2 is available now on Xbox One and PlayStation 4.\n\nSource: Bungie]" time="1.828"><properties><property name="score" value="0.9439132809638977" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Approximate size 4.75\u201d x 3.5\u201d\n\nUse both sides\n\nHeavy-duty material for superior durability\n\nProtect your firearms and valuables\n\nMade in USA\n\nColor \u2013 Black\n\nUse both sides to take notes or hold your gun with a large holster. Black finish. Approximate size 4.75\u201d x 3.5\u201d. Heavy-duty material for superior durability. Made in USA.]" time="0.227"><properties><property name="score" value="0.9996210336685181" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[My wife is a big fan of using oils in her hair. So, I bought a couple different ones to try and compare. This is one of them. It smells very good and it makes her hair nice and soft. If you want a lot of shine, this one is not going to do it. But if you want a lot of softness, this is your oil.]" time="0.230"><properties><property name="score" value="0.9997840523719788" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Dear Reader, As you can imagine, more people are reading The Jerusalem Post than ever before. Nevertheless, traditional business models are no longer sustainable and high-quality publications, like ours, are being forced to look for new ways to keep going. Unlike many other news organizations, we have not put up a paywall. We want to keep our journalism open and accessible and be able to keep providing you with news and analyses from the frontlines of Israel, the Middle East and the Jewish World.\n\nThe proposal for a binational state on Israel\u2019s future borders with the Palestinians, put forth by outgoing Prime Minister Ehud Olmert, is no longer an acceptable solution, former Knesset speaker Avraham Burg said Wednesday.\n\n\n\nSpeaking in an interview with Channel 10, Burg said that the proposal, which the Olmert presented last week to Palestinian Authority President Mahmoud Abbas, was \u201cna\xefve and counterproductive.\u201d\n\n\n\n\n\nJPOST VIDEOS THAT MIGHT INTEREST YOU:\n\n\u201cIsraelis don\u2019t want to live in a binational state. One can\u2019t create a state of Jews and Arabs, and not even a state that is a nation of all its citizens,\u201d Burg said. \u201cI don\u2019t see any logic in it, even less so now than five years ago.\u201dThe former Kadima party leader argued that a similar situation exists in Ireland, with the Irish living alongside the English, but not in a single state.Similarly, he said, one could not build a future Palestinian state within the borders of Israel, but one can build a state of all its citizens in the West Bank, Gaza Strip and eastern Jerusalem.If the Israeli government doesn\u2019t agree to this, he warned, the two-state solution is no longer attainable.\u201cWhat we need is an understanding that [an agreement with the Palestinians] will be built on borders and not on a state of all its citizens. We\u2019ll build two states on the 1967 borders, one Jewish and one Palestinian,\u201d Burg said.\u201cIsraelis need to know that [in the future] there will be two states,\u201d he continued. \u201cThe Palestinian people needs to know that [in the future] there will be a Palestinian state.\u201dThe two-state solution is the only way to prevent the State of Israel from becoming the State of the Jewish people, the Knesset speaker warned. \u201cThe future of the State of Israel depends on the existence of two states,\u201d he said.\u201cI\u2019m for a binational state only as a stage, not as a solution,\u201d he added. \u201cIn a binational state, we will never live together, but there will be an agreement. If there\u2019s no agreement, there will be a civil war. We have to prepare for a binational state as a transition stage to the two-state solution.\u201dWhen asked about the present leadership\u2019s decision to allow MK Otniel Schneller (Kadima) to bring a vote of no-confidence in Defense Minister Ehud Barak, Burg accused Barak of \u201centhusiastic nationalism\u201d and of being \u201cna\xefve.\u201d\n\nJoin Jerusalem Post Premium Plus now for just $5 and upgrade your experience with an ads-free website and exclusive content. Click here&gt;&gt;\n\n\n\n]" time="1.626"><properties><property name="score" value="0.9066943526268005" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[JOYNER LUCAS \u2013 JOYNER\u2019S WORLD \u2013 PREMIERE\n\n#JOYNERSWORLD PREMIERE:\n\n\n\nI\u2019m telling you right now, JOYNER LUCAS has been fucking with his album so hard. That\u2019s right, his album. What I mean is that we know about his movement, but his album really stands for something. His \u201cJoyner\u2019s World\u201d is about the transition to adulthood. It\u2019s about being a young artist making waves on the charts and the fan support he\u2019s getting. The song \u201cI\u2019m Not Racist\u201d really gives off the vibe of Joyner\u2019s]" time="0.428"><properties><property name="score" value="0.9998251795768738" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Yet another reason why we love Mary Tyler Moore: The Golden Girl will donate the proceeds of her upcoming appearance on Dancing With the Stars to St. Jude Children's Research Hospital.\n\nMoore's reps confirmed that the donation will come from the profits of Moore's Dancing With the Stars' appearance, in which she and professional dance partner Tristan MacManus perform a cha-cha to Donna Summer's &quot;Hot Stuff.&quot;\n\nWatch Mary Tyler Moore's 'Hot Stuff' Dancing With the Stars Performance\n\n&quot;Mary Tyler Moore and her team have been supportive of St. Jude Children's Research Hospital for many years and she will be donating a portion of her proceeds from Dancing with the Stars to support St. Jude,&quot; the charity said in a statement to E! News. &quot;Her support is critical to the hospital and we are grateful for her help.&quot;\n\nMoore is competing on the show's 12th season with former General Hospital star Ingo Rademacher. Last night, Moore and Rademacher scored the first 10 of the season from judge Len Goodman.\n\nWhile Moore will donate her Dancing With the Stars appearance, some celebrities don't -- even though that's what's requested. They still get a share of the show's profits from voting. Last year, Katie Holmes said she wouldn't do Dancing With the Stars but she still did. She did donate the appearance fee she would have received from the show to the Children's Hospital of Los Angeles.\n\nIn 2010, Jane Seymour donated the proceeds of her Dancing With the Stars appearance to the Cedars-Sinai Medical Center. Also in 2010, season 11 winner Donny Osmond donated his Dancing With the Stars appearance fee to Children's Miracle Network Hospitals.\n\nDancing With the Stars airs Mondays at 8/7c on ABC.]" time="0.916"><properties><property name="score" value="0.046796441078186035" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[PHILADELPHIA (CBS) \u2013 The 76ers fell to 5-12 overall and 0-5 at home with a 99-94 loss to the Milwaukee Bucks.\n\nThe Bucks outscored the Sixers 22-16 in the fourth quarter to overcome an 18 point deficit.\n\nHollis Thompson led the Sixers with a season high 21 points off the bench.\n\n\u201cI\u2019m happy for him. I think he\u2019s earned the opportunity to get more minutes. I thought he was a pretty complete player tonight, he guarded, he rebounded. I think he\u2019s taking the first step,\u201d Sixers coach Brett Brown said.\n\nThe Sixers held a seven point lead after one quarter but couldn\u2019t hold the momentum.\n\n\u201cWe had such an amazing first quarter. We really put a dent in their defensive scheme,\u201d Brown said.\n\nBucks coach Jason Kidd, who played with the Sixers for a short time, on a night where his team set a franchise record with 14 three pointers made.\n\n\u201cWe were missing a lot early and then the last quarter we really played well and got it going. We were able to put some big points up and get some stops when we needed it,\u201d Kidd said.\n\nMichael Carter-Williams led the Sixers with 19 points and six assists.]" time="0.668"><properties><property name="score" value="0.8974822461605072" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[...Rather than simple self-help, this book represents what it should be - a collaborative work between the writer and his/her therapist to ensure maximum understanding and success. - Robert M. Corry, M.D., in his review in Psychotherapy\n\n\u2026Clear, practical, and insightful. - Frederick Chapman, M.D., Harvard Medical School\n\nDavid Cohen, M.D. has written a manual on depression that will help many who suffer with this debilitating condition to recognize that there is hope for recovery. More importantly, his description of effective treatments available today is evidence of the tremendous advances in this field. He has done a wonderful job in providing the reader with easy to understand information about depression and how to overcome it. - Daniel J. Vitiello, M.D.\n\nEditorial Board Member, Journal of Affective Disorders\n\n...Presents a wealth of ideas that are written in an easy to read and accessible style. - Library Journal\n\n...From description to treatment, there is something in this book for all readers. - John Guilford, M.D., Clinical Professor of Psychiatry and Behavioral Sciences, Duke University Medical Center, Durham, NC\n\n...Dr. Cohen's book is the first on the market to deal in detail with the nonmedical therapies available for the treatment of depression and to be accessible to both patients and their families. It is a valuable contribution to the literature and deserves to be widely read. - E.E. Riley, Ph.D., in his Foreword\n\n...Cohen discusses in simple terms the pathophysiology of depression, the cognitive/behavioral theories of depression, the treatment of depression, and provides information about organizations that may be able to help. The author's descriptions of medications, both antidepressants and anti-anxiety medications, are accurate and detailed. The reader can feel confident that the medication section is accurate, without the endless descriptions found in other books. The key point to remember is that the medications are not the &quot;cure&quot; for depression. Instead, they are used to provide a way of breaking through a &quot;state of learned helplessness&quot; and providing a method to help the patient engage in a treatment program. This is the key point in the book. In order to help those suffering with depression, Dr. Cohen believes that the medical profession has a responsibility to refer patients to a &quot;team&quot; which will help them get over their depression and learn to cope with the problem. This book will give the layperson a much-needed introduction to the field of depression, the various treatments, and how they work. It is written in an easy-to-understand language and provides references for the reader to pursue if the reader wishes to know more about the topic. - Richard H. Price, Ph.D., Past President, Academy of Cognitive Therapy\n\n...All those who are dealing with depression will appreciate this practical guide. It is written in a clear, direct manner. Those who are serious about becoming well will find in Dr. Cohen a knowledgeable and empathic guide. - Rev. Marybeth Twichell, Ph.D.\n\nThe excerpt from the Foreword, &quot;The Practical Guide for Persons With Depression,&quot; by Dr. E.E. Riley is reprinted with permission of the author. The excerpt from the Foreword, &quot;A Friend to Help Along the Way,&quot; by Dr. Daniel J. Vitiello is reprinted with permission of the author. The excerpt from the Foreword, &quot;The Working Alliance,&quot; by Dr. Robert M. Corry is reprinted with permission of the author. The excerpt from the Foreword, &quot;A Different Perspective on Depression,&quot; by Dr. John Guilford is reprinted with permission of the author. The excerpt from the Foreword, &quot;Keeping the Ball Rolling,&quot; by Dr. Frederick Chapman is reprinted with permission of the author. The excerpt from the Foreword, &quot;A Review of Common Medications,&quot; by Dr. Richard H. Price is reprinted with permission of the author. The excerpt from the Foreword, &quot;Excerpts from 'Depression - What Every Clinician Should Know',&quot; by Rev. Marybeth Twichell is reprinted with permission of the author. The excerpt from the Foreword, &quot;Fantastic Resource for Persons Who are Coping with Depression,&quot; by Dr. Daniel J. Vitiello is reprinted with permission of the author. The excerpt from the Foreword, &quot;The Treatment of Depression,&quot; by Rev. Marybeth Twichell is reprinted with permission of the author. The excerpt from the Foreword, &quot;Dr. Cohen's Story of Success,&quot; by Dr. E.E. Riley is reprinted with permission of the author. The excerpt from the Foreword, &quot;Respecting Depression and its Sufferers,&quot; by Dr. Robert M. Corry is reprinted with permission of the author. The excerpt from the Foreword, &quot;A New Paradigm of Depression,&quot; by Dr. John Guilford is reprinted with permission of the author. The excerpt from the Foreword, &quot;Facing Depression Together,&quot; by Dr. Frederick Chapman is reprinted with permission of the author. The excerpt from the Foreword, &quot;Depression: The Last Great Challenge,&quot; by Dr. Richard H. Price is reprinted with permission of the author.]" time="2.680"><properties><property name="score" value="0.9217539668083191" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[City workers fill up sandbags for residents after torrential rains on January 21, 2019 in Biloxi, Mississippi. Getty Images\n\nThe storms killed at least six people across the U.S. Gulf Coast.\n\nTropical Storm Barry formed on Thursday evening, and at 11 p.m. ET it was a tropical depression, according to the National Hurricane Center.\n\nBut it's expected to regain strength as it continues to move across the warm waters of the Gulf of Mexico.\n\nThe storm, which has maximum sustained winds of 45 mph, is expected to dump anywhere between 4 to 8 inches of rain across Louisiana and neighboring states.\n\nThe National Hurricane Center issued a hurricane warning for southeastern Louisiana and coastal Mississippi as the storm nears land.\n\nLouisiana Governor John Bel Edwards declared a state of emergency on Thursday.\n\n&quot;We will do all that we can to help people prepare for this storm,&quot; Edwards said at a news conference.\n\nHe added: &quot;I urge all Louisianians to remain vigilant and to stay informed on weather conditions and possible storm impacts through the entire weekend.&quot;\n\nEdwards warned residents to be prepared for potential road flooding and power outages.\n\nThe Weather Channel reported that flooding from the storm is a concern for a number of towns along the Louisiana coastline, including Lafayette and Morgan City.\n\nThe NHC says that &quot;life-threatening storm surge is likely along portions of the Mississippi and southeastern Louisiana coastlines&quot; as the storm nears landfall on Saturday night.\n\nWeather models have]" time="0.765"><properties><property name="score" value="0.9991845786571503" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[#1 Maintaining momentum. While this might be my blog, I know you guys have a lot going on. We all do. That\u2019s why it\u2019s important to make sure you\u2019re building on what you\u2019ve started and staying on track. It\u2019s so easy to get derailed. The most important thing is to be consistent and you\u2019ll find that the results you want will come naturally.\n\nIf you enjoyed the article please share it on social media. I am the author of 4 books, all available on Amazon.com or Barnesandnoble.com. For more information, please visit www.wright-well.com.]" time="0.335"><properties><property name="score" value="0.9990864992141724" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Surging.\n\nI was a fan of Steppenwolf's Magic Carpet Ride when I was younger, and my tastes have not changed.\n\nI have been listening to Steppenwolf non-stop.\n\nBut this is not a tribute band, they are completely on their own turf.\n\nThere is a burning wildness in the vocals, an urgency in the songs, that you don't see in tribute bands.\n\nAnd the lead guitar is stellar. I don't even have to describe it because the sound is unmistakable, like the perfect mix of Carlos Santana and David Gilmour.\n\nThe group members hail from all over the planet: Stacey Blades is from New York City, Marco Panascia from Brazil, Marco Alfonso, of Caracas, Venezuela, and Juan Pablo Torres from Guayaquil, Ecuador.\n\nTheir bandcamp page gives you a good overview of their sound:\n\nhttps://steppenwolfband.bandcamp.com/]" time="0.479"><properties><property name="score" value="0.999795138835907" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[As global oil prices continue to rise, some observers are warning that rising gas prices may cause a major economic downturn in 2008. What's more, the recent natural gas pipeline explosion in California, which prompted a moratorium on new development, will exacerbate the problem.\n\nOver the next five years, the United States will have to import nearly half of the energy it uses, according to the most recent forecast by the U.S. Energy Information Administration. As of now, the United States is the largest energy importer in the world, importing more energy than any other country. We import 3.5 million barrels of oil per day, and more than 80 percent of our oil comes from foreign countries. This puts us at the mercy of OPEC nations, and increasingly Russia, who we are now trying to court to invest in our own energy infrastructure.\n\nThis dependence is a matter of great concern for many energy analysts. Some say that it's time to cut our losses and invest in the most energy-efficient technology possible while we still can.\n\nOur continued dependence on foreign oil could have a long-term impact on the economy, they say, because the price of oil will continue to rise with the demand, and the price of energy could skyrocket in the next decade.\n\n'We Need to Start Thinking Now'\n\n&quot;Energy is not just a topic that's come up in the last few months,&quot; says Matthew Ranson, an analyst with the American Council for an Energy-Efficient Economy. &quot;It's been discussed in a serious way for decades now.&quot;\n\nAccording to Ranson, for a long time, people weren't really paying attention to energy issues in this country. But then, he says, &quot;the economy collapsed in the last year or so and the [energy] debate started up again in full force.&quot;\n\nBut even now, he says, &quot;we need to start thinking about it now. We can't just wait.&quot;\n\nThe economy will be hurt most by high energy prices, says Ranson. &quot;Energy, unlike other goods and services, is necessary for people to have jobs and make a living and grow the economy,&quot; he says.\n\nThe crisis is already in the works. &quot;Gas prices are expected to increase at the rate of $3 a year,&quot; Ranson says. &quot;At some point, it's going to make people reconsider car ownership, which we are seeing already.&quot;\n\nOther analysts have noted that the worst economic effects will come in the next few years as oil prices begin to hit consumers' wallets.\n\n&quot;It's really important for people to understand that we don't have to sit around and wait for it to hit us,&quot; Ranson says. &quot;We can do things now to change the way we make energy choices.&quot;\n\nThe United States needs to cut back on our reliance on oil, he says, and start developing clean, domestic energy resources like solar and wind power.\n\n&quot;We need to think differently about the choices that we're making,&quot; he says. &quot;If we have a choice to import oil from places that don't like us, or if we have a choice to invest in clean domestic energy that's good for us, we need to make the choice.&quot;\n\nThe Golden Years\n\nRanson admits that it's not going to be easy to move toward a clean energy economy. But there are things that can be done, such as improving efficiency and investing in new energy technologies. &quot;We can't do everything at once,&quot; he says.\n\nTo help reduce energy demand, Congress has passed a law to require the federal government to improve energy efficiency in its buildings.\n\nThere are also efforts under way to address the energy crisis in cities, where people can be more dependent on energy, according to Ranson.\n\n&quot;A lot of cities are thinking about this in different ways,&quot; he says. &quot;In New York, they're talking about ways to reduce car ownership, like getting people out of cars and into public transportation, making public transportation more convenient, and adding bike paths. You can start with that and see what happens.&quot;\n\nThe Natural Gas Supply\n\nAnother important energy issue to keep an eye on is the country's natural gas supply. Natural gas accounts for nearly 20 percent of the country's energy, and nearly 60 percent of the nation's electricity generation.\n\n&quot;We don't really have a clear understanding of how the supply of natural gas will be affected by what's going on in California,&quot; says Fred Millar, senior fellow at the Washington-based think tank, Energy Security Initiative.\n\nMillar believes that the California situation may signal &quot;troubles ahead&quot; for the country's natural gas supply. &quot;There are real questions about how much natural gas is going to be produced by the industry in the coming years,&quot; he says.\n\nThere's been a dramatic drop in natural gas prices in the last few years, and that has encouraged a lot of companies to start new drilling operations, says Millar. But natural gas prices were going up last week, Millar says, and &quot;the question now is, has the bubble burst?&quot;\n\nA recent explosion at a Southern California natural gas pipeline, which took the lives of two workers, has put a halt to natural gas production in the state. The pipeline was owned by the Texas-based company Sempra Energy, and the blast occurred near Bakersfield. It is still unclear what caused the explosion.\n\nThe blast caused a temporary halt to the development of the Puente Valley Pipeline, a major new pipeline project.\n\nMillar says that the halt on the pipeline will have a major effect on the state's energy supplies. &quot;It could cause a major problem for California's electricity supplies,&quot; he says.\n\nMillar adds that the price of natural gas is likely to rise in the coming weeks and months because of the halt.\n\nOil and the Economy\n\nHigh energy prices are also a major concern for economists. &quot;The most immediate effect of a large price increase is inflation,&quot; says David Hughes, a geologist who has written extensively on energy issues. &quot;We can expect energy prices to increase for gasoline, electricity, natural gas, and other fossil fuels.&quot;\n\nBut economists are still debating how the economy will be affected. Some economists say that higher energy prices could have a positive impact on the economy, while others say the impact could be negative.\n\n&quot;It's too early to tell,&quot; says Millar. &quot;We're still in the stage where the price increases are in the process of playing out.&quot;\n\n&quot;The economy is more fragile than it was in the 1970s, and energy is more integrated into the economy, and prices are higher,&quot; says Ranson.\n\nBut he believes that the economy is better positioned to deal with energy prices now than it was in the 1970s. &quot;I don't think the impact of high energy prices will be nearly as bad this time,&quot; he says.\n\nA slowdown in the economy and a possible recession, Ranson says, could cause energy prices to spike even more.\n\nBut even if the economy takes a downturn, Ranson says that energy prices will keep going up. &quot;People may stop spending on some things,&quot; he says, &quot;but energy prices are going to keep going up.&quot;\n\nRanson says that people will have to make choices about how they live, and the government will have to decide how much of a subsidy it's willing to provide for the alternative energy industry.\n\nHughes believes that, ultimately, high energy prices will bring about some kind of economic crisis. &quot;In my view, this is a symptom of a much bigger problem, which is our oil-based economy,&quot; he says. &quot;We can no longer afford the fossil fuel based economy we have had for the past century.&quot;\n\nCopyright 2007 E&amp;E Publishing. All Rights Reserved.]" time="3.843"><properties><property name="score" value="0.9985998272895813" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[On the first play from scrimmage, Washington quarterback Jake Browning went deep for Dante Pettis, who picked up 50 yards on the play, but then injured his knee. Browning threw a touchdown pass to Chico McClatcher on the next play, but Browning later came out of the game and it was announced that he had a right leg injury. No further details were given on his condition, but his return]" time="0.254"><properties><property name="score" value="0.9979606866836548" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[I\u2019m back in Shoreditch and work starts again tomorrow so I\u2019m trying to get things sorted before that. However, this was my second big trip of the year and I\u2019m feeling a bit worn out. It feels like it was longer than a week though.\n\nI went to Seoul with Jess for two weeks. We flew to Seoul Incheon on Cathay Pacific and stayed at the Hyatt in Seoul Station, just outside the hotel station (it was actually our last day before we flew out). We visited the Han River and the Hongdae area (including bars, restaurants and music venues) and went to the Seoul Film Festival. We spent a lot of time walking around too as we don\u2019t really know the city that well and were pretty tired by the end of each day.\n\nWe visited a temple near Cheongdam called Angukdaegyo which was beautiful. We also visited the grounds of the 63 Building which is quite close to the hotel. The following day, we went to Jogyesa Temple which is where the head of the Jogye Order of Korean Buddhism has his residence. We also went to Cheongwadae which is the residence of the President of Korea and the Blue House. I really wanted to go inside the Blue House but it was closed when we were there. The area around the Blue House was really beautiful.\n\nWe had a really nice lunch at the Daehangno area. We had a really nice lunch at the Dongmyo Station near Gyeongbokgung Palace. The area is quite nice. We then went to Insadong where there were lots of small shops. We also visited Namsangol Hanok Village and had a nice lunch at the Hanok Village. It was also near Gyeonghuigung. We walked up the hills near Gyeonghuigung, which were quite steep and also quite a nice walk. We also visited Seoul Tower which was lovely and there was a great view of Seoul.\n\nWe then travelled to Busan for a day to see the Ulsan Bijudong Cable Car and to take the ferry to Geoje Island. Busan was nice and I can definitely see why people live there. There was a great view of the coast from the cable car. There were quite a lot of islands in Geoje Island. We stayed in the Geoje Island Ferry Cruise which is a small hotel. It was very pleasant there and the food was good. We walked around the island and saw the Minho lighthouse. It was quite cloudy so we couldn\u2019t see the islands around. We stayed for dinner and saw the sunset before we returned to Busan.\n\nOur next destination was Jeju Island. We flew from Busan to Jeju Island on a flight that took about 1 hour. There was also a harbour just outside the hotel. We stayed at the Full Moon Hotel. It was quite beautiful. We visited Hallasan Mountain and the Jeju Haenyeo Museum of Women\u2019s Contributions to Life in the Jeju Island in which there was a lovely view of Hallasan Mountain. We also visited Samcheok Beach. We had a great dinner at the Full Moon Hotel in the Samcheok Bay. There was a beautiful view of the moon rising over the beach from our room.\n\nThe next day, we went to Seongsan Ilchulbong Peak, which was very beautiful. It is a volcanic crater on a small island that was completely formed from lava. It was cloudy and rainy while we were there. We went to Seogwipo and had a lovely lunch at Jin Jin Gop. We had a really nice seafood lunch and it was just a couple of minutes away from Seogwipo station. We went to the Seogwipo Lotte Shopping Center and bought some gifts.\n\nWe then went to Hallim Park. It was a very hot day but it was very pleasant there. We went on a cable car to Mount Baekhwa, which was really nice. It is the highest mountain on Jeju Island and the views were nice. We also went to Mokcheongyeon cave and we visited the memorial museum and the huge cave. The cave was really nice and it was a good experience. We then went to Jeju Olle Trail which is a series of pathways that cross Jeju Island. They have been designed to showcase the most beautiful spots on Jeju. We went to the Jeju Olle trail which was near the entrance of Hallim Park. The weather was really nice so it was pleasant to be outside.\n\nWe went to Seongsan Sunrise Peak which was a nice place to see the sunrise. There was a really nice sunset too. The weather was very nice and it was a pleasant place to be. We also went to Mt Sine, which was very pretty. It was a bit foggy and we couldn\u2019t see Mt Halla but it was a nice walk. We also went to the 7 treasures of Jeju which was nice. We then went to Biyang-ri, Seogwipo, which was a bit more rural and very nice.\n\nOur last destination was the Jeju World Cup Stadium. It was a really nice stadium with lots of seating. It had lots of places to eat and drink too.\n\nWe then flew back to Seoul to fly home. It was a lovely place and I can definitely see why people would want to live there.]" time="2.719"><properties><property name="score" value="0.6026448249816895" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[In her interview with Popnography, Kate Micucci discussed her two new projects, BoJack Horseman and another animated series on Yahoo! Screen called Other Space. Other Space will be an anthology comedy following the crew of the space shuttle Venturi 2, created by Paul Feig.\n\nOther Space will be the first show I do where the cast is mostly male. I\u2019m a big feminist and I\u2019m not sure how I feel about that, but my character, Natalie, is the only girl on the ship and the youngest, which makes her feel isolated from the rest of the crew. She\u2019s shy and reserved. She\u2019s a huge sci-fi nerd and very book smart but not street smart. She\u2019s obsessed with books and fantasy but there are too many people on the ship and she has a hard time focusing.\n\nKate plays one of the main characters, Natalie, a woman with aspirations of space travel who works on the shuttle, Venturi 2. The series has been compared to an early 90s show called Space Cases, which also takes place on a space shuttle, however, it seems that Other Space will be a little bit more adult and will likely also delve into the science fiction genre as well as the comedy. Micucci will be joined by a great ensemble of actors including The Hunger Games actor, Sam Richardson, as Commander Glen. Richardson and Micucci previously worked together on Saturday Night Live, and Kate shared how much she loves working with him.\n\nI love working with Sam. He\u2019s so wonderful. He\u2019s so cute. I think my character Natalie really brings him out of his shell. She\u2019s the only one who really has the courage to push him. I think she also brings out a really kind, sensitive side in him that he probably doesn\u2019t have a lot of. It\u2019s been a really lovely experience to get to work with him. He\u2019s just a very warm, kind person.\n\nAlso joining the cast are Rebecca Romijn as the ambitious first officer, Kelly, and Enrico Colantoni as Captain Zalian. Other Space will be directed by Joe Russo and will be a half hour comedy show with ten episodes. The show is set to premiere on Yahoo! Screen on April 14, 2015.]" time="1.177"><properties><property name="score" value="0.0826774537563324" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The 24th annual World Press Photo Contest has announced its winners. One of the most prestigious and respected contests for photography, winners were selected from 82,951 images submitted by 5,775 photographers from 124 countries. Below are some of the winners in the photojournalism category. All captions were written by the individual photographers and are published as they were received.]" time="0.219"><properties><property name="score" value="0.9680722951889038" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Scoop First, an economic-development nonprofit organization based in Mobile, will present \u201cShoals Business Snapshots\u201d at 11 a.m. March 28 at the Crowne Plaza Hotel, 4880 Montlimar Drive.\n\nThe program features short presentations about businesses throughout Mobile and Baldwin counties, including businesses, schools and nonprofit organizations.\n\nThe event is free and open to the public.\n\nA light lunch will be provided.]" time="0.223"><properties><property name="score" value="0.7829578518867493" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.78295785&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.78295785
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Home &gt; Table Lamp Base &gt; Pearl / Crystal &gt; Round Crystal Table Lamp Base\n\nRound Crystal Table Lamp Base Code: 316 Price: US$ 20.00 New from: US$ 20.00 30 day money back guarantee\n\nMade to order, usually ships within 10 to 14 days Quantity in Basket: None Round Crystal Table Lamp Base\n\n\n\nA round base with polished silver or gold finish and a bright clear crystal knob that adds a sparkle to any table lamp.\n\n\n\nTo order:\n\nEnter the number of bases you wish to order into the box above the quantity selection and press the 'Add to Cart' button.\n\n\n\nFor quantities of less than 12 we will invoice you on receipt of your order and ask for payment to be made before your order is sent.\n\nThere are no products to list in this category.]" time="0.384"><properties><property name="score" value="0.9628957509994507" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Post by waltzmania \xbb Sun Nov 24, 2017 7:25 pm\n\nJudging by what you're hearing, I'm not sure the recording was ever mixed properly. My best guess is that it was mixed for the wrong kind of playback device. The first movement of the symphony is long and in four sections. If there are abrupt changes in sound quality, it is almost certain that]" time="0.215"><properties><property name="score" value="0.9717992544174194" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Der Pr\xe4sident des Verfassungsschutzes, Hans-Georg Maa\xdfen, ist von Bundeskanzlerin Angela Merkel (CDU) entlassen worden. Wie das Bundeskabinett am Mittwoch nach Teilnehmerangaben beschloss, wird Maa\xdfen zum Innen-Staatssekret\xe4r im Bundesinnenministerium bef\xf6rdert. Seine bisherige Stellvertreterin, Pr\xe4sidentin des Bundesamtes f\xfcr Verfassungsschutz (BfV), Susann Wimmer, wird dann seinen Posten als Beh\xf6rdenleiter \xfcbernehmen. Wie die &quot;S\xfcddeutsche Zeitung&quot; berichtet, haben die drei Parteichefs der Gro\xdfen Koalition, Angela Merkel (CDU), Horst Seehofer (CSU) und Andrea Nahles (SPD), die Personalie Maa\xdfen am Dienstagabend einstimmig beschlossen. Merkel habe demnach in einem &quot;Auftragskommando&quot; an die drei Parteichefs auf die schnelle L\xf6sung gedrungen. Im Laufe des Tages soll der Bundestag Maa\xdfen noch abschlie\xdfend \xfcber die Bef\xf6rderung abstimmen.\n\nMaa\xdfen soll dann als Staatssekret\xe4r f\xfcr europ\xe4ische und internationale Aussagen im Bundesinnenministerium eingesetzt werden, wie es am Mittwoch aus Regierungskreisen hie\xdf. Demnach wird er Staatssekret\xe4r f\xfcr die Bereiche Bauen, Wohnen und Energie.\n\nMerkel hatte am Dienstag erkl\xe4rt, die Koalition wolle den Fall Maa\xdfen noch heute abschlie\xdfend kl\xe4ren. Nachdem der Verfassungsschutzpr\xe4sident angeblich Zweifel an den Aussagen des j\xfcdischen Mannes ge\xe4u\xdfert hatte, der in Chemnitz gefilmt worden war, hatte der &quot;Spiegel&quot; berichtet, Maa\xdfen habe die AfD-Spitze in Sachsen gewarnt, eine Einbindung des Verfassungsschutzes in die Aufarbeitung des Chemnitzer T\xf6tungsdeliktes anzustreben. Zudem soll Maa\xdfen im Zusammenhang mit den Vorf\xe4llen in Chemnitz gegen\xfcber der &quot;Bild&quot;-Zeitung gesagt haben, dass ein Video von dem \xdcbergriff auf einen Mann &quot;gef\xe4lscht&quot; sein k\xf6nnte. Die Generalstaatsanwaltschaft Dresden bezeichnete diese Aussagen zu einem angeblichen Fake-Video als &quot;irref\xfchrend&quot;.\n\nAm Dienstag hatte der Fall Maa\xdfen bereits f\xfcr Unruhe in der Bundesregierung gesorgt, nachdem das Kabinett offenbar \xfcberraschend \xfcber die Zukunft des Verfassungsschutzpr\xe4sidenten entschieden hatte. Nach Informationen von NDR, WDR und &quot;S\xfcddeutscher Zeitung&quot; sollen SPD-Chefin Nahles und CSU-Chef Seehofer die Entscheidung der Kanzlerin zum sofortigen Rauswurf Maa\xdfens umgehend blockiert haben. Laut Informationen der &quot;S\xfcddeutschen Zeitung&quot; hat Merkel den Koalitionspartnern gegen\xfcber zudem ihren Willen bekr\xe4ftigt, Maa\xdfen als Staatssekret\xe4r in ihrem Kabinett zu halten. Merkel sei ver\xe4rgert gewesen \xfcber die Weigerung der SPD-Parteispitze, den konservativen Kritiker der Migrationspolitik als Staatssekret\xe4r im Innenministerium zu akzeptieren, schreibt die Zeitung unter Berufung auf SPD-Kreise.\n\nMaas: &quot;Irrelevante Debatte&quot;\n\nVor dem Hintergrund der Debatte um Maa\xdfen hat Au\xdfenminister Heiko Maas (SPD) die Bundesregierung aufgefordert, eine &quot;tiefgreifende und un\xfcbersehbare Antwort auf die Herausforderungen der Migration und die Probleme mit der Integration&quot; zu geben. &quot;Es geht nicht um eine Personaldiskussion, sondern um das Schicksal der Demokratie&quot;, sagte Maas am Mittwoch in Berlin. Die Debatte \xfcber Maa\xdfen sei nicht nur eine rein parteipolitische, sondern vor allem eine &quot;irrelevante&quot;.\n\nDas Verh\xe4ltnis der gro\xdfen Koalition zueinander sei &quot;durch die Vorg\xe4nge in der letzten Woche in keiner Weise verbessert&quot;, sagte Maas. &quot;Was wir in den letzten Tagen erlebt haben, ist ein erschreckender Mangel an Selbstverst\xe4ndnis, der manchmal nur noch unterstreicht, wie wenig sich die gro\xdfe Koalition selbst wertsch\xe4tzt.&quot; Dass es &quot;in den vergangenen Wochen&quot; dar\xfcber diskutiert worden sei, &quot;ob sich Maa\xdfen noch halten l\xe4sst&quot;, k\xf6nne niemanden erfreuen. Die Entscheidung, Maa\xdfen entlassen zu wollen, sei dar\xfcber hinaus von Personaldebatten \xfcberlagert worden.\n\nAufgeregte Stimmen\n\nBei der Debatte zum Haushalt des Bundesinnenministeriums im Bundestag wurde im Anschluss an die Haushaltsrede von Minister Horst Seehofer (CSU) die Stimmung in den Reihen der Regierungsparteien am Mittwoch sehr aufgeregt. Die Opposition im Bundestag sieht in den Spannungen der Koalition \xfcber die Zukunft des Verfassungsschutzpr\xe4sidenten Hans-Georg Maa\xdfen auch ein eigenes Versagen beim Streit.\n\n&quot;F\xfcr das, was hier gerade passiert, tragen Sie alle, die Sie hier sitzen, die politische Verantwortung&quot;, sagte die SPD-Fraktionsvorsitzende Andrea Nahles. &quot;Sie tragen daf\xfcr die Verantwortung, dass es in dieser Woche \xfcberhaupt zu einem solchen Haushaltsgesetz kommen kann, wie Sie es hier machen.&quot; Nahles forderte Seehofer auf, die Vorg\xe4nge \xfcber Maa\xdfen zu erkl\xe4ren, weil sie sich selbst von der Regierung verraten f\xfchle. &quot;Die Regierung und die politische F\xfchrung dieses Landes muss jetzt den Weg der Einheit beschreiten&quot;, forderte sie.\n\nKritik \xfcbte Nahles daran, dass das Bundesinnenministerium bei der Einrichtung von Erstaufnahmeeinrichtungen in Anforderungen an die jeweiligen Bundesl\xe4nder nun &quot;weich&quot; geworden sei. Zudem gebe es zu wenig Pl\xe4tze f\xfcr die Abschiebung abgelehnter Asylbewerber und zu wenig R\xfcckf\xfchrungen nach Afghanistan. Es sei ein &quot;Skandal&quot;, dass so viele Menschen in Deutschland festsitzen.\n\nAfD: Maa\xdfen soll Verfassungsschutz-Chef bleiben\n\nDie AfD h\xe4lt dagegen. Fraktionschef Alexander Gauland begr\xfc\xdfte Maa\xdfen als neuen Staatssekret\xe4r im Innenministerium. Er sei \xfcberzeugt, dass Maa\xdfen in der neuen Funktion seine Erfahrungen als Pr\xe4sident des Verfassungsschutzes in die Politik einbringen k\xf6nne, sagte]" time="5.564"><properties><property name="score" value="0.997104823589325" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99710482&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99710482
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Establishing a GitHub Account\n\nLog into your UCSC Gmail account. If you need to create a new UCSC Gmail account, visit this webpage. Once you have logged into your UCSC Gmail account, you can begin the process of creating a GitHub account. Please use your UCSC username as your GitHub username.\n\nTo create a GitHub account:\n\nIn the upper-right corner of the Gmail interface, click the menu icon, then click Settings. On the left side of the settings window, click Accounts and Import. In the lower-left corner of the Accounts and Import window, click Create an account. In the first field, enter your GitHub username. This is your UCSC username. In the second field, enter a GitHub password. This will be your username@github.com email address. In the third field, enter a password for your GitHub account. Please remember to remember this password. This password is only to access your GitHub account from the web. To access your account on your local machine, you will still need to use SSH. In the last field, enter your email address. This will be the email address your collaborators will use to send you commits and other notifications. Click the Sign Up button.\n\nYour UCSC GitHub account will be created. To access your new GitHub account, simply visit GitHub.com and log in with your GitHub username and password.\n\nSet up your GitHub Account\n\nIf you want to allow others to view or contribute to your project, you will need to set up your account to allow this.\n\nTo set up your GitHub account, go to the project you want to contribute to, click on the &quot;Settings&quot; button in the upper right corner of the GitHub page, and under &quot;GitHub profile&quot; click &quot;Edit&quot;. In the profile settings, you will see the &quot;Contributions&quot; tab. You can change the settings in this tab to either allow others to contribute to your project or to view your contributions to others' projects. Click on the &quot;Next&quot; button to save your changes.\n\nGitHub provides a detailed guide for the contribution settings here.]" time="1.016"><properties><property name="score" value="0.9998265504837036" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Myanmar currency and coins\n\nMyanmar Banknotes\n\nMyanmar has a local currency, the Kyat, or K, named after King Thibaw, a former king of Burma. The kyat was introduced in 1923, replacing the Indian rupee which had been in use in Burma for about two centuries. One kyat is subdivided into 100 pyas, and one kyat is subdivided into 100 cents. Coins have been minted in denominations of 5, 10, 25, 50, 100, 500 and 1000 kyats, and 1, 5, 10 and 25 pyas, though the highest four denominations are rarely used.\n\nThe coins and notes currently in circulation are:\n\nBanknotes Coins\n\nKyat\n\n50, 100, 500, 1000 Kyat\n\n5, 10, 25, 50, 100, 500 Pyas\n\n1, 5, 10, 25, 50, 100 Cents\n\n50, 100, 500, 1000 Cents\n\n5, 10, 25, 50, 100\n\nCheck out the latest exchange rates!\n\nMyanmar Currency History\n\nDuring the last half of the 19th century, large quantities of Indian coins circulated in Burma. In 1885, Burma began to issue its own coinage, in a new decimal system. The rupee and the double rupee were equivalent to one kyat. These coins were not for circulation outside of Burma and were not legal tender in India. A similar coinage was issued in the last year of the 19th century, using slightly different weights and diameters for the three silver coins and for the copper coins.\n\nDuring World War I, the Burmese Government began issuing a series of paper notes for 1, 5, 10, 20, 100 and 1000 rupees. They were overprinted on banknotes of the Imperial Bank of India, the currency used for exchange in Burma until the mid-1930s.\n\nDuring the 1920s, the paper money of several countries was in circulation in Burma. Since the demand for small denomination notes was greatest, the British Government began to issue India notes in 1925. They were of low denomination, being in the denominations of 5, 10 and 20 rupees. The existing paper money was then demonetised.\n\nIn 1940, the Japanese occupied Burma and issued their own notes in place of the British issues. These notes continued to be in use until 1945, when the British re-established control of the country. In 1942, emergency notes were introduced, in the denominations of 1, 5, 10, 20 and 100 rupees. These notes were overprinted on paper notes issued by the Reserve Bank of India.\n\nIn 1946, a new series of notes was issued in Burma. This consisted of Indian notes issued by the Reserve Bank of India, overprinted B.E.P. and the State Bank of Burma. These were in denominations of 1, 5, 10, 20, 100 and 1000 rupees.\n\nThe present series of notes began in 1951. These notes are in denominations of 1, 5, 10, 20, 50, 100, 500, 1000 and 5000 kyats.\n\nThe &quot;Burmese&quot; rupee, issued by the British between 1885 and 1896, was a fixed exchange rate at 1 shilling 4 pence (two and one-third rupees).\n\nThe U.S. dollar was established as the national currency of Burma in 1885 and is still the predominant currency in use in Myanmar.\n\nBanking in Myanmar\n\nMyanmar has an international banking system, with at least 7 major banks operating there. These include the state-owned Myanmar Foreign Trade Bank (MFTB) and the private Myanmar Economic Bank (MEB), National Bank of Commerce (NBC), Myanma Foreign Trade Bank (MFTB), Myanma Investment and Commercial Bank (MICB), Myanmar Mayflower Bank (MMB), Myanma Ahlia Bank (MAB), and Myanma Security Bank (MSB).\n\nMost commercial banks use the SWIFT system to transfer funds, and many banks in Myanmar accept electronic banking services through mobile phone banking, Internet banking, and ATMs. Some smaller banks still deal only in cash, but since cash is used mainly for larger transactions and shopping, and with so many ATMs in service now, it is no longer a problem.\n\n\n\nMyanmar currency info\n\nMyanmar info\n\nGuide to Myanmar\n\nCurrency information on country homepages]" time="2.201"><properties><property name="score" value="0.012721210718154907" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01272121&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01272121
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[This week at Marvel there\u2019s an \u201cArmor Wars\u201d as the New Warriors reunite! Will the Illuminati save our universe from the all-powerful \u201cManhattan Projects\u201d? And that\u2019s just the beginning! The Big 2 also has its second full week of August books, including DC\u2019s \u201cInjustice: Gods Among Us\u201d #11, \u201cTrinity of Sin: Pandora\u201d #3, and \u201cBatman Eternal\u201d #19.\n\nDC has the following this week: \u201cGreen Lantern\u201d #24, \u201cInjustice: Gods Among Us\u201d #11, \u201cThe Flash\u201d #24, \u201cJustice League of America\u201d #9, \u201cBatman and Robin\u201d #24, \u201cDetective Comics\u201d #24, \u201cRed Hood and the Outlaws\u201d #24, \u201cCatwoman\u201d #24, \u201cTeen Titans\u201d #24, \u201cThe Green Team: Teen Trillionaires\u201d #3, \u201cNew 52: Future\u201d #3, \u201cThe Movement\u201d #6, \u201cStormwatch\u201d #24, \u201cHarley Quinn\u201d #7, \u201cHe-Man: The Eternity War\u201d #2, \u201cInjustice: Gods Among Us\u201d #12, \u201cTrinity of Sin: The Phantom Stranger\u201d #19, \u201cSuperman/Wonder Woman\u201d #8, \u201cDeathstroke\u201d #24, \u201cSinestro\u201d #3, \u201cAll-Star Western\u201d #24, \u201cWorlds\u2019 Finest\u201d #19, \u201cThe Multiversity\u201d #1, \u201cBatgirl\u201d #24, \u201cGreen Lantern Corps\u201d #24, \u201cNew Suicide Squad\u201d #9, \u201cEarth 2\u201d #22, \u201cCatwoman\u201d #25, \u201cThe Green Team: Teen Trillionaires\u201d #4, \u201cConstantine\u201d #4, \u201cThe Movement\u201d #7, \u201cJustice League Dark\u201d #24, \u201cGreen Lantern New Guardians\u201d #24, \u201cRed Lanterns\u201d #24, \u201cAquaman\u201d #24, \u201cWonder Woman\u201d #24, \u201cHe-Man: The Eternity War\u201d #3, \u201cBatman \u201966\u201d #12, \u201cAll-Star Western\u201d #25, \u201cAction Comics\u201d #24, \u201cTeen Titans\u201d #25, \u201cCatwoman\u201d #25, \u201cStormwatch\u201d #25, \u201cBatgirl\u201d #25, \u201cAll-Star Western\u201d #26, \u201cHarley Quinn\u201d #8, \u201cDeathstroke\u201d #25, \u201cAquaman\u201d #25, \u201cJustice League 3000\u201d #8, \u201cSupergirl\u201d #24, \u201cSmallville Season 11\u201d #10, \u201cBatman \u201966\u201d #13, \u201cSuperman/Wonder Woman\u201d #9, \u201cSwamp Thing\u201d #24]" time="1.775"><properties><property name="score" value="0.9240812361240387" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[FIG. 1 is a perspective view of a virtual reality system in accordance with a preferred embodiment of the present invention;\n\nFIG. 2 is a top plan view of the virtual reality system of FIG. 1;\n\nFIG. 3 is a front plan view of the virtual reality system of FIG. 1;\n\nFIG. 4 is a side elevational view of the virtual reality system of FIG. 1;\n\nFIG. 5 is a perspective view of a two-handed tracking device in accordance with a preferred embodiment of the present invention;\n\nFIG. 6 is a perspective view of the two-handed tracking device of FIG. 5 with an array of LEDs mounted thereto;\n\nFIG. 7 is a perspective view of a three-dimensional position sensing device in accordance with a preferred embodiment of the present invention;\n\nFIG. 8 is a perspective view of a two-handed tracking device in accordance with another preferred embodiment of the present invention;\n\nFIG. 9 is a perspective view of a three-dimensional position sensing device in accordance with another preferred embodiment of the present invention;\n\nFIG. 10 is a perspective view of a three-dimensional position sensing device in accordance with another preferred embodiment of the present invention;\n\nFIG. 11 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 12 is a top plan view of the virtual reality system of FIG. 11;\n\nFIG. 13 is a front plan view of the virtual reality system of FIG. 11;\n\nFIG. 14 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 15 is a top plan view of the virtual reality system of FIG. 14;\n\nFIG. 16 is a front plan view of the virtual reality system of FIG. 14;\n\nFIG. 17 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 18 is a top plan view of the virtual reality system of FIG. 17;\n\nFIG. 19 is a front plan view of the virtual reality system of FIG. 17;\n\nFIG. 20 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 21 is a top plan view of the virtual reality system of FIG. 20;\n\nFIG. 22 is a front plan view of the virtual reality system of FIG. 20;\n\nFIG. 23 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 24 is a top plan view of the virtual reality system of FIG. 23;\n\nFIG. 25 is a front plan view of the virtual reality system of FIG. 23;\n\nFIG. 26 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 27 is a top plan view of the virtual reality system of FIG. 26;\n\nFIG. 28 is a front plan view of the virtual reality system of FIG. 26;\n\nFIG. 29 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 30 is a top plan view of the virtual reality system of FIG. 29;\n\nFIG. 31 is a front plan view of the virtual reality system of FIG. 29;\n\nFIG. 32 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 33 is a top plan view of the virtual reality system of FIG. 32;\n\nFIG. 34 is a front plan view of the virtual reality system of FIG. 32;\n\nFIG. 35 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 36 is a top plan view of the virtual reality system of FIG. 35;\n\nFIG. 37 is a front plan view of the virtual reality system of FIG. 35;\n\nFIG. 38 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 39 is a top plan view of the virtual reality system of FIG. 38;\n\nFIG. 40 is a front plan view of the virtual reality system of FIG. 38;\n\nFIG. 41 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 42 is a top plan view of the virtual reality system of FIG. 41;\n\nFIG. 43 is a front plan view of the virtual reality system of FIG. 41;\n\nFIG. 44 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 45 is a top plan view of the virtual reality system of FIG. 44;\n\nFIG. 46 is a front plan view of the virtual reality system of FIG. 44;\n\nFIG. 47 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 48 is a top plan view of the virtual reality system of FIG. 47;\n\nFIG. 49 is a front plan view of the virtual reality system of FIG. 47;\n\nFIG. 50 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 51 is a top plan view of the virtual reality system of FIG. 50;\n\nFIG. 52 is a front plan view of the virtual reality system of FIG. 50;\n\nFIG. 53 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 54 is a top plan view of the virtual reality system of FIG. 53;\n\nFIG. 55 is a front plan view of the virtual reality system of FIG. 53;\n\nFIG. 56 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 57 is a top plan view of the virtual reality system of FIG. 56;\n\nFIG. 58 is a front plan view of the virtual reality system of FIG. 56;\n\nFIG. 59 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 60 is a top plan view of the virtual reality system of FIG. 59;\n\nFIG. 61 is a front plan view of the virtual reality system of FIG. 59;\n\nFIG. 62 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 63 is a top plan view of the virtual reality system of FIG. 62;\n\nFIG. 64 is a front plan view of the virtual reality system of FIG. 62;\n\nFIG. 65 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 66 is a top plan view of the virtual reality system of FIG. 65;\n\nFIG. 67 is a front plan view of the virtual reality system of FIG. 65;\n\nFIG. 68 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 69 is a top plan view of the virtual reality system of FIG. 68;\n\nFIG. 70 is a front plan view of the virtual reality system of FIG. 68;\n\nFIG. 71 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 72 is a top plan view of the virtual reality system of FIG. 71;\n\nFIG. 73 is a front plan view of the virtual reality system of FIG. 71;\n\nFIG. 74 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 75 is a top plan view of the virtual reality system of FIG. 74;\n\nFIG. 76 is a front plan view of the virtual reality system of FIG. 74;\n\nFIG. 77 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 78 is a top plan view of the virtual reality system of FIG. 77;\n\nFIG. 79 is a front plan view of the virtual reality system of FIG. 77;\n\nFIG. 80 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 81 is a top plan view of the virtual reality system of FIG. 80;\n\nFIG. 82 is a front plan view of the virtual reality system of FIG. 80;\n\nFIG. 83 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 84 is a top plan view of the virtual reality system of FIG. 83;\n\nFIG. 85 is a front plan view of the virtual reality system of FIG. 83;\n\nFIG. 86 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 87 is a top plan view of the virtual reality system of FIG. 86;\n\nFIG. 88 is a front plan view of the virtual reality system of FIG. 86;\n\nFIG. 89 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 90 is a top plan view of the virtual reality system of FIG. 89;\n\nFIG. 91 is a front plan view of the virtual reality system of FIG. 89;\n\nFIG. 92 is a perspective view of a virtual reality system in accordance with another preferred embodiment of the present invention;\n\nFIG. 93 is a top plan view of the virtual reality system of FIG.]" time="4.801"><properties><property name="score" value="0.9751322070757548" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[To show their support for Venezuela, five prominent Tibetan Buddhist teachers from the United States will be joining a gathering in Caracas, Venezuela on September 24, 2015. The purpose of this gathering is to support the effort of the Venezuelan people to bring peace and prosperity to the country.\n\nWe will be holding teachings and prayers in three different locations. The first event will be held at 1:30 p.m. at the Cine Teatro Teresa Carreno in Caracas. This venue can seat 900 people. The second and third events will be held at the Jos\xe9 Antonio Paez Theatre, which can seat around 400 people.\n\nIf you are in Venezuela, please attend one of these events. And if you cannot come in person, please join us in spirit by thinking prayers for Venezuela, especially during the times of the teaching and prayers.\n\nThis event is being organized by Geshe Tsultrim Gyeltsen, president of the Foundation for the Preservation of the Mahayana Tradition (FPMT). He will be joined by other FPMT teachers, including Ven. Robina Courtin, Ven. Tenzin Chonyi, Ven. Tsondue Tsomo, and Ven. Marta Meana.\n\nThe event is being supported by the Embassy of the Bolivarian Republic of Venezuela to the United States, and the Embassy of the Bolivarian Republic of Venezuela to the United Nations.]" time="0.748"><properties><property name="score" value="0.9988296627998352" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[more-in\n\nJust ahead of the International Yoga Day on June 21, the Gujarat High Court on Monday observed that the Prime Minister Narendra Modi-led government's decision to declare June 21 as International Yoga Day was against the spirit of the Constitution.\n\nHearing a PIL against the Centre's decision, the High Court said that though there was nothing wrong in promoting yoga, but that should not be at the cost of other forms of physical exercise. The Centre's decision to declare International Yoga Day was against the spirit of Article 51A (h) of the Constitution, which emphasises on the promotion of scientific temper, humanism and the spirit of inquiry and reform.\n\n&quot;We will have to consider various aspects like whether the decision to declare June 21 as International Yoga Day is against the Constitution of India. A decision has to be taken as to whether scientific temper, humanism and spirit of inquiry is being followed by the Centre or not,&quot; observed Justice A.J. Desai.\n\nS.P. Tamang, the petitioner, had sought a direction to the Centre to hold scientific discussions before declaring a holiday. The petitioner also questioned the need for declaring June 21 as International Yoga Day and claimed that various festivals and events were already being celebrated by India on the same day.]" time="0.560"><properties><property name="score" value="0.09942474961280823" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.09942475&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.09942475
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Open file to use in Google Earth (KML)\n\nCreate Waypoint or Locator from the KML loaded.\n\n\n\nFile &gt; &quot;Open file to use in Google Earth&quot;\n\n\n\nExample\n\nC:\\Program Files (x86)\\Google\\Google Earth\\GoogleEarth.exe\n\nC:\\Program Files (x86)\\Google\\Google Earth\\kmz\\Dhahran.kml\n\nC:\\Program Files (x86)\\Google\\Google Earth\\kmz\\Dhahran.kml\n\nC:\\Program Files (x86)\\Google\\Google Earth\\kmz\\Dhahran.kml\n\nC:\\Program Files (x86)\\Google\\Google Earth\\kmz\\Dhahran.kml\n\n...\n\nOptions - Jumps to KML file in file chooser]" time="0.374"><properties><property name="score" value="0.9997754693031311" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Schools\n\n\nThis report focuses on the experiences of LGBT youth in the nation's public and private schools. This report shows that anti-LGBT victimization in school is a serious problem for a large number of students. It also provides data on other forms of bullying and discriminatory treatment that LGBT students experience. Some of the findings of this report are particularly disturbing. For example, over half (53.1 percent) of LGBT students experienced feeling unsafe in school because of their sexual orientation, over one-quarter (26.1 percent) because of their gender expression, and over one-third (35.2 percent) because of both their sexual orientation and gender expression. This harassment and discrimination takes a toll on these students' well-being. The impact of the harassment and discrimination experienced by LGBT students is serious. LGBT students were more likely than non-LGBT students to report several adverse outcomes. For example, they were more likely than non-LGBT students to report high levels of depression (33.1 percent of LGBT students compared to 19.8 percent of non-LGBT students), to have seriously considered attempting suicide (32.2 percent of LGBT students compared to 13.2 percent of non-LGBT students), and to have made a suicide attempt (21.8 percent of LGBT students compared to 4.3 percent of non-LGBT students). Students who were more highly victimized were also more likely to experience these negative outcomes. For example, compared to students who experienced little or no anti-LGBT victimization, students who experienced a great deal of anti-LGBT victimization were more than three times as likely to have been threatened or injured with a weapon on school property (10.0 percent compared to 3.6 percent), more than three times as likely to have been threatened or injured with a weapon off school property (14.7 percent compared to 4.3 percent), more than three times as likely to have experienced school violence (35.2 percent compared to 10.6 percent), more than four times as likely to have attempted suicide (25.0 percent compared to 5.7 percent), and more than six times as likely to have made a suicide attempt that required treatment from a doctor or nurse (10.4 percent compared to 1.8 percent).\n\nIntroduction\n\n\nThis report presents data from the National School Climate Survey (NCS), which collected data from over 7,000 lesbian, gay, bisexual, and transgender (LGBT) students ages 13\u201321 in public and private schools in every state and the District of Columbia during the spring of 2015. The report focuses on the experiences of LGBT students in public schools, but private school students' experiences are briefly discussed to provide context and additional information about the experiences of LGBT students in schools across the country.1 This report also includes a limited analysis of the data on the experiences of bisexual youth. This report provides the first national data on LGBT students' experiences in our nation's schools.\n\n\nLGBT students face a number of serious challenges at school, and this report provides a comprehensive look at the extent of the problem and its impact on LGBT students. Many LGBT students have experienced verbal harassment, physical assault, and sexual violence at school. For example, over half (53.1 percent) of LGBT students have felt unsafe at school because of their sexual orientation, over one-quarter (26.1 percent) because of their gender expression, and over one-third (35.2 percent) because of both their sexual orientation and gender expression. In addition, over one-third (35.2 percent) of LGBT students have been verbally harassed and 19.6 percent have been physically attacked at school in the past year because of their sexual orientation or gender expression. These figures are even higher for transgender students: over half (53.8 percent) have felt unsafe at school because of their gender expression and over one-third (36.2 percent) because of their gender identity. Furthermore, over one-quarter (27.2 percent) of LGBT students have missed at least one entire day of school in the past month because they felt unsafe or uncomfortable, and the prevalence of truancy increases further for transgender students: 40.1 percent of transgender students have missed at least one day of school in the past month because they felt unsafe or uncomfortable.\n\n\nThis report also shows that the harassment and discrimination that LGBT students experience can lead to serious and negative outcomes for these students, including lower school engagement, lower levels of happiness and life satisfaction, and higher levels of depression and even higher levels of suicidal behavior. Over one-third (34.7 percent) of LGBT students reported missing school for at least one day during the month before the survey because they felt unsafe or uncomfortable. Students who experienced more victimization because of their LGBT status were also more likely to report this type of adverse outcome. For example, compared to students who experienced little or no victimization, students who experienced a great deal of victimization were almost four times as likely to have missed at least one day of school in the month before the survey because they felt unsafe or uncomfortable (45.6 percent compared to 12.3 percent).\n\n\nThe negative impact of this victimization on LGBT students can be seen in their reports of lower levels of engagement in school activities, lower levels of happiness and life satisfaction, and higher levels of depression. LGBT students who experienced high levels of victimization because of their sexual orientation, gender expression, or both were more likely than students who experienced low levels of victimization to feel that they could not be themselves at school (68.6 percent compared to 42.2 percent) and to feel unhappy or depressed at school (40.7 percent compared to 19.5 percent). Students who experienced high levels of victimization because of their sexual orientation, gender expression, or both were also more likely to report symptoms of depression than students who experienced low levels of victimization. For example, compared to students who experienced little or no victimization, students who experienced high levels of victimization were almost twice as likely to report feeling unhappy or depressed (38.7 percent compared to 20.2 percent) and almost three times as likely to report feeling less happy than other students (14.3 percent compared to 5.4 percent).\n\n\nData from the 2015 NCS show that many students still face serious and pervasive mistreatment and discrimination at school. These experiences can have a serious negative impact on students' well-being. This is especially true for transgender and gender non-conforming students who are most likely to]" time="3.298"><properties><property name="score" value="0.6440417000225612" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The band's creative soul and ringleader, Robert Pollard, reflects on the wild success of Guided By Voices as well as his own triumphs and tragedies.\n\nRobert Pollard \u2014 the creative soul and ringleader of Guided By Voices \u2014 is a living embodiment of the enduring power of rock \u2018n\u2019 roll. Pollard, whose band just released its 15th album, August By Cake, has lived a life that\u2019s been as varied as it has been eventful. From crafting classic lo-fi pop records and guiding his beloved group to victory after victory for over two decades, to overcoming multiple tragedies that threatened to undo him, Pollard has endured and endured well.\n\nFrom rock \u2018n\u2019 roll heartbreak to musical triumphs and to the stories of his childhood friends who took their own lives, the tales behind Pollard\u2019s life are as varied as the songs he\u2019s written over the years. They\u2019re full of sweetness, heartache and inspiration \u2014 the kind that only music can provide.\n\nHere, Pollard reflects on the enduring and wonderful power of rock \u2018n\u2019 roll.\n\nOn being inspired by Bob Dylan\n\n\u201cHe made rock \u2018n\u2019 roll feel alive to me, you know? He just did things no one else was doing. The first time I heard his music, I was just in awe of it. I thought he was one of the only people doing something different in music and it really inspired me to make a rock band.\u201d\n\nOn working at Tower Records\n\n\u201cI worked at Tower Records on Sunset Boulevard in the Eighties. I was in a band that was signed to a major label and we were going to go on tour. We were going to make our first big move out to L.A. and the band broke up in Pittsburgh. I was heartbroken. I went to L.A. anyway and got a job at Tower and the band started to do really well in Pittsburgh. I started to realize that I was being a moron by working at a job that didn\u2019t have anything to do with music. I should have been playing music.\n\n\u201cThen, I met people that worked at the store \u2014 people that worked in the music section \u2014 and they were people that were into the same kind of music that I was. They were all very into the new wave thing that was going on at that time and I was very into the punk scene and I was discovering all this underground music. That was a good time because I was discovering so many great bands and all these great people.\u201d\n\nOn putting together Guided By Voices\n\n\u201cI was living in this little house on Cass Avenue in Detroit and I would just record these songs in the basement. I didn\u2019t think anyone would be interested. I just thought I was doing this for myself. Then, people started calling me and saying they wanted to play on the songs and asking if I could make copies of them. Eventually, I had enough to make a record. I was putting the band together around the same time.\n\n\u201cIn the beginning, it was just one song. The whole point was to just have one song on a compilation of different artists from Michigan. That was the point of the whole thing. It was like, \u2018Oh, that\u2019s fun. That\u2019s really fun. Let\u2019s do another song.\u2019 It just became something else. Then, I started working with the other people. The other people were my friends. My friends were just my buddies that I\u2019d played music with. The first couple of years was just me and my buddies doing these songs. Then, it started to become a bigger thing and we started to tour and it became something else. It just turned into something else. It was an accident that it happened.\u201d\n\nOn his songwriting approach\n\n\u201cMy songs have always come from an unconscious, naive place. I really have to focus on a song and work on a song in order for it to have a definite melody or a definite idea. The words, I kind of write more like a novel and I try to write songs that work like a story. When I\u2019m doing them, I\u2019m trying to do it like a movie. I\u2019m trying to think of it in terms of a movie and I\u2019m trying to create characters and set the scenes and tell the story and make it a little bit deeper than what you would normally expect from a pop song. I\u2019m trying to take it to a different level in some ways. That\u2019s always been my goal and that\u2019s always been my style.\n\n\u201cI don\u2019t think I\u2019m trying to reinvent the wheel, I just think that I\u2019m trying to take things that are already in existence and make them bigger and more beautiful and a little more complete. I don\u2019t know. I don\u2019t really try to analyze it that much. I\u2019ve always written songs in the same way and the only thing that\u2019s changed is the way I record them. I used to record them all on a boombox and then, as we went along, we started to get more and more into studio work.\u201d\n\nOn recording songs at home\n\n\u201cI like to record in a place that\u2019s not a studio. I\u2019ve always liked to record in a place where I could have access to as much stuff as possible. I really enjoy it when I have a lot of cool things to play with. I always like to record at home. I think it\u2019s important to make the place you record at like a laboratory, you know? I like to have it be a lab. If I could, I\u2019d have every piece of equipment ever made and be able to put it all in one room and be able to experiment with every sound I could possibly make. It\u2019s so much fun and it\u2019s so much work at the same time, too. I love it.\u201d\n\nOn his songwriting process\n\n\u201cThe way I work, I just keep track of my thoughts and feelings. I think about it when I\u2019m walking or when I\u2019m eating or I\u2019m just doing stuff. I just keep track of stuff and I write down a lot of notes and I try to pay attention to things. My life is so boring, really, you know? I don\u2019t do anything that exciting. I just pay attention to things and I think about stuff and I write down little ideas. I do that all the time. I write little ideas all the time and I just try to remember them and try to put them in a folder.\n\n\u201cIt\u2019s kind of like a filing system. The filing system is in my head. I try to remember everything I think about and I try to write it down and then, eventually, it becomes a song. Then, I try to look at the songs that are done and I try to pick the ones that seem to be the best and start thinking about them in terms of a story or a character or a scenario or something that would be entertaining. I have to think about all the things that are going on in my life and try to piece things together like a collage.\n\n\u201cThen, when it comes to the recording, I\u2019m thinking about the whole thing as a complete package. I\u2019m thinking about the sound of it and I\u2019m thinking about how I\u2019m going to mix the album and everything. I\u2019m thinking about the lyrics and the voice. Everything is kind of part of a bigger thing and the sound and the performance is really important.\u201d\n\nOn keeping it simple\n\n\u201cIt\u2019s easier to write when you\u2019re just keeping it simple. When you start to complicate things, it becomes too big of a thing to think about. That\u2019s when the song becomes too precious and you can\u2019t finish it or you start getting scared of it or something like that. Then, you get caught up in the emotion of it and you can\u2019t do it. You have to remember that it\u2019s just something you\u2019re doing to pass the time. It\u2019s just something you\u2019re doing for fun. That\u2019s what it is for me. It\u2019s just fun. I\u2019m not trying to be any big shot or anything like that. I\u2019m just trying to do something that\u2019s fun and I\u2019m just trying to make something that\u2019s going to give me some satisfaction. I\u2019m just trying to do something that\u2019s going to be fun to listen to and fun to make.\u201d\n\nOn his songwriting being like therapy\n\n\u201cYeah. I\u2019m not trying to be a serious writer or anything like that. I\u2019m not trying to be pretentious at all. I just want to make a record that I\u2019d like to listen to.\u201d\n\nOn the joy of performing\n\n\u201cIt\u2019s a good feeling, you know? There\u2019s something very satisfying about it. You just can\u2019t stop it, you know? I\u2019ve been doing it for so long and it just doesn\u2019t go away. You can\u2019t stop it. I just keep on doing it and I can\u2019t stop it. I think that\u2019s the reason why I do it, you know?\u201d\n\nOn discovering The Beatles\n\n\u201cThe Beatles made]" time="4.951"><properties><property name="score" value="0.9997810274362564" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[CHICAGO (AP) \u2014 An American Airlines pilot has been removed from flying duties after he temporarily parked a plane on a taxiway where four other airliners were waiting to take off at Chicago's O'Hare International Airport.\n\nAn investigation into the Saturday incident is ongoing, airline spokesman Ross Feinstein said Sunday.\n\nFeinstein said the pilot has been taken off flying duty &quot;pending the outcome of the investigation.&quot; He didn't specify how long the pilot would be grounded.\n\nNo other crew members were disciplined, and the incident won't affect the scheduling of the roughly 3,000 flights a day the airline operates around the world, Feinstein said.\n\nHe declined to release the name of the pilot or the number of years he had been flying for the airline.\n\nThe Chicago Department of Aviation said in a statement that the pilot of the American Airlines plane, which had arrived from Nashville, Tenn., was supposed to go to a designated taxiway but instead went to a parallel taxiway where four other airliners were waiting for clearance to take off.\n\nIt wasn't immediately clear how close the plane came to taking off.\n\nAccording to radar information cited in the aviation department statement, the aircraft came within 100 feet of an aircraft on the ground and approximately 200 feet from another aircraft.\n\nThe aviation department is investigating the incident along with the Federal Aviation Administration.\n\nAmerican Airlines has endured two crashes this year and five in 2015. Saturday's incident is unlikely to do much to allay the fears of passengers and the traveling public.\n\n&quot;It's obviously one of those things that happen occasionally. It's a pretty serious mistake,&quot; said Henry Harteveldt, founder of the Atmosphere Research Group. &quot;It's a reminder that airline pilots are human beings and that despite their extensive training, sometimes they make mistakes.&quot;\n\n___\n\nAssociated Press writer Don Babwin contributed to this report.]" time="0.847"><properties><property name="score" value="0.02879965305328369" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[As a manufacturer, it is easy to produce one-of-a-kind pieces, but even easier to make a product for everyone, because every piece is made to the same specifications. What happens when you decide to make a new product and to have a bigger impact? You might think that you have to start with an already existing product and create something new. But this is not always true.\n\nSometimes you need to start from scratch and create something new \u2013 a new form, a new material, new applications or new customer experiences. This is what happened when Ikea created the L\xd6MSK light collection \u2013 a range of products made of plaster and leather, using two materials that we usually associate with other products.\n\nL\xd6MSK: it\xb4s easy being a one-of-a-kind light\n\nTwo elements were the starting point of the L\xd6MSK light collection: an existing material and a very unusual use for it. Plaster and leather are very traditional materials that are linked to very different products. We know plaster for its use in wallpapers, ceilings and sculptures, but it is not very common to use it in products such as lights or furniture. Leather has a much stronger association with leather goods, clothes, bags or shoes. And in fact, leather is one of the most used materials in the furniture industry.\n\nTwo products, plaster and leather, that are very common to different types of industries but never used together.\n\nIkea wanted to combine these two materials in a collection of lights, lamps and other decorative elements, in order to make a complete lighting system that can be easily incorporated into any interior and made more personal.\n\nA creative team started exploring ways of using these two materials in a complete lighting collection. They started experimenting with different elements such as the light bulb and the suspension, to combine them and to make sure they can adapt to the maximum number of situations. The development of the light was not a process of making a more perfect product, but rather a process of exploring new ways of using existing products.\n\nAs a result, they developed three series of products. One made of plaster and leather, and two others made of plaster. The leather light is one of the most surprising elements in the collection, as it is hard to believe that this kind of leather is the same that we see in jackets or shoes.\n\nThe leather used in the light is dyed in different colours in order to add interest to the design and it is used to create the \u201cU\u201d shape that holds the light bulb. This element was designed to be a flexible element in the light, allowing the user to change the direction of the light and adapt it to different uses.\n\nFinally, the leather light can also be suspended with a wire, making it ideal for reading lights.\n\nThe second series, made of plaster, is the plateroom series, and the third series is a collection of lamps made of plaster with different shades. The aim was to use plaster to make different types of lamps that can adapt to the different sizes of space. This is achieved by combining the plaster with different shades and creating elements such as the plateroom bench and the plateroom table, with different dimensions and shades.\n\nThere is an important message that Ikea wants to communicate with this collection: that it is possible to create something new and original.\n\nYou need to start with something simple but at the same time creative and unusual. This is what Ikea wants to achieve with this collection: creating a new product that is easy to understand and appreciate, but at the same time has an innovative design and can be used in different types of interior.\n\nThis can be achieved by starting with a very simple element and having an open mind, as the designers did with this collection. Ikea has created a very unusual product that is adaptable to different situations and can be adapted to any kind of interior or architecture.]" time="2.032"><properties><property name="score" value="0.9851841926574707" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Good for price\n\nBy Anonymous\n\n06 Apr 2012\n\nShimano PD-T400 SPD Pedals\n\n3 / 5\n\nI recommend this product\n\nCons rubber grip strip wears quickly, not made to wear a cleat off road or in the mud Pros inexpensive, comfortable, easy to adjust, replaceable\n\nMy first set lasted me 4 years on the commuter bike and MTB with up to 50 mile/week rides. The rubber grips are made to wear down quickly, and don't offer a lot of friction to hold on to (that's good and bad), but replaceable. The cleat is replaceable too, which is nice if you walk on the pedals a lot (they do have walkable cleats).\n\n\n\nOverall, a great value pedal for commuter bikes. The tension is adjustable, but it is easy to over-tighten and bend the pedal.]" time="0.401"><properties><property name="score" value="0.9497540593147278" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.94975406&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.94975406
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Movie details\n\n7.5 / 10 ( 189 )\n\nAn old friend of Archie's has been released from prison and needs a place to stay. Archie offers his hospitality to &quot;Leroy&quot; but has a hard time adjusting to this out-of-place visitor. The ...]" time="0.158"><properties><property name="score" value="0.9476650953292847" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.9476651&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.9476651
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The Pennsylvania governor's race is getting a jolt from two longshot third-party candidates \u2014 Libertarian Ken Krawchuk and Green Party candidate Paul Glover.\n\nThe two men were on the ballot with Republican Tom Corbett and Democrat Dan Onorato until the parties removed them. Onorato sued to keep Krawchuk on the ballot, but the state Supreme Court on Monday said he could be dropped.\n\nKrawchuk says he's energized by the attention. He plans a statewide tour and calls Corbett &quot;the worst candidate for governor&quot; in the state's history.\n\nGlover is campaigning in Philadelphia and plans a statewide tour with the Green Party, but he's unlikely to win.\n\nCopyright Associated Press]" time="0.329"><properties><property name="score" value="0.7722986340522766" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Karen &amp; Ryan, 9-6-08\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n]" time="0.090"><properties><property name="score" value="0.7698278427124023" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Jubal Early\n\nJubal Anderson Early (July 14, 1816July 2, 1894) was a Confederate general in the American Civil War and a politician after the war. Although he served almost the entire war in the eastern theater, his most significant contributions were in the Valley Campaigns of 1864, which ended in the stalemate at Petersburg, Virginia, and the capture of the Union garrison at Lynchburg, Virginia. Early is considered among the best Confederate generals who never served in the Army of Northern Virginia.\n\nEarly was born in Southampton County, Virginia, the seventh of ten children of Ruth (Anderson) Early and Judge John Early, a former governor of Virginia and member of the United States Congress. In 1825, Early attended Charlotte Hall Military Academy. He graduated second in his class at the United States Military Academy at West Point in 1838, placing him above such future generals as Stonewall Jackson and Ulysses S. Grant.\n\nEarly was appointed a second lieutenant in the 4th U.S. Artillery. He served as an aide to General Winfield Scott in the War with Mexico and fought in the Battle of Molino del Rey and the Battle of Chapultepec. Early served as the acting chief of artillery in the assault on the city of Mexico, and received two brevets for his bravery in these battles. He was one of the first officers to enter Mexico City, and he served as an artillery instructor at West Point from 1847 to 1852. Early served as a major in the United States Army in the Seminole Wars in 1855, and as a captain in the United States Dragoons in 1858.\n\nEarly entered the Virginia Military Institute in 1852, and became a professor of mathematics at the Institute, where he taught until the Civil War. He was known for his severity as a disciplinarian. Early became known as a superb military tactician and administrator, as well as an able administrator and strategist.\n\nIn 1860, Early entered politics, running as a Democrat for one of the two seats in the Virginia House of Delegates from Southampton County, but he was defeated. The state seceded from the Union on April 17, 1861, and Early was commissioned as a colonel of artillery in the Virginia militia on May 14. When Virginia militia units were activated for Confederate service, Early was appointed a brigadier general of Virginia forces on May 27.\n\nEarly initially commanded a brigade in the brigade of Brig. Gen. Richard B. Garnett. When Garnett died at the Battle of Mill Springs, Early was promoted to brigadier general on January 16, 1862. His brigade was assigned to Maj. Gen. Edmund Kirby Smith's division in the Army of Northern Virginia. Early fought in the Peninsula Campaign, including the Siege of Yorktown, during which his brigade was an important factor in repulsing a Union assault on the right flank of the Confederate entrenchments at Savage's Station. Early received praise for his actions at Hanover Court House.\n\nEarly distinguished himself during the Seven Days Battles. He received his commission as a major in the Confederate Regular Army on June 9, 1862, but was later returned to his previous rank of colonel in the reserves. During the Northern Virginia Campaign, he received plaudits for his actions at the Second Battle of Bull Run.\n\nEarly was promoted to brigadier general in the Confederate regular army on October 11, 1862. He was given command of a brigade in Maj. Gen. Robert E. Rodes's division of A.P. Hill's III Corps. During the Maryland Campaign, Early performed well at the Battle of South Mountain.\n\nDuring the Battle of Fredericksburg, Early was severely wounded in the right thigh and temporarily assigned to administrative duties. He returned to the field in time to defeat Union Maj. Gen. Joseph Hooker's offensive at the Battle of Chancellorsville. At Chancellorsville, Early performed ably in combat and earned praise for his conduct as a brigade commander. Early was sent to western Virginia, where he defeated Union Maj. Gen. William S. Rosecrans at the Battle of Mill Springs. After Chancellorsville, Early was promoted to the permanent rank of brigadier general in the Confederate Regular Army on May 21, 1863, and given command of a division.\n\nEarly fought under Maj. Gen. Thomas J. &quot;Stonewall&quot; Jackson during the Valley Campaigns of 1864. Early's old brigade was reassigned to Maj. Gen. Richard H. Anderson. The Second Battle of Kernstown was his most significant contribution during the Valley Campaigns. Following the Valley Campaigns, Early was assigned to Maj. Gen. John C. Breckinridge's command in the Shenandoah Valley.\n\nHe conducted the Valley Campaigns of 1864. Following Early's victory at the Battle of Lynchburg in June 1864, Early's army became separated by Union Maj. Gen. David Hunter's raid. Hunter's force was driven from the Valley in the Battle of Lynchburg by Early's troops. During the Valley Campaigns, Early defeated Union Maj. Gen. George Crook at the Battle of Rutherford's Farm and the Battle of Tom's Brook, drove Maj. Gen. David Hunter from the Valley in the Second Battle of Kernstown, and defeated Union Maj. Gen. Philip Sheridan at the Third Battle of Winchester.\n\nEarly defeated Union Maj. Gen. David Hunter's Valley Campaign in the Second Battle of Kernstown in July 1864. He then crossed into Maryland and threatened Washington, D.C., forcing Grant to send reinforcements to the capital area. During the subsequent Valley Campaigns of 1864, Early defeated Maj. Gen. Philip Sheridan at the Battle of Opequon, the Third Battle of Winchester, and the Battle of Fisher's Hill. Sheridan's victory at the Battle of Cedar Creek forced Early to retreat back to the Valley.\n\nEarly was assigned to the Department of the West and ordered to reinforce the Confederate Army of Tennessee, but Maj. Gen. John C. Breckinridge convinced him to resist Maj. Gen. William T. Sherman's advance. At the beginning of the Atlanta Campaign, Early was given command of the newly created Third Corps of the Army of Tennessee. During the Battle of Resaca, he launched an unsuccessful assault against a Union fortification on the western bank of the Oostenaula River, suffering heavy casualties. During the subsequent battles of Adairsville and Cassville, his corps attempted to flank Union Maj. Gen. William T. Sherman's left flank, but his subordinates were slow in attacking and failed to coordinate their movements with other Confederate forces. After the Confederates suffered defeat at the Battle of Peachtree Creek, Early retreated with his corps to Lovejoy's Station. He ordered Maj. Gen. Benjamin F. Cheatham's division to open the Battle of Atlanta by attacking the Union force under Maj. Gen. James B. McPherson. Early was in the town of Marietta when he learned that Lt. Gen. John B. Hood was wounded during the Battle of Atlanta, and he returned to Atlanta to take command of the army.\n\nDuring the subsequent Siege of Atlanta, the Army of Tennessee lacked provisions and was forced to withdraw from Atlanta. The army was defeated during the subsequent pursuit, and Early's corps was heavily engaged at the Battle of Ezra Church. In November 1864, the Army of Tennessee was reorganized and Early was given command of the new Second Corps. At the Battle of Franklin, he was killed by a shot through the heart, and his corps suffered significant casualties. The Second Corps was commanded by Maj. Gen. William B. Bate after Early's death. Bate was also wounded and captured, and the corps was virtually destroyed. Bate's capture led to another shakeup of the army's command structure, and Lt. Gen. Alexander P. Stewart assumed command of the Second Corps.\n\nEarly was buried at the Old Methodist Cemetery in Alexandria, Virginia, but the cemetery closed and his remains were moved to Richmond National Cemetery.\n\nThe Virginia Historical Society operates the Jubal Early Memorial Cemetery near New Market, Virginia, where he is buried.\n\nEarly was known as &quot;Old Jubal&quot; to his men and &quot;Fighting Joe&quot; to his friends. Confederate General Robert E. Lee was a friend of Early's, and named his youngest son, Jubal Anderson, after him.\n\nIn the film &quot;Gettysburg&quot;, Early was portrayed by actor Martin Sheen. The movie presented Early as a vain, glory-seeking but weak commander who bickered with fellow Confederate generals Joseph E. Johnston and James Longstreet over who was in command of the Army of Tennessee.\n\nHe is also featured in Jeff Shaara's &quot;The Last Full Measure&quot; trilogy, where he is portrayed by Jon Voight.\n\nThe James River Squadron was a Confederate Navy organization during the American Civil War which operated on the James River, Virginia.\n\n\n\n\n]" time="4.573"><properties><property name="score" value="0.738131602605184" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Bangalore United\n\nBangalore United is a men's team of the Bangalore United Football Club.\n\nIn the past Bangalore United Football Club had two men's teams, Bangalore FC and Bangalore United.\n\nBangalore United FC was a founder member of the I-League in 2007 and competed in the first season of the competition, finishing in third place in the Southern Group. The club withdrew from the 2008\u201309 season for financial reasons and since then have played in the Karnataka State Football Association's Leagues and the Bangalore Super Division.\n\n\nThe following players are registered for the 2014\u201315 I-League.\n\n&quot;As of 5 January 2015&quot;\n\n]" time="0.309"><properties><property name="score" value="0.9923925399780273" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Colorimetric Detection of Bovine Serum Albumin by Sodium Dodecyl Sulfate-Boronate Affinity Chromatography with Conventional Staining: Optimization by Factorial Design\n\nD. J. Lindstedt, F. C. Hanft, and J. P. Weese\n\nCereal Chemistry\n\nVolume 67(4), 380-383\n\nJune 1990\n\nThe detection of serum albumin in bovine milk by sodium dodecyl sulfate-borate affinity chromatography was investigated with the use of Con A stain. After optimization, it was determined that a total of 6.6 ml of bovine plasma (0.5 mg protein/ml) and 1.2 ml of milk sample were sufficient for the analysis of 0.2 ml of milk. A total of 35 ml of buffer A (20 mM sodium phosphate pH 7.0) containing sodium dodecyl sulfate, boronate, Con A, and 0.5 M sodium sulfate were added to an 8 \xd7 100 mm borosilicate glass column. Bovine plasma and milk samples were passed through the column at a flow rate of 1.5 ml/min. The flow-through was collected and fractions were obtained by passing 1.5 ml of fresh buffer A through the column. Bovine serum albumin was detected in the fractions after staining with the Con A. The mean recovery was 2.78 \xb1 0.01% for bovine plasma and 1.28 \xb1 0.07% for bovine milk.\n\nGo to the article in Cereal Chemistry\n\nKeywords: Milk and dairy products, Bioactive compounds, Animals, Mammals, Bovines, Colostrum, Proteins, Bovine, Dried-milk products, Dairy products, Whole milk, Milk fat, Milk-fat globule membranes, Caseins, Sodium dodecyl sulfate, Serum albumin, Affinity chromatography, Sodium dodecyl sulfate-borate, Concanavalin A stain, Affinity chromatography, Affinity separations, Immobilized enzymes]" time="1.063"><properties><property name="score" value="0.16801482439041138" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Hi Guys,We need your help.We are writing an essay about a topic of our choice and we need help with a paragraph.The topic is:&quot;The more mobile and connected we are, the more isolated we feel. Is it because the more we interact the more we expect? Is it because our friends are not always available? We are flooded with so many information but we don\u2019t know how to manage it and, in a way, are lost in it. The real question is, how can technology provide more human connections and make us less lonely? &quot;Please feel free to share any ideas you have.Thanks and have a great day.Thesis = + http://www.plenglish.com/index.php?option=com_content&amp;view=article&amp;id=9288:2014-04-21-22-52-20&amp;catid=40:tecnologia-ciencia&amp;Itemid=51]" time="0.513"><properties><property name="score" value="0.9997774958610535" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Junior Potchefstroom Boys\n\nBoys' Junior Cup\n\n04 April 2019\n\nGrade: Gauteng North (South)\n\nDivision: Pool A\n\nPool: A\n\nResults\n\nGame 1 (09:30)\n\nPlayStation/Yello Pockets v Athlone (Kingfisher Blue 2)\n\nWin: 3 wickets\n\nAthlone won the toss and elected to field\n\nPlayStation/Yello Pockets - 75/10 in 10 overs\n\nKhawu v Chamuhiya 2 runs\n\nChirag v Alam 8 runs\n\nQalbya v Singh 8 runs\n\nRyan v Daggupan 6 runs\n\nSantlal v Taufik 5 runs\n\nTiaan v Alam 0 runs\n\nScorecard\n\nAthlone - 68/10 in 9.4 overs\n\nAwal v Ryan 16 runs\n\nSeqib v Santlal 4 runs\n\nImran v Qalbya 4 runs\n\nShahir v Chirag 2 runs\n\nMubarak v Kole 2 runs\n\nAhmed v Chamuhiya 4 runs\n\nFarook v Kabbya 5 runs\n\nNadeem v Ahli 6 runs\n\nScorecard\n\nPoints: Athlone 2, PlayStation/Yello Pockets 2\n\nGame 2 (10:15)\n\nClub Greenheights v Kelvin Stars (Kingfisher Blue 1)\n\nWin: 0 wickets\n\nKelvin Stars won the toss and elected to bat\n\nKelvin Stars - 131/6 in 20 overs\n\nAsela v Johnson 7 runs\n\nSmith v Shakoor 5 runs\n\nPaliwal v Hallahan 4 runs\n\nDipoo v Dabhi 4 runs\n\nFernando v Dabhi 3 runs\n\nJo v Rajoo 9 runs\n\nStrydom v Kumar 10 runs\n\nGoswami v Basheer 1 run\n\nScorecard\n\nClub Greenheights - 76/6 in 12 overs\n\nVusumuzi v Smith 16 runs\n\nAubrey v Jo 6 runs\n\nAlen v Fernando 3 runs\n\nChase v Dabhi 3 runs\n\nHaque v Johnson 3 runs\n\nNjoo v Fernando 4 runs\n\nAhmed v Dabhi 1 run\n\nSheehy v Gupta 3 runs\n\nScorecard\n\nPoints: Kelvin Stars 2, Club Greenheights 2\n\nGame 3 (11:00)\n\nClub Greenheights v Athlone (Kingfisher Blue 2)\n\nWin: 3 wickets\n\nAthlone won the toss and elected to bat\n\nAthlone - 90/7 in 10 overs\n\nAwal v Chamuhiya 2 runs\n\nSeqib v Chase 5 runs\n\nAhmed v Njoo 1 run\n\nFarook v Sheehy 0 runs\n\nAhmed v Basheer 3 runs\n\nShahir v Singh 1 run\n\nFernando v Dabhi 2 runs\n\nMubarak v Kumar 0 runs\n\nScorecard\n\nClub Greenheights - 92/3 in 9.2 overs\n\nVusumuzi v Dabhi 9 runs\n\nAubrey v Kumar 7 runs\n\nChase v Seqib 1 run\n\nAhmed v Fernando 3 runs\n\nKeshi v Gupta 5 runs\n\nNjoo v Patel 2 runs\n\nSheehy v Rajoo 3 runs\n\nHaque v Ahmed 1 run\n\nScorecard\n\nPoints: Athlone 2, Club Greenheights 2\n\nGame 4 (12:00)\n\nPlayStation/Yello Pockets v Kelvin Stars (Kingfisher Blue 1)\n\nWin: 3 wickets\n\nKelvin Stars won the toss and elected to bat\n\nKelvin Stars - 72/10 in 10 overs\n\nDabhi v Sheehy 0 runs\n\nSmith v Hasan 5 runs\n\nPaliwal v Rahman 1 run\n\nFernando v Hasan 5 runs\n\nJo v Chase 2 runs\n\nDipoo v Njoo 1 run\n\nDabhi v Shaheed 5 runs\n\nSmith v Shaheed 3 runs\n\nScorecard\n\nPlayStation/Yello Pockets - 76/4 in 8.4 overs\n\nKhawu v Dabhi 10 runs\n\nChamuhiya v Smith 7 runs\n\nAhli v Dabhi 2 runs\n\nRyan v Dipoo 4 runs\n\nSantlal v Shaheed 7 runs\n\nQalbya v Fernando 5 runs\n\nTiaan v Jo 1 run\n\nKole v Fernando 2 runs\n\nScorecard\n\nPoints: Kelvin Stars 2, PlayStation/Yello Pockets 2\n\nGame 5 (13:00)\n\nPlayStation/Yello Pockets v Kelvin Stars (Kingfisher Blue 2)\n\nWin: 0 wickets\n\nKelvin Stars won the toss and elected to bat\n\nKelvin Stars - 95/7 in 10 overs\n\nAhmed v Chamuhiya 0 runs\n\nFarook v Aucho 3 runs\n\nAhmed v Aucho 5 runs\n\nShahir v Chamuhiya 5 runs\n\nFernando v Aucho 2 runs\n\nMubarak v Hasan 5 runs\n\nSmith v Hasan 0 runs\n\nDabhi v Njoo 1 run\n\nScorecard\n\nPlayStation/Yello Pockets - 98/5 in 8.4 overs\n\nKhawu v Ahmed 3 runs\n\nChamuhiya v Farook 7 runs\n\nAhli v Hasan 7 runs\n\nRyan v Shaheed 3 runs\n\nSantlal v Fernando 3 runs\n\nQalbya v Shaheed 7 runs\n\nTiaan v Dabhi 3 runs\n\nKole v Dabhi 3 runs\n\nScorecard\n\nPoints: Kelvin Stars 2, PlayStation/Yello Pockets 2\n\nGame 6 (14:00)\n\nClub Greenheights v Athlone (Kingfisher Blue 2)\n\nWin: 3 wickets\n\nAthlone won the toss and elected to bat\n\nAthlone - 89/9 in 10 overs\n\nAwal v Shaheed 3 runs\n\nAhmed v Dabhi 1 run\n\nShahir v Rajoo 3 runs\n\nFarook v Shaheed 2 runs\n\nAhmed v Patel 3 runs]" time="3.217"><properties><property name="score" value="0.03418537974357605" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Andr\xe9 Le N\xf4tre \u2013 Andr\xe9 Le N\xf4tre, also known as Le N\xf4tre, was a landscape architect, responsible for the design of the gardens of many of the most famous French ch\xe2teaux of the 17th century, including Versailles. He is popularly known as the father of the profession and he was born in Paris, the son of Pierre Le N\xf4tre, a painter of miniatures and Marie Le N\xf4tre. He was trained in his fathers occupation, he did some paintings by the time he was sixteen. He travelled to Italy with Pierre Berchet in 1640, and worked with the architect Jacques Lemercier, from 1646 he was occupied with architectural drawing for the Louvre. He was made a member of the Acad\xe9mie at Mans in 1650, Le N\xf4tres first known landscape project was at the ch\xe2teau of Vaux-le-Vicomte, for his distant relative, Nicolas Fouquet. Here he was more open to the imagination, transforming the plan into a Picturesque and Romantic garden. His work showed a tendency towards the natural and asymmetrical, Le N\xf4tre was unable to visit Italy on account of wars with the Italian states. His influence was limited to parts of the northeast of France, Le N\xf4tres first projects in this style were done for the court of the prince de Cond\xe9. The ch\xe2teau of Chantilly, belonging to the princess of Cond\xe9, Fran\xe7ois, Le N\xf4tre achieved the most at the Ch\xe2teau de Saint-Germain-en-Laye, for Louis XIV and his duc de Bourgogne, later the duc dOrl\xe9ans. Le N\xf4tre created a estate at Marly, and, at the Trianon, for the young Louis XIV. His work transformed the old fort into a palace, more suitable for the residence of the king. He is also responsible for the landscapes of the gardens of the Ch\xe2teau de Maisons and Ch\xe2teau de Saint-Cloud, Le N\xf4tre died in 1681, and was succeeded by his son Jean-Charles Le N\xf4tre, to whom he had taught his encyclopedic style, and his son-in-law Fran\xe7ois Dunan. The design of the gardens of the Ch\xe2teau de Marly was a project of Le N\xf4tres. They were designed to be seen from the top of the terrace, the perspectives and size of the gardens were calculated to surprise the spectators. Le N\xf4tre planned the gardens following the Renaissance principles of regularity, symmetry and he was also responsible for the modifications made to the original plan, done by Charles Le Brun and the gardener Martin Charbonnier, who remade the terrace and the parterres. The final result was the ancestor of modern formal gardens, in 1660 and 1661, Le N\xf4tre built two mansions on the Place Royale in Paris, for the jewellers Louis de Lancy and Balthazar Claquin. In 1664, Le N\xf4tre and Claude Perrault collaborated on a project for the west side of the Louvre. Le N\xf4tres last landscape was the garden around the ch\xe2teau de Sceaux, a project which was incomplete at the time of his death in 1681\n\n9. Gardens of Versailles \u2013 The Gardens of Versailles occupy part of what was once the Domaine royal de Versailles, the royal demesne of the ch\xe2teau of Versailles. Situated to the west of the palace, the gardens cover some 800 hectares of land, in addition to the meticulous manicured lawns, parterres of flowers, and sculptures are the fountains, which are located throughout the garden. Dating from the time of Louis XIV and still using much of the network of hydraulics as was used during the Ancien R\xe9gime. On weekends from spring to early autumn, the administration of the museum sponsors the Grandes Eaux \u2013 spectacles during which all the fountains in the gardens are in full play. In 1979, the gardens along with the ch\xe2teau were inscribed on the UNESCO World Heritage List, one of thirty-one such designations in France. With Louis XIII\u2019s final purchase of lands from Jean-Fran\xe7ois de Gondi in 1632 and his assumption of the role of Versailles in the 1630s. Records indicate that late in the decade Claude Mollet and Hilaire Masson designed the gardens and this early layout, which has survived in the so-called Du Bus plan of c.1662, shows an established topography along which lines of the gardens evolved. This is evidenced in the definition of the main east-west. From this point forward, the expansion of the gardens of Versailles followed the expansions of the ch\xe2teau, accordingly, Louis XIVs building campaigns apply to the gardens as well. At every stage the prescribed tour was managed, under the Sun Kings directions. First building campaign In 1662, minor modifications to the ch\xe2teau were undertaken, existing bosquets and parterres were expanded and new ones created. Most significant among the creations at this time were the Orangerie, the Versailles Orangery, which was designed by Louis Le Vau, was located south of the ch\xe2teau, a situation that took advantage of the natural slope of the hill. It provided an area in which orange trees were kept during the winter months. The Grotte de Th\xe9tys, which was located to the north of the ch\xe2teau, formed part of the iconography of the ch\xe2teau, the grotto would be completed during the second building campaign. By 1664, the gardens had evolved to the point that Louis XIV inaugurated the gardens with the f\xeate galante called Les Plaisirs de l\u2019\xcele Enchant\xe9e. The event, which officially was to celebrate his mother, Anne d\u2019Autriche, guests were regaled with fabulous entertainments in the gardens over a period of one week. As a result of this f\xeate \u2013 particularly the lack of housing for guests, Louis realized the shortcomings of Versailles and began to expand the ch\xe2teau and the gardens once again. With this new phase of construction, the gardens assumed the topographical and iconological design vocabulary that would remain in force until the 18th century. \u201d\n\n10. French language \u2013 French is a Romance language of the Indo-European family. It descended from the Vulgar Latin of the Roman Empire, as did all Romance languages, French has evolved from Gallo-Romance, the spoken Latin in Gaul, and more specifically in Northern Gaul. Its closest relatives are the other langues do\xefl\u2014languages historically spoken in northern France and in southern Belgium, French was also influenced by native Celtic languages of Northern Roman Gaul like Gallia Belgica and by the Frankish language of the post-Roman Frankish invaders. Today, owing to Frances past overseas expansion, there are numerous French-based creole languages, a French-speaking person or nation may be referred to as Francophone in both English and French. French is a language in 29 countries, most of which are members of la francophonie. As of 2015, 40% of the population is in Europe, 35% in sub-Saharan Africa, 15% in North Africa and the Middle East, 8% in the Americas. French is the fourth-most widely spoken mother tongue in the European Union, 1/5 of Europeans who do not have French as a mother tongue speak French as a second language. As a result of French and Belgian colonialism from the 17th and 18th century onward, French was introduced to new territories in the Americas, Africa, most second-language speakers reside in Francophone Africa, in particular Gabon, Algeria, Mauritius, Senegal and Ivory Coast. In 2015, French was estimated to have 77 to 110 million native speakers, approximately 274 million people are able to speak the language. The Organisation internationale de la Francophonie estimates 700 million by 2050, in 2011, Bloomberg Businessweek ranked French the third most useful language for business, after English and Standard Mandarin Chinese. Under the Constitution of France, French has been the language of the Republic since 1992. France mandates the use of French in official government publications, public education except in specific cases, French is one of the four official languages of Switzerland and is spoken in the western part of Switzerland called Romandie, of which Geneva is the largest city. French is the language of about 23% of the Swiss population. French is also a language of Luxembourg, Monaco, and Aosta Valley, while French dialects remain spoken by minorities on the Channel Islands. A plurality of the worlds French-speaking population lives in Africa and this number does not include the people living in non-Francophone African countries who have learned French as a foreign language. Due to the rise of French in Africa, the total French-speaking population worldwide is expected to reach 700 million people in 2050, French is the fastest growing language on the continent. French is mostly a language in Africa, but it has become a first language in some urban areas, such as the region of Abidjan, Ivory Coast and in Libreville. There is not a single African French, but multiple forms that diverged through contact with various indigenous African languages, sub-Saharan Africa is the region where the French language is most likely to expand, because of the expansion of education and rapid population growth\n\n11. Garden \u2013 A garden is a planned space, usually outdoors, set aside for the display, cultivation, and enjoyment of plants and other forms of nature. The garden can incorporate both natural and man-made materials, the most common form today is known as a residential garden, but the term garden has traditionally been a more general one. Zoos,]" time="5.036"><properties><property name="score" value="0.7297635475794474" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.72976355&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.72976355
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[SINGAPORE - The Central Narcotics Bureau (CNB) said on Monday (June 19) that it had seized more than 80,000 methamphetamine pills worth an estimated street value of more than $3.1 million during two separate operations.\n\nIn a Facebook post on Monday, CNB said that the drugs had been seized at Tuas Checkpoint on May 24.\n\nIn the first operation, CNB officers were deployed at the checkpoint when a Singaporean male traveller was found to be acting suspiciously at about 9.15pm.\n\n&quot;When challenged, the traveller was found to be concealing two plastic packets inside his rectum,&quot; CNB said. &quot;Upon further search of his vehicle, officers found an additional three packets containing a total of 40,160 meth tablets hidden inside a sling bag.&quot;\n\nCNB officers arrested the 33-year-old man. He has been assessed and found to be mentally sound. He will be charged in court on Tuesday (June 20).\n\nIn a second operation that took place at about 11.45pm that day, CNB officers detected two suspicious-looking men acting suspiciously at a pedestrian walkway near the Customs, Immigration and Quarantine (CIQ) complex.\n\nAfter conducting a search, CNB officers found a total of 17,280 meth tablets hidden in the roof lining of the car.\n\nThe two men, aged 34 and 50, are Singaporeans.\n\n&quot;Investigations are ongoing,&quot; CNB said.]" time="0.765"><properties><property name="score" value="0.10113489627838135" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Hello there! If you are new here, you might want to subscribe to the RSS feed for updates on this topic.\n\nAurorawe\n\nWith all the police man killings in the news lately it\u2019s a good thing we have the security of a president who is in complete control of his body and doesn\u2019t have to worry about a trigger happy cop.\n\nBenjamin Smith II, 34, shot dead a 34-year-old black man and a 37-year-old Hispanic man on Friday.\n\nA police chief says a gunman had told him he was \u201cupset about Black Lives Matter\u201d, and that he was targeting white people in a series of random shootings that struck fear in the heart of Chicago.\n\nThe gunman killed three people and injured 16 others, all of them apparently chosen at random, before being shot dead by police, according to authorities.\n\nPeople spoke of their terror after the shootings at various sites in the city\u2019s downtown area on Friday evening.\n\nHe shot and killed a man on the street, a train station and a fast food restaurant \u2013 his final two victims were shot in front of a Chicago police station.\n\nThe suspect was identified as a white man in his 30s, according to the Chicago police superintendent, Eddie Johnson, who said he had no information about the gunman\u2019s background.\n\nThe suspect had no criminal record, was known to police only for minor traffic violations and was not on the radar of federal law enforcement, a law enforcement official told Reuters.\n\nPolice found a weapon in the man\u2019s car, Johnson said, adding that a bomb squad had also been called to the scene.\n\nEarlier, Mayor Rahm Emanuel said in a statement: \u201cAny act of violence on a police officer is an attack on our entire community, and requires a swift and thorough investigation.\u201d\n\nWitnesses described scenes of chaos as people scattered for cover after hearing gunshots and watching police run towards the shooter, who they described as a white man.\n\n\u201cI was working, the next thing I know all these people are running and I heard gun shots. I heard at least five,\u201d said Elijah Rodriguez, 17, who works near the shooting scene.\n\nIn the chaos of the moment, a woman, also described as being in her 30s, and a man, were wounded as bystanders rushed to escape the gunfire, Johnson said.\n\nAn officer who arrived at the scene was shot in the head at point blank range, but his vest saved his life, Johnson said.\n\n(click for source)]" time="1.278"><properties><property name="score" value="0.8459406693776449" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Still loading...\n\nA B C D E F G H I J K L M N O P Q R S T U V W X Y Z AA AB AC AD AE AF AG AH AI AJ AK AL AM AN AO AP AQ AR AS AT AU AV AW AX AY AZ BA BB BC BD BE BF BG BH BI BJ BK BL BM BN BO BP BQ BR BS BT BU BV BW BX BY BZ CA CB CC CD CE CF CG CH CI CJ CK CL CM CN CO CP CQ CR CS CT CU CV 1 Eintr\xe4ge bis zum 20.04.2015 \u2013 l\xf6schen bitte bei neuen Eintr\xe4gen 2 Vorname (allg.) Nachname (allg.) Verein / Firma E-Mail-Adresse Anzahl der Tickets/Ausgaben wie viel \u20ac pro Ticket? f\xfcr welche Spiele 3 Tim Wollinger Tim (TIM) Wollinger Kr\xfcger U. Wiesner GmbH u. Co. KG wollinger@krueger-wiesner.de 2 x 1. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 16,00 4 Benedikt Bauck Benedikt Bauck 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 5 Heiko Thelen Heiko Thelen 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 6 Lars Kirchner Lars Kirchner 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 7 Joachim Sch\xf6nfeld Joachim Sch\xf6nfeld 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 8 Ingo Zillmer Ingo Zillmer 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 9 Gunnar Schneider Gunnar Schneider 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 10 Heiko Nitschke Heiko Nitschke 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 11 Sebastian Thomas Sebastian Thomas 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 12 Thomas Arnold Thomas Arnold 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 13 Michael Gr\xe4f Michael Gr\xe4f 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 14 Thomas Lehn Thomas Lehn 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 15 Matthias Wojtalla Matthias Wojtalla 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 16 Max Schaap Max Schaap 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 17 Daniel Senkbeil Daniel Senkbeil 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 18 Mark B\xf6hm Mark B\xf6hm 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 19 Christian Hanusch Christian Hanusch 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 20 Stefan B\xf6lger Stefan B\xf6lger 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 21 Markus Vogl Markus Vogl 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 22 Mario Jankowski Mario Jankowski 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 23 Jens Gredele Jens Gredele 1 x 2. Reihe Stehpl\xe4tze plus Ticket auf allen Heim- und Ausw\xe4rtsspielen im Schnitt 40,00 24 Thomas Kohler Thomas Kohler 1]" time="3.166"><properties><property name="score" value="0.521485224366188" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.52148522&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.52148522
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[SOLD\n\nLot #: 1311\n\n-Chevrolet Bel Air, 1955-1956, Restored. 250 HP, 283 C.I. 4-Barrel Engine, 2-Speed Automatic Transmission, 12 Bolt Rear End, 2-Wheel Power Drum Brakes. White/Blue Interior. Very Well Restored in 2001. Beautifully Reupholstered White Bench Seat Interior with Door Panels, Upholstered Headliner and Side Curtains. Glass Pack Dual Exhaust System.]" time="0.260"><properties><property name="score" value="0.9375905990600586" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.9375906&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.9375906
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[We've had a fire with smoke smell that has not disappeared. The contractors worked hard to solve the problem, but unfortunately, it was a difficult case. We ended up spending more than initially planned and I can't say it was money well spent. I am disappointed. The one problem was that the company was rather unresponsive and I had to follow up more than I would have liked. I would hire again.]" time="0.239"><properties><property name="score" value="0.9904504418373108" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Pueblo County and a water provider say they have come to a \u201cmutual understanding\u201d on a controversial decision last month that curbed water supply to a community south of Pueblo.\n\nLast month, the Colorado Springs Water Works announced it would cut water supply to two water districts that buy its water because the Pueblo Reservoir was too low to pump water into the Pueblo County\u2019s Jimmie Creek Reservoir.\n\nThat reservoir, which supplies water to about 20,000 people south of Pueblo, had been on a rotating \u201cwater conservation\u201d restriction for more than a year.\n\nJim May, manager of the CSWW, said the restrictions were intended to help the reservoir fill faster, but the cuts to the districts south of Pueblo appeared to exacerbate the problem.\n\nAfter the announcement last month, residents and water officials in Pueblo and Pueblo County wrote to the state saying they believed the restrictions were discriminatory. They argued that residents in the water districts that purchase the water were being punished for water overuse by residents of Pueblo and Pueblo County, even though they don\u2019t buy the water.\n\nThe Pueblo County commissioners sent a letter to CSWW last month that said, in part: \u201cWe do not consider that the Pueblo Reservoir can be drained without impact to the water service area, and it is our position that this is discriminatory.\u201d\n\nThe commissioners sent another letter Tuesday saying the county is committed to working with the Colorado Springs Water Works on this issue and future water issues.\n\nCSWW agreed to issue a second letter to water districts in Pueblo and Pueblo County clarifying that water would be available.\n\nThe districts that had been receiving water have already begun taking the water again, said Dennis Tenorio, general manager of the Northern Water Users Association. He said the water districts bought a $200,000 storage tank to hold water when there is a shortage.\n\nMay said CSWW has worked with the districts to understand why the Pueblo Reservoir\u2019s level is low and how to address the problem.\n\n\u201cThere\u2019s a huge, huge amount of water being used up there,\u201d May said. \u201cBut, with this mutual understanding, I\u2019m confident we\u2019ll be able to make this a long-term fix for everyone.\u201d\n\nTim Hoover: 303-954-1626, thoover@denverpost.com or twitter.com/timhoover]" time="1.236"><properties><property name="score" value="0.8671181201934814" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The layout is interesting, and the selection of buttons and values are nicely suited for a steampunk background.\n\nThe mix of cold blue and warmer browns works well, and overall I think it would be very effective on a phone.]" time="0.189"><properties><property name="score" value="0.9991303086280823" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[BAGHDAD \u2014 Prime Minister Haider al-Abadi on Monday presented a proposed new Cabinet to parliament for approval, which is expected to include a defense minister and an interior minister.\n\nAbadi said in a statement that a number of the ministers would remain in their posts, while some would get new assignments and one would be replaced.\n\nThe move would boost his government\u2019s credibility as it takes on the Islamic State group, which swept across much of northern and western Iraq last summer. It would also help the prime minister gain influence with a parliament elected in April.\n\nLawmakers met Monday afternoon in a closed session to hear from the prime minister and approve the new Cabinet.\n\nIraqi lawmakers already have approved acting ministers in the foreign and defense portfolios in anticipation of a new Cabinet.\n\nThe proposed new defense minister is Khaled al-Obeidi, the chief of staff of the armed forces.\n\nThe candidate for the interior portfolio is Mohammed al-Ghabban, an official in the predominantly Shiite paramilitary forces known as the Popular Mobilization Forces. The forces were created last summer after Iraq\u2019s most powerful Shiite cleric, Grand Ayatollah Ali al-Sistani, issued a fatwa, or religious decree, calling for Iraqis to fight the extremists.\n\nThe new Cabinet members will have to undergo further screening before they can be sworn in.\n\nAn aide to the prime minister told The Associated Press that the interior minister would replace Mohammed Salem al-Ghabban, who has been acting minister since March. The aide spoke on condition of anonymity because he was not authorized to speak to the media.\n\nThe process of forming the Cabinet has been drawn out, with delays because of disputes between Shiite, Sunni and Kurdish blocs over who would hold key posts.\n\nIt was not clear if the interior minister would be a Sunni or a Shiite. It also wasn\u2019t clear if the defense minister would be a Sunni or a Shiite.\n\nWhen the new Cabinet is sworn in, the president will be able to name a new Cabinet as demanded by the constitution, giving him greater leverage in the new legislature.\n\nAbadi, a Shiite, has struggled to overcome his dependence on political parties, some of which receive support from neighboring Iran.\n\nDuring the U.S. occupation, he served as minister of communications under the Shiite-led government of then-Prime Minister Nouri al-Maliki. He was tapped as prime minister last summer after his predecessor was ousted amid an Islamic State offensive that overran about a third of the country.\n\nAbadi promised to unite Iraq, which is divided among Sunnis, Shiites and Kurds, but has struggled to overcome his dependence on political parties, some of which receive support from neighboring Iran.\n\nThe State Department in Washington announced the appointment Monday of Stuart Jones, a career diplomat, as the new U.S. ambassador to Iraq, replacing Douglas Silliman.]" time="1.409"><properties><property name="score" value="0.21092957258224487" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.21092957&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.21092957
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[November 11th, 2012\n\nWhile we\u2019re waiting for the official Apple iPod nano 2G Review to come, I wanted to share my review and give you my thoughts. I\u2019ve now had the second generation of the iPod nano for about a week and wanted to give you my thoughts about it.\n\n(Last Updated: Nov 12th, 2012)\n\nThe Physical Design\n\nI think Apple got this design right with the nano. It\u2019s very thin and light, which I like. It\u2019s almost identical to the 1st gen nano, but thinner, and with a larger screen. The big deal for me is that it is so much thinner and smaller than the previous nano. I don\u2019t know why I like that so much, but it does make it so much easier to carry around.\n\nThe Screen\n\nI was initially concerned about the small size of the screen. My fear was that I wouldn\u2019t be able to see the screen as well, or that it would be harder to navigate the menus. I\u2019m very happy to report that the screen size is perfect for what I want. I think the new screen is just large enough to easily navigate menus and change songs.\n\nThe Sound Quality\n\nI was very skeptical about the new sound quality. I don\u2019t have a lot of time to listen to music, and when I do, I usually do it while at the gym. So for me, sound quality is not a huge priority, but I was concerned about it with the new nano.\n\nWell, I am very happy to say that the sound quality is much better than the first generation nano. It is not better than my 4th gen iPod touch, but it is about the same, and sounds really good to me.\n\nThe Battery Life\n\nThe battery life of the new nano is what I was most worried about. The previous nano only got about 5 hours of battery life when playing music, and this is the same battery as the first generation nano. But I was pleasantly surprised by the new nano. It only lasts about 8 hours, but that\u2019s fine for me. My goal is to carry a nano that can give me 8 hours of listening time at the gym. The only time I need longer than that is when I\u2019m traveling for work, and I have my iPhone with me for music.\n\nThe Software\n\nI\u2019m not going to go into too much detail here, but I am very happy with the new software. I think the new interface looks great and I love that I can easily make the album art fit the screen perfectly.\n\nConclusion\n\nOverall, I\u2019m very happy with the new iPod nano 2G. It is the perfect size and weight for me. It sounds good, and the battery life is good enough for me. It\u2019s nice to have an iPod that\u2019s small enough to fit in my pocket.\n\nI would definitely recommend the new iPod nano. If you already have a 1st gen nano, the only reason to upgrade is for the size.\n\nOne thing to keep in mind, is that Apple has said that]" time="1.582"><properties><property name="score" value="0.9997828205426534" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Advanced Biological Warfare (ABW)\n\nThere is an urgent need for &quot;offensive&quot; biological warfare (BW) capability, to complement defensive measures against an attack with biological weapons (BW). Existing defensive measures provide only limited protection against BW, and their usefulness would be significantly reduced in some circumstances. No acceptable defenses have been developed against BW agents.\n\nThis lack of adequate defenses against biological weapons is a major vulnerability in the security of the United States. The increasing danger that biological weapons will be used for hostile purposes poses an urgent need to counter the threat.\n\nModern biological warfare could include development and use of pathogenic organisms and toxins, as well as incapacitating agents. The former might well be the only BW agent of concern because of their potential for causing mass casualties.\n\nThe widespread use of modern technology has increased the availability of biological materials for use as BW agents. Current medical knowledge has increased the potential to create and spread biological weapons. This trend is expected to continue.\n\nEarly development and testing of a pathogenic agent for use in an offensive BW program would not be easily detectable. A biological agent can be disseminated in many ways and to many places with little warning. It is difficult to effectively monitor, detect, and decontaminate a wide area after an attack.\n\nStrategic Offensive Biological Warfare\n\nStrategic offensive BW provides the United States with a major deterrent against potential adversaries. As a deterrent weapon, strategic offensive BW can have a significant impact on the overall military posture of the United States and its potential adversaries. It is in the national interest of the United States to develop a military biological capability to deter an adversary from using biological weapons.\n\nStrategic offensive BW will be used against enemy personnel or facilities to inflict maximum casualties, degrade military capability, or destroy life-sustaining resources or capabilities. In military operations, strategic offensive BW would be used as part of a coordinated effort, in conjunction with other conventional and nuclear forces, and it would require integration of tactical, operational, and strategic capabilities.\n\nStrategic offensive BW could be employed by the United States to complement and support its defensive and non-military (e.g., diplomatic, political, economic, and legal) measures in the event of a biological attack. Its primary purpose would be to minimize the impact of a biological attack on the United States or its forces. Strategic offensive BW may also have applications in countering adversaries that are seeking to develop and maintain a BW capability.\n\nNon-Strategic Offensive Biological Warfare\n\nNon-strategic offensive BW may be employed in support of specific military operations against specific adversary targets. It may also be used in support of other military operations. It would be used to complement and support defensive and non-military measures in the event of a biological attack.\n\nNon-strategic offensive BW would be limited in scope and duration, but it would still be considered offensive because it would result in the infliction of casualties. Non-strategic offensive BW would be used against specific targets, rather than against a large area. It would normally be employed in peacetime as a form of anticipatory self-defense, rather than as a weapon of last resort. Non-strategic offensive BW could be conducted clandestinely to avoid a general war.\n\nThe possession of non-strategic offensive BW capabilities would enhance the deterrent posture of the United States. It would enhance US options to counter biological threats. These options include military action and other non-military actions, which could preclude the need for a military response.\n\nThe National Military Strategy (NMS) endorses non-strategic offensive BW as a credible deterrent to aggression against US interests. The NMS identifies two levels of non-strategic offensive BW for deterrence: A highly credible threat to use BW agents in response to biological attacks or a threat to use BW agents in response to a biological attack. The NMS also identifies a low-level but still credible threat to use BW agents to counter limited biological attacks, sabotage, and hostage-taking.\n\nGlobal Issues\n\nThe US BW program is important to national security and international stability. A fundamental premise of US policy is that offensive BW capabilities are legitimate instruments of national power, especially if the United States is threatened by chemical or biological weapons. The United States has an interest in dissuading the proliferation of BW programs. The development of offensive capabilities by countries without chemical or biological weapons programs or doctrines would be destabilizing and increase the likelihood of their use. The growth of indigenous BW programs in countries with other WMD programs would have similar destabilizing effects.\n\nInternational stability is jeopardized when there is uncertainty as to whether nations will abide by international legal commitments regarding BW. The US BW program is not intended to influence foreign countries, but to demonstrate the US commitment to deter the use of biological weapons and to respond if deterrence fails. BW will be maintained as a credible deterrent, but not one that would have to be used. The United States should make clear its continued intention to pursue an effective BW program to deter adversaries, if necessary.\n\nUS non-strategic offensive BW is an important element of national power and US leadership in arms control. The US BW program plays a crucial role in US leadership in arms control. Non-strategic offensive BW is necessary for full compliance with existing and emerging arms control and non-proliferation agreements. It is important for successful implementation of the CWC, the BWC, and other regional and international agreements. This would prevent the proliferation of BW programs in countries with other WMD programs.\n\nInterdiction\n\nThe inherent characteristics of biological agents make them suitable for clandestine delivery and easy to disseminate. Biological agents could be delivered clandestinely to specific personnel or facilities in areas under control of a potential adversary, without risk to the agents\u2019 handlers.\n\nThe ability of an adversary to clandestinely deliver a biological agent makes the effectiveness of a traditional containment strategy, such as is used in the management of nuclear, chemical, and missile proliferation, problematic. A defensive strategy for biological weapons will have limited utility.\n\nInterdiction of a clandestine BW attack would require significant intelligence resources and considerable time. It would be difficult to find the attacking agent, assess the magnitude of the attack, and identify the source of the attack. If it is found, the agents are difficult to detect and are difficult to destroy. In the unlikely event that a clandestine BW attack were detected and stopped before any damage occurred, the United States would have to consider the political and military implications of conducting an extensive search for the source of the BW attack.\n\nIf an adversary does attempt to use BW, the United States would have to consider the implications of an extensive search for the source of the attack, whether it should be attempted, and what political and military actions might result. US policy is to discourage development of BW programs in countries without other WMD programs, but to take BW programs in friendly countries into account. Therefore, the United States would have to consider how the public and international community would react to the use of US resources to search for a biological attack.\n\nBiological Weapons Control\n\nUnited States policy is to discourage countries from developing and using BW and to discourage the proliferation of BW.\n\nThe United States wants to minimize the risk of WMD proliferation and to prevent or mitigate the harmful consequences of proliferation. Because BW programs are easier to develop and conceal than are nuclear, chemical, and missile programs, US policy is to take appropriate measures to minimize the risk of proliferation of BW. In the case of US friends and allies, it will consider the importance of the BW program to the defense of the country and its overall relations with that country.\n\nBiological weapons could be used in conjunction with nuclear, chemical, or other weapons. This possibility increases the potential threat posed by BW. The United States has concluded that the worldwide prohibition of BW would not eliminate the threat of their use. Although the worldwide elimination of BW is a long-term goal, there is no realistic means to reach it in the foreseeable future. The United States will continue to work toward that goal, but in the meantime it will maintain an effective BW deterrent.\n\nFormal arms control, including international verification of BW programs, would not be sufficient to prevent the clandestine development and use of BW. Biological weapons can be developed and used without a massive infrastructure. Interdiction of clandestine BW programs may not be successful. A country with both BW and chemical weapons programs would be able to attack with biological weapons without any risk of exposure of its chemical weapons program. The United States will seek to work with friendly countries to prevent proliferation of BW.\n\nThe United States would oppose a global ban on the development, possession, or use of BW because it is not a realistic possibility. Such a ban would not eliminate the threat of BW. It would be easy to evade a ban and difficult to enforce.\n\nThe United States will continue to work with the international community to reach an effective and acceptable solution to the BW problem.\n\nTechnical Issues\n\nThe United States does not want to be vulnerable to BW, and it must be able to meet the challenge of a biological attack. However, the US BW program will not be capable of protecting the United States against a chemical or nuclear attack, nor will it be able to protect against all types of biological weapons.\n\nThere is no defense against BW. The United States does not want to be vulnerable to BW. However, if an adversary did use BW, the United States would be willing to use its offensive BW capability to counter it.\n\nThere are no weapons systems or defensive measures capable of protecting against all types of biological warfare. However, a significant deterrent can be developed to dissuade potential adversaries from using BW. This would require BW programs to develop and maintain a viable capability to inflict casualties, degradation of military capability, or destruction of life-sustaining resources or capabilities.\n\nThe likelihood that a biological attack would be effective is low. The impact of biological weapons would be low compared to that of nuclear, chemical]" time="4.962"><properties><property name="score" value="0.813485942103646" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Punto Fijo Municipality\n\nPunto Fijo is one of the 21 municipalities (municipios) of the state of Yaracuy, Venezuela. The municipality's capital is the town of Punto Fijo.\n\nThe municipality is one of several in Venezuela named &quot;Punto Fijo Municipality&quot; in honour of independence hero Antonio Jos\xe9 de Sucre.\n\nThe municipality's area is , and the population is 88,115 (2010 est.).\n\nThe municipality is the site of the Bicentennial Aqueduct.\n]" time="0.296"><properties><property name="score" value="0.9948403239250183" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99484032&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99484032
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[University of Notre Dame Australia\n\nThe University of Notre Dame Australia (formerly Notre Dame University) is an Australian Roman Catholic university established in 1989. The main campus is located at the northern end of the Monash Freeway in Craigieburn, 25 kilometres north-west of the Melbourne central business district in Victoria, Australia. The university has approximately 1,800 students.\n\nNotre Dame Australia is a tertiary institution of the Congregation of Holy Cross, a religious community of priests and brothers founded in 1837 in Le Mans, France by Father Pierre Coudrin. The Congregation of Holy Cross was formally recognised by Pope Leo XIII in 1885. In Australia the Congregation of Holy Cross established the predecessor of Notre Dame University, the Holy Cross College, at Box Hill, Victoria in 1951. Holy Cross College was incorporated as a company in 1978 and was approved by the Victorian Government as a Catholic University College in 1981.\n\nIn January 1989 the Holy Cross College and the St. Columba's Church in Bulla moved to new premises in Craigieburn. At that time the college was renamed as the Notre Dame Institute of the Arts and the College of Commerce, Science and Technology.\n\nIn 1991 the Notre Dame Institute of the Arts, the Notre Dame Institute of the Arts (Interim), the College of Commerce, Science and Technology and the Holy Cross Seminary amalgamated to become a university college.\n\nIn 1994 the Notre Dame College of Education and the College of Science amalgamated to become the Notre Dame College of Education and Science.\n\nIn 1996 the Notre Dame Institute of the Arts was amalgamated with the College of Education and Science.\n\nIn 1997 the University of Notre Dame Australia was formally established.\n\nIn 2003 the university changed its name from the University of Notre Dame Australia to Notre Dame University Australia. The change in name reflected the development of the university's research and postgraduate research and development capabilities.\n\nIn 2005 the university changed its name back to the University of Notre Dame Australia.\n\nThe university's crest was designed by Michael Malone of Pentagram, Melbourne, and is based on the &quot;Cross of the Holy Spirit&quot;, which was designed by Leonard McCoy of the University of Notre Dame, in the United States.\n\nIn 2006 Notre Dame became the first university in Australia to gain ISO 14001 certification for its environmental management systems.\n\nIn 2008 the University of Notre Dame Australia took part in the international Interfaith Center of Beijing and in the International Symposium on the contribution of religions to sustainable development, held at Shandong University in Jinan, Shandong Province, China.\n\nThe University of Notre Dame Australia has two major sites: the Craigieburn campus and the Werribee Park campus.\n\nCraigieburn campus\n\nThe Craigieburn campus, located at the northern end of the Monash Freeway, houses the Schools of Architecture, Business, Communication and Social Sciences, Education, Health, Humanities and Law, Music, Science and Sport. The campus is also home to the University's administration and a student village.\n\nWerribee Park campus\n\nThe Werribee Park campus is home to the School of Education and Social Work. The campus is also the site of the University's Sport Centre.\n\nUniversity library\n\nThe University of Notre Dame Australia Library holds over 43,000 print volumes and over 2,000 CD-ROMs. The collection covers theology, philosophy, the arts, social sciences and sciences.\n\nThe library's periodical collection contains more than 15,000 current serials including complete runs of the major theological journals. It also holds more than 2,000 current audiovisual titles, 2,000 electronic books and 3,000 electronic journal articles. The library is fully networked to connect to the Internet and to link with all parts of the University, and all staff and students have access to its facilities. The library is a member of the Council of Regional Libraries in Victoria (CRLV) and of the National Library of Australia's Research Collections Network (RCLN).\n\nThe library also houses the university's Special Collections and the Archives, including the history of the Congregation of Holy Cross in Australia, Ireland, the United Kingdom and the United States, the history of the Holy Cross College (now the University of Notre Dame Australia), and personal papers of members of the Congregation of Holy Cross and its educational institutions. The Archives are open to the public.\n\nResearch centres and institutes\n\nThe university has research centres and institutes including the Australian Institute for Bioengineering and Nanotechnology (AIBN), the Australian Institute of Music, the National Institute for Education Development (NIEd), the Australian Institute for Indigenous Development, and the Aloysius JinEndean Lidcombe Catholic Library.\n\nThe university also has research alliances with other universities and institutions.\n\nIn 2011, the University of Notre Dame Australia was awarded the \u201cVice-Chancellor\u2019s Award for Excellence in Corporate Reporting\u201d by the Australasian Centre for Corporate Responsibility. In the same year the University of Notre Dame Australia received the second highest score in the Excellence in Learning category in the Australian Government's Excellence in Higher Education Award, a system for measuring and benchmarking universities in Australia.\n\nThe University of Notre Dame Australia has an undergraduate intake of around 1,500 students and an average intake of about 200 new postgraduate students a year.\n\nIn 2006 the University of Notre Dame Australia began offering graduate certificates and diplomas.\n\nIn 2011 the university introduced a new academic program that allowed its students to study in the United States of America.\n\nThe university has two residential colleges, Cormac and St. Brigid's.\n\nCormac College was opened in 2000, replacing the college that existed in the Werribee Park campus from 1997 to 2000. The College's name is derived from Cormac MacCarthy, the first Bishop of Melbourne, who was a brother of the Congregation of Holy Cross. The College was founded by the Rev. Brian Finn SC, a brother of the Congregation.\n\nSt. Brigid's College was opened in 2001. The College is named after St. Brigid of Kildare, who is the Patron Saint of Ireland. The College was founded by the Rev. Brian Finn SC, a brother of the Congregation.\n\nSt Brigid's and Cormac College form a community known as Cormac Brigid's. Students in the colleges form their own representative councils.\n\nThe University of Notre Dame Australia is a private university with four schools and one institute:\n\n\nThe university is governed by a board of governors, with the Bishop of the Diocese of Sandhurst as the president of the university.\n\n\n\n]" time="3.529"><properties><property name="score" value="0.9487294895308358" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Prays cingulata\n\nPrays cingulata is a moth of the family Plutellidae. It is found in Taiwan.\n\nThe wingspan is 15\u201317\xa0mm. The forewings are blackish, irrorated with pale fuscous. The basal area is overlaid with deep purple-fuscous, the edge evenly curved. There is a blackish streak, on the inner side of which is an oblique blackish dash. The area between the basal area and the base of the costa is deep purple-fuscous and is traversed by an indistinct oblique fuscous strigula. The first discal stigma is indistinct and the plical and second discal forming a transverse purple-fuscous line. There is a group of indistinct blackish dots around the posterior part of the costa and termen. The hindwings are light grey, in males suffused with whitish on the posterior half.\n\nThe larvae feed on &quot;Ficus superba&quot;. They mine the leaves of their host plant.\n]" time="0.545"><properties><property name="score" value="0.9979859590530396" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Bunny boiler\n\nA bunny boiler (sometimes bunny-boiler) is a person who is obsessed with another person to the point of stalking, harassment, or violence. It was popularised by a number of films in the 1990s including &quot;Fatal Attraction&quot;, in which Glenn Close's character Alex Forrest is known for boiling bunnies as part of her psychosis.\n\nThe film &quot;Fatal Attraction&quot; featured the villain Alex Forrest, who works as a photographer and engages in an affair with her married lover, Dan Gallagher. When she discovers that Dan is having second thoughts about the affair, she begins to stalk him and his family, which culminates in an attempted seduction of Dan's daughter, which causes Alex to be murdered by Dan.\n\nIn the 2010 video game &quot;Fatal Frame&quot;, the primary antagonist is a psychopath named Miu Hinasaki. During the course of the game, she kidnaps several young girls and holds them prisoner at her house, where she sadistically mutilates them. Throughout the game, it is implied that she has been doing this for a long time, and that many other girls have disappeared without a trace.\n\nThe trope was also used in an early episode of &quot;Glee&quot; when Rachel Berry was terrorised by a former classmate, Quinn Fabray. The term has been used to refer to women in many fictional works, and has been]" time="0.663"><properties><property name="score" value="0.1438104510307312" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[\n\nEdition: Dominaria\n\n\n\nSort:\n\nRare\n\nCommon\n\nUncommon\n\nCommon (2 versions)\n\nUncommon (2 versions)\n\nRare (2 versions)\n\nCommon (1 versions)\n\nUncommon (1 versions)\n\nRare (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nUncommon (1 versions)\n\nUncommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nUncommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nUncommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nUncommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nUncommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (1 versions)\n\nCommon (]" time="4.329"><properties><property name="score" value="0.9998179078102112" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Casa Italia (band)\n\nCasa Italia was a short-lived Italian band.\n\nCasa Italia was born in 1988 in Bologna as a project for the creation of music without any commercial purpose. The initial core of the band is formed by Daniele Chiodi, Franco Stalteri and Valerio Varesi, and the experience is enriched by other musicians who will be only present for live performances.\n\nIn the years from 1988 to 1995, Casa Italia published three full-length albums, &quot;Infinito in te&quot; (1991), &quot;Sono come sono&quot; (1992), &quot;Gli dei della gioia&quot; (1995), the last two of which were released by the MTM Music label.\n\n\n]" time="0.417"><properties><property name="score" value="0.9943723678588867" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Edward Ashworth (politician)\n\nEdward Ashworth (13 November 1794 \u2013 27 January 1879) was an English Whig politician.\n\nAshworth was elected Whig MP for Preston in 1847, but did not stand again in 1852.\n]" time="0.170"><properties><property name="score" value="0.9383976459503174" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.93839765&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.93839765
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Syntomodrillia concinna\n\nSyntomodrillia concinna is a species of sea snail, a marine gastropod mollusk in the family Drilliidae.\n\nThe length of the shell varies between 5.3\xa0mm and 9.3\xa0mm.\nThis marine species occurs off the Bahamas and the Florida Keys, USA; also off Colombia, Venezuela, Trinidad and Tobago, Surinam, Costa Rica and French Guiana\n\n]" time="0.221"><properties><property name="score" value="0.9691162109375" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.96911621&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.96911621
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Mike Patterson (Canadian football)\n\nMike Patterson (born April 10, 1965) is a former professional Canadian football defensive lineman who played 12 seasons in the Canadian Football League (CFL). He was drafted by the BC Lions in the second round of the 1987 CFL Draft and spent eight seasons with the Lions. He played college football at Western Illinois.\n]" time="0.165"><properties><property name="score" value="0.9927598237991333" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Olszyca, Greater Poland Voivodeship\n\nOlszyca () is a village in the administrative district of Gmina Budzy\u0144, within Chodzie\u017c County, Greater Poland Voivodeship, in west-central Poland. It lies approximately south-east of Chodzie\u017c and north of the regional capital Pozna\u0144.\n\n&lt;br&gt;\n]" time="0.224"><properties><property name="score" value="0.6599848866462708" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Charlie Higgs\n\nCharles Higgs (born 23 April 1871 \u2013 unknown) was an English professional footballer who played as a wing half. He played in the Football League for Glossop, Bolton Wanderers and Newton Heath. He also made appearances in the Scottish League for Heart of Midlothian, Motherwell, Celtic and Dunfermline Athletic.\n]" time="0.183"><properties><property name="score" value="0.6134288907051086" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Abraham Bredius\n\nAbraham Bredius (1855\u20131927) was a Dutch art collector and author of art books, known for his critical approach. He was especially a specialist in Dutch old masters, and the attribution of paintings.\n\nBredius was born on 2 July 1855 in Amsterdam, the son of a diamond merchant. He studied law at the University of Amsterdam and received his doctorate in 1879. He became a member of the Amsterdam Stock Exchange in 1878. He and his brother, Pieter Frederik Bredius, ran the diamond business for a time after their father's death in 1895. In 1899 he and his brother retired from business and moved to the town of Laren. His brother committed suicide in 1905.\n\nIn 1911, Bredius published his first art book, &quot;Rembrandt: een monografie&quot; (Rembrandt: a monograph), a scholarly study of Rembrandt's oeuvre, based on careful research of the existing works and contemporary documentation. His other publications include &quot;Jan Steen&quot; (1914), &quot;Frans Hals&quot; (1916), &quot;De groote schilderijen in het Mauritshuis&quot; (The large paintings in the Mauritshuis) (1923), &quot;A descriptive catalogue of the works of Peter Paul Rubens&quot; (1925) and &quot;De schilderselders in Nederland gedurende de XVIIe en XVIIIe eeuw&quot; (The painters of the Netherlands in the 17th and 18th centuries) (1927). The last book was published posthumously by his daughter.\n\nBredius was married to a daughter of Cornelis Hofstede de Groot, the art historian. Their daughter, Wilhelmina Bredius, was also an art historian.\n\nBredius was an art collector, and owned works by such artists as Frans Hals, Rembrandt and Jan Steen. He purchased many paintings and drawings at auction, and was considered to have the best collection of old masters in the Netherlands. The collection was put up for sale after his death, and many of the paintings found their way into museum collections. Some of his books are still available, as is the catalogue of his collection.\n]" time="1.297"><properties><property name="score" value="0.8649034202098846" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Chongqing Medical University\n\nChongqing Medical University () is a public university based in Chongqing, China.\n\nThe university is organized into the following colleges and schools:\n\n\nThe university's &quot;Health Research and Clinical Application Center&quot; (HRCC) is one of three bases of China's International Centre for Infectious Diseases (ICID). The university is also a participant in the China-Australia-Japan Trilateral Infectious Disease Laboratory (CAT-IDL).\n\n\n]" time="0.288"><properties><property name="score" value="0.9996054768562317" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Sesalpinia\n\nSesalpinia is a genus of flowering plants in the pea family, Fabaceae. The genus is native to the Americas, where it is distributed from the southern United States (Florida, Texas) to the northernmost part of Argentina (Tucum\xe1n). Species are known by many common names, including crown-peas, princess flower, Barbados pride, guizotia, maua, grama grass, whitetick, sandburs, Mexican ticks, pinquitos, piquitos, barbas de Indias, and zorillo.\n\nThe species of &quot;Sesalpinia&quot; are annual or perennial herbaceous vines, often with a woody base, reaching a height of 3 m. Leaves are bipinnate or tripinnate, made up of 10-30 leaflets, often with reddish margins. Flowers are in racemes, and are usually white to pink, though yellow and lavender flowers are known. The fruit is a pod, often with a distinctive hooked beak.\n\nThe genus is currently classified in the subfamily Mimosoideae, but the taxonomy is in need of a comprehensive phylogenetic study. The five species that have been traditionally recognized in &quot;Sesalpinia&quot; are not monophyletic. While the closest relatives of the genus are in the tribe Ingeae in subfamily Mimosoideae, &quot;Sesalpinia&quot; is not most closely related to any of the genera in that tribe.\n\n\n\n\n\n\n\n\n\n]" time="0.848"><properties><property name="score" value="0.14490091800689697" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Ion beam studies of materials\n\nIon beam studies of materials (also known as ion beam analysis or IBA) is a discipline that aims to determine the microstructure of solids. Ion beam studies of materials use mass spectrometry to study materials. Ion beam analysis is the name used for the techniques developed to study the solid materials with either a gas or a particle beam. Particle beams are commonly produced by ion guns or by secondary ion mass spectrometry. For high resolution, single particle mass spectrometry (called single particle IBA) or time of flight (TOF) instruments are employed. For many studies, electron microscopes can be used with appropriate detectors.\n\nSingle particle mass spectrometry was invented by Herb Guyer and Edward A. Heineman at Cornell University in 1953. This technique was the foundation of the field of ion beam analysis. This technique has been used to study a wide range of materials including metals, ceramics, glasses, organics and polymers.\n\nIn this technique the ion beam is focused to a small spot size on the sample, either by focusing in a magnetic or an electrostatic field or by being focused by electrostatic fields in a multielectrode system. The beam is then scanned over the surface of the sample. For each sample point, a mass spectrum of the sample material is obtained. The signals from the individual elements in the sample are then compared to each other in order to determine the composition and/or structure of the sample. This is because the signals are characteristic of the elements from which they are produced. The process of matching up the signals is called deconvolution. In this technique, the detector signal is measured only when a particle of the beam is detected.\n\nThe disadvantage of this technique is that it requires much more time to measure each sample point compared to scanning probe microscopes (which use electron beams).\n\nThe electron microscope with a detector attached to it can be used as an IBA tool. It is very useful to study the microstructure of ceramics and glasses.\n\nThe technique was developed at Cornell University and its name was coined by Thomas Vale in 1959. In electron microscope IBA, the sample is placed on a thin window and a beam of electrons is directed at it. The beam of electrons is either the electron beam itself or the secondary electron beam produced by a beam of primary electrons.\n\nWhen the electron beam is directed at the sample, some of the electrons are absorbed by the sample. This produces electrons inside the material and some ions. The ions are extracted by an electric field and are accelerated. The electrons emitted from the ions are detected. The detector may be a microchannel plate (MCP) detector or an electron multiplier.\n\nIn these systems the electron beam is accelerated between two plates. As]" time="1.489"><properties><property name="score" value="0.9998011390368143" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[A police officer suffered facial injuries when he was sprayed with acid during a &quot;very nasty&quot; street robbery on Monday evening.\n\nThe Metropolitan Police said a member of the public used water to wash out the officer's eyes. He has been taken to hospital.\n\nThe policeman was on duty in Clapton, east London, when he was confronted by a man who pulled a corrosive substance from a pouch and sprayed it into his face.\n\nThe officer was then forced to hand over his personal possessions.\n\nMetropolitan Police spokesman, Detective Sergeant Gary Brown said: &quot;This was a really nasty robbery, where the officer has approached a male and been sprayed with a substance - which we believe to have been an acid - in his face.\n\n&quot;He has then been further assaulted and had personal belongings stolen.&quot;\n\nDet Sgt Brown said police were not aware of any particular motive behind the attack, adding: &quot;It is probably a robbery that has gone wrong.&quot;\n\nThe officer is based at Stoke Newington Police Station, in Hackney, east London, and is a member of the borough's safer transport team, which was set up to reduce crime around Hackney's taxi and private hire trades.\n\nScotland Yard said the man responsible for the attack is described as a white male, in his late teens to early 20s.\n\nHe is about 5ft 11in, slim, with short blond hair and a small, blond beard. He was wearing a red, grey and black jacket, with blue jeans and white trainers.]" time="0.770"><properties><property name="score" value="0.9690220654010773" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Rapper says he wants to release songs that are &quot;like f**k Donald Trump&quot;\n\nLil B has said that he wants to release a mixtape titled \u2018Black Man\u2019s World\u2019, featuring songs about \u201cunity\u201d and \u201cequality\u201d.\n\nSpeaking to Billboard, the rapper, who released new track \u2018AfricAryan\u2019 in January, said he is \u201cgetting ready to drop the \u2018Black Man\u2019s World\u2019 mixtape\u201d.\n\n\u201cThat\u2019s about unity and equality, about togetherness, love,\u201d he said. \u201cI\u2019m working on it right now, and hopefully it will come out in June.\u201d\n\nThe rapper also revealed that the project will contain songs about \u201cstuff like taking over the world, like f**k Donald Trump, like making history\u201d.\n\nLil B\u2019s last mixtape, \u2018Black Ken\u2019, was released in 2016. The release was followed by an album of the same name in 2017, which was released without his consent.\n\nBriefly jailed in January for punching a woman in the face, the rapper later blamed the incident on a result of his \u201cmental disorder\u201d and said he planned to stop releasing music altogether.]" time="0.621"><properties><property name="score" value="0.9997389912605286" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[EAT-25 Test System\n\nThe Panasonic EAT-25 Test System is a complete one-box system for testing audio/video (A/V) signals, enabling designers of A/V equipment to perform critical measurements quickly and conveniently.\n\nOur comprehensive system includes EZ-25A and EZ-19A signal generators, both of which can generate the same high-quality A/V test signals, EZ-D9A dual digital pattern generator, EZ-10R high-speed data pattern recorder, and EZ-S8 data logger. You can use the system in your own lab or rent it on a daily or monthly basis through our worldwide sales network.\n\nEZ-25A\n\nThe EZ-25A signal generator can generate a wide range of test signals and is the perfect solution for designers who need to generate test signals for A/V equipment such as digital receivers, AV power amplifiers, etc.\n\nEZ-19A\n\nThe EZ-19A signal generator is ideal for designers who need to generate video test signals for the latest A/V equipment such as liquid-crystal displays (LCD), plasma displays, etc. It is also ideal for assessing test and measurement capabilities for other equipment such as laser displays and printers.\n\nEZ-D9A\n\nThe EZ-D9A is a dual digital pattern generator that can generate high-quality A/V test signals. It is ideal for use in the development of portable AV devices, as well as for assessment of measurement capabilities for other devices such as printers, etc.\n\nEZ-10R\n\nThe EZ-10R high-speed data pattern recorder is a flexible test system that can record high-speed data patterns, as well as graphic images. It can be used to measure jitter and data loss that is occurring during transmission of A/V signals, and is ideal for measuring the performance of external data-compression devices.\n\nEZ-S8\n\nThe EZ-S8 data logger can record signals such as sound and video. It is ideal for users who want to log events such as sound and video in a non-stop mode.]" time="1.116"><properties><property name="score" value="0.9998226165771484" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[$25.00\n\nThe City and Guilds Body Care Science Level 2 qualification is a highly effective means of providing the learner with the practical skills required to use a variety of products and procedures in the provision of beauty treatments.]" time="0.201"><properties><property name="score" value="0.9965519905090332" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Embarrassing tax mistake gives a false impression that tax gap is falling at a time when it is rising\n\nThe government\u2019s much heralded fall in the tax gap may have been boosted by an embarrassing tax \u2018error\u2019, which has skewed official figures to suggest that the amount of tax which goes unpaid is falling.\n\nHM Revenue and Customs (HMRC) has admitted that a major error in the way it handles unpaid taxes from firms using the banking system has overstated its tax gap \u2013 the amount of tax which goes unpaid \u2013 and understated the number of people working under the taxman\u2019s radar.\n\nIn a document on its website, it admitted that it has been failing to record taxes which should have been received but were not, but said it did not \u2018understate\u2019 the size of the tax gap.\n\n\u2018The effects of this error are that the statistics do not represent a full year\u2019s tax gap and do not fully represent the HMRC compliance work,\u2019 it said.\n\nThe error means that the tax gap fell from \xa335 billion in 2014/15 to \xa333.2 billion last year, when in fact the correct figure would have been \xa339.5 billion.\n\nThe mistake means that the tax gap is now out by more than \xa35 billion.\n\nThe admission follows a series of problems for the HMRC. The organisation was accused of losing sensitive personal data on more than 10 million people in an unencrypted CD in 2015, and was also accused of sending the wrong letters to half a million people.\n\nIt has also been criticised for a raid on the offices of investigative reporters at The Guardian.\n\nAn HMRC spokesperson said: \u2018It has become clear that our compliance data is not fully comparable over time, and we will not be publishing it in this form again. This does not affect the overall tax gap and our published estimates of the tax gap for the years 2014-15 and 2015-16 have not changed.\u2019\n\nA statement on the HMRC website said that the data published on 31 August \u2018estimated the total tax gap to be \xa333.2 billion for the year 2015-16. This figure represents a decrease of \xa34.7 billion on the 2014-15 tax gap of \xa338.9 billion, or a 14 per cent decrease\u2019.\n\nBut it admitted that the figure included \u2018incorrect estimates\u2019 on income tax, VAT and customs duty.\n\n\u2018We have corrected this error and published revised estimates for income tax, VAT and customs duty. These estimates now estimate the total tax gap to be \xa339.5 billion for the year 2015-16, which is an increase of \xa34.3 billion on the 2014-15 tax gap of \xa335.2 billion,\u2019 it said.\n\n\u2018This increase in the tax gap is not a true reflection of what has happened in compliance. It is because the 2015-16 estimates have included the correction of the error which has overstated the 2014-15 tax gap.\u2019]" time="1.580"><properties><property name="score" value="0.999380866686503" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[9 ATC ? What is an ATC and how do I use one? An ATC is an automatic transfer switch. It is used to transfer power to the boiler in the event of a power failure. These switches are &quot;automatic&quot; in the sense that they monitor the power flow through the generator and as soon as the power goes off, they engage the transfer switch (in parallel to the house breaker) to power the boiler. When the power is restored, the ATC turns off and disconnects the power to the boiler, restoring the house power. (We are not responsible for damages caused by not following these instructions). Before starting to install the system, you need to find the right place to put your ATC switch. Some considerations include: 1. You should put your ATC as close to the back of the generator as possible. This will make it easier for you to wire it up. 2. It is important to find a place that is dry. You can't use an ATC outside. It will just short out if it gets wet. 3. It needs to be in an accessible place (inside your home). It doesn't need to be near the breaker box (but it should be near a wall outlet). 4. The ATC should be installed at least 3 feet from the nearest source of water (sump pump, water heater, etc.) 5. If your generator is in a cabinet, you should put the ATC switch on the side of the cabinet so it is easy to]" time="0.789"><properties><property name="score" value="0.9974614083766937" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[One of my best friends, Rachel, is pregnant and is due in April, so of course I had to make her something. She loves to read, so I made her a book gift! It has 2 to choose from; The Mad Tea Party by Lauren Child and Mary Poppins.\n\nI used the Victoria Plum and The Grey from the Victoria Plum fabric line.\n\nI found the bookmarks at a scrapbooking store and thought they were adorable.\n\nI wish I could remember where I found the quote, I think I searched Etsy and bought it there.\n\nI thought it would be cute to make little tags to put the names on so she could just hand them out.\n\nI love the bright colors on the back!\n\nThank you for looking!]" time="0.349"><properties><property name="score" value="0.9997884631156921" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The original photographic image is composed of five photos stitched together with Photoshop. Each photo is on a different curve in order to show the fullest scope of the incredible reach of the fire.\n\n[See &quot;more photos&quot; below for further info and versions.]]" time="0.130"><properties><property name="score" value="0.994228720664978" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The map below shows the approximate location of every reported gorilla location between October 2010 and September 2011.\n\nThe red circles show where the gorillas have been reported.\n\nTo see the previous year's map, click here.\n\nMapping gorillas is difficult because of the rugged and thick vegetation in the forests of Congo, where they are found.\n\nOften the trackers do not actually see the gorillas themselves, only hear their distinctive screams, and have to make their estimates based on signs such as footprints and vegetation trampled by the large apes.\n\nHowever, the gorilla range is thought to have expanded significantly in the last five years.\n\nThe good news is that the population is now thought to be in excess of 4,000 gorillas.\n\nThere are many more threats to gorilla survival than loss of habitat, however, and none are yet under control.\n\nThe gorilla is Africa's most threatened mammal, and faces threats from poaching and disease, in addition to the loss of their habitat, and the practice of hunting them for &quot;bushmeat&quot;.\n\nThe Rwandan genocide of 1994 also left a terrible legacy, with the influx of refugees leaving a legacy of poaching and poverty.\n\nThe official policy of the Republic of Congo is that all gorillas should be left in peace in their natural habitat, which is now also under pressure from logging.\n\nThe WWF is helping to train a team of gorilla trackers to continue to monitor the gorillas' progress.\n\nYou can support their work by donating to the WWF.]" time="0.769"><properties><property name="score" value="0.9998009204864502" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[A neighborhood staple since 2003, Shawarma Grill offers quality Mediterranean cuisine, including an extensive menu of vegetarian and vegan options. Customers rave about the huge portions, reasonable prices, and friendly staff. If you're looking for healthy options, Shawarma Grill has salads, burgers, and baked potatoes on the menu. The vegetarian gyro is another hit, and the shawarma is sure to please any meat lover.]" time="0.219"><properties><property name="score" value="0.9917821884155273" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Related articles: 1/3\n\n2/3\n\n3/3\n\nAre you looking for a way to improve your website\u2019s search engine rankings? How about better user experience? There are many ways you can accomplish both.\n\n1. Reduce your page\u2019s load time\n\nGoogle uses the page load time as one of the factors when determining your search engine rankings. You can use PageSpeed Insights to measure the page load time for any URL.\n\nYou should aim for a score of 90/100 and lower. It\u2019s worth pointing out that although this factor is being used as a ranking factor, a slow site will have a negative impact on user experience, which in turn will affect your brand image, search results, and conversions.\n\n2. Enable GZIP compression\n\nGZIP compression compresses the HTML, CSS, and JavaScript files in your web pages, which reduces the size of those files and the total data transferred to the user. Google uses this factor to determine page load time.\n\nPageSpeed Insights has a compression test that allows you to measure the level of compression achieved by your website.\n\nFor more information on how to enable GZIP compression, see this article.\n\n3. Enable HTTPS\n\nHTTP Strict Transport Security (HSTS) allows websites to tell browsers to avoid interacting with any sites except the ones that use HTTPS. This is useful in preventing man-in-the-middle attacks.\n\nAs of this writing, Google considers HTTPS as a ranking factor in its search algorithm.\n\nThe problem is that although enabling HTTPS is pretty easy (for WordPress websites), it\u2019s not so easy when you have to replace all of your URLs with HTTPS URLs, and make sure that your backlinks are using HTTPS.\n\nIf you\u2019re not ready to migrate, you can use the free tool from CloudFlare, which allows you to encrypt your WordPress website and even apply SSL to all of your images, making the switch to HTTPS easy and effective.\n\n4. Check your internal links\n\nInternal links are links that point to another page on your own website. Google uses these links to determine the value of the pages in your website, and how well your site is structured.\n\nLink-building specialists will tell you that there are a lot of ways to build good quality links.\n\nFor example, you can use social bookmarking sites, forums, and blogs. For links pointing to your website, you can ask friends, relatives, clients, and business partners to post links to your website on their websites.\n\nBut, you should do this within reason, and avoid looking like a link-building spammer.\n\n5. Include internal links in your blog posts\n\nYour blog posts should include internal links, to ensure that Google can detect the relationships between your posts. This will also improve your readers\u2019 experience by allowing them to quickly find the posts that are most relevant to them.\n\n6. Create an XML sitemap\n\nAn XML sitemap is a page on your website that contains a list of all the URLs on your website. If you\u2019ve been blogging for some time, you\u2019ve probably already created a sitemap to allow search engines to crawl your site.\n\nThe problem is that search engines do not use the sitemap file on your website to crawl your website. In other words, sitemap file is useless to search engines.\n\nFortunately, there\u2019s a free plugin from Yoast that can automatically generate an XML sitemap and also add the sitemap to your website.\n\n7. Install social sharing plugins\n\nSocial media is a great way to market your website and brand. Not only that, you can also use social media as a way to improve your search engine rankings.\n\nWhen you share your content on social media, you can make the content more accessible to your visitors. Not to mention the fact that social sharing plugins also provide users with a link to your content, which is great for SEO.\n\nSome of the best social sharing plugins for WordPress websites include Facebook, Pinterest, and Twitter.\n\n8. Install and enable W3 Total Cache\n\nW3 Total Cache can greatly improve your website\u2019s load speed by enabling caching for your website\u2019s pages, posts, and images.\n\nYou can read my review of W3 Total Cache to learn more.\n\n9. Install and enable Yoast SEO\n\nYoast SEO allows you to configure your search engine settings for each post and page on your website. It\u2019s the most powerful SEO plugin for WordPress.\n\nFor more information on how to configure Yoast SEO for optimal SEO, see my tutorial on SEO for WordPress.\n\n10. Install and configure Google XML Sitemaps\n\nThis is another free plugin from Google that generates an XML sitemap for your website. This sitemap will allow search engines to crawl your website faster and more efficiently.\n\nIt also helps you to build quality backlinks to your website. The links in the XML sitemap are much stronger than the links in your sitemap file, because they are permanent links.\n\nYou can read more about how to configure Google XML Sitemaps for optimal SEO.\n\n11. Install and enable Google Analytics\n\nGoogle Analytics is a free analytics tool from Google that lets you monitor your website\u2019s performance, including the number of visitors you\u2019re getting, where they\u2019re coming from, and what they\u2019re doing on your website.\n\nI have my own tutorial on how to configure Google Analytics for optimal SEO.\n\n12. Install and enable Google Webmaster Tools\n\nWebmaster Tools is another free tool from Google that lets you verify your website and configure a wide range of settings, including your preferred domain, how your website is indexed, what the canonical URLs are, whether you allow Google to crawl dynamic pages, whether you allow web crawlers to follow your links, and whether you allow Google to scan your pages for malware.\n\nWebmaster Tools is especially useful for websites that have a lot of dynamic pages, and pages that are indexed using non-standard URLs.\n\n13. Install and enable Google Analytics by Yoast\n\nGoogle Analytics by Yoast is another free plugin that allows you to synchronize your Google Analytics account with the most popular SEO plugin for WordPress, Yoast SEO.\n\nWith this plugin, you can easily add your Google Analytics tracking code to all your posts, pages, and sidebars, without having to manually add the tracking code for each post.\n\n14. Submit your website to Google\n\nIf you have a new website, you should submit it to Google Webmaster Tools, which will allow Google to index your website and create a site map.\n\n15. Install a contact form plugin]" time="3.526"><properties><property name="score" value="0.9772511294909886" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[As the fashion world is taking notice of Athleisure (a new style trend of casual attire that incorporates sports-luxe style elements), a new video called The Slits, from G. Garvin (a C-level marketing executive), perfectly illustrates why we love this trend. The video features an athleisure-clad woman who uses the sporty-swag inspired outfit to complete a variety of day-to-day tasks in style.\n\nThe music to the video is performed by Miley Cyrus and written by Cyrus and Pharrell Williams. The music video is part of a larger promotion for the G. Garvin brand.]" time="0.327"><properties><property name="score" value="0.999817430973053" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[.\n\n\n\nAttenzione:\n\nL'ingresso al museo \xe8 gratuito.\n\nLa visita ai locali espositivi e ai musei ad esso collegati \xe8 a pagamento.\n\nIl costo della visita \xe8 di \u20ac 6,00 per adulto e di \u20ac 3,00 per bambini fino ai 12 anni.\n\n\n\nOrario\n\nAperto il marted\xec, mercoled\xec, venerd\xec e sabato dalle 9:00 alle 17:30.\n\nGioved\xec dalle 9:00 alle 22:00.\n\nDomenica dalle 10:00 alle 17:30.\n\n\n\n]" time="0.377"><properties><property name="score" value="0.9997382760047913" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99973828&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99973828
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[In today\u2019s digital age, many people look to the internet to learn about and purchase products. These include retail shoppers who look to Amazon for new toys or appliances, travelers who look to Expedia for airline tickets, and even patients who look to the internet to learn about a disease or find a doctor. This presents unique challenges for brand owners. While brand owners can claim ownership of certain content on their sites and apps, much of the internet is controlled by other entities. This includes social media sites, search engines, content syndicators, content aggregators, and other web sites. A brand owner who does not protect its name and trademarks on these sites risks losing control of its brand online.\n\nIn the past, brand owners who found a third-party using its trademarks or similar trademarks had to prove that the third-party was using the mark in a confusing manner. This required proving that the third-party\u2019s use of the mark was likely to cause confusion in the mind of the average consumer, i.e. that it was likely to be mistaken for the mark of the brand owner. Today, however, that burden of proof is no longer necessary.\n\nA little over one year ago, the Ninth Circuit Court of Appeals decided the landmark case of Global-Tech v. eBay, 562 F.3d 982 (9th Cir. 2009). The Global-Tech decision held that a plaintiff can bring a claim against a third-party for cybersquatting by using the new procedures set forth in the Anti-Cybersquatting Consumer Protection Act of 1999 (\u201cACPA\u201d). Previously, under the ACPA, a brand owner had to prove confusion to make a cybersquatting claim. Under the new regime, the burden has been placed on the third-party to show that its use of a domain name is legitimate.\n\nThe burden shift was the most significant change in the Global-Tech case. Before Global-Tech, a cybersquatting plaintiff had to prove confusion. After Global-Tech, a cybersquatting plaintiff only needs to prove that a defendant has registered a domain name similar to the plaintiff\u2019s mark with bad faith intent to profit.\n\nAccording to the Ninth Circuit, \u201cthe ACPA\u2019s [bad faith] requirement means little more than that a domain name registrant must intend to profit from the goodwill associated with the plaintiff\u2019s mark.\u201d\n\nOnce a cybersquatting plaintiff has proven this intent, a plaintiff need only show that it owns a trademark that is similar to the defendant\u2019s domain name in order to prevail on its claim. This is the essence of the burden shift. A plaintiff who successfully proves that the defendant is cybersquatting on a domain name that is similar to its trademark does not need to prove that consumers are being confused. The cybersquatter must show that its use of the domain name is legitimate.\n\nThe Global-Tech case also made it easier for a plaintiff to win a cybersquatting case because of the additional burdens placed on a defendant.\n\nFirst, the Global-Tech case made it harder for a cybersquatting defendant to defend itself on the merits. Before Global-Tech, a defendant could defend a cybersquatting claim by arguing that its use of the domain name was not confusing to consumers. However, after Global-Tech, a defendant can no longer rely on its good faith use of the domain name to show that it is not cybersquatting. Instead, a defendant must show that its use of the mark is legitimate. This has led to claims of cybersquatting where the mark and the domain name are similar, but where there is no evidence that consumers are being confused.\n\nSecond, the Global-Tech case made it easier for a cybersquatting plaintiff to obtain injunctive relief. Traditionally, a cybersquatting plaintiff had to show that it was likely to prevail on its claim in order to obtain an injunction to prevent a defendant from using its domain name. After Global-Tech, a plaintiff need only show that the defendant is likely to profit from its use of the domain name to obtain an injunction. Since an injunction may be obtained before trial, this allows plaintiffs to eliminate the use of their marks by defendants before they can defend their actions in court.\n\nThe Global-Tech case has led to a shift in the burden of proof in cybersquatting cases. This shift has created an environment where cybersquatting cases are becoming more difficult to defend.]" time="2.594"><properties><property name="score" value="0.13550121784210206" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.13550122&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.13550122
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Rated 5 out of 5 by Bullockie79 from The only thing better than this shirt is the price I\u2019m a bigger guy and I like to buy shirts that fit my body. This shirt is long enough in the body and the sleeves, so the shirt fits properly. I\u2019m very pleased with the weight of the shirt. It\u2019s not too heavy. So I don\u2019t get too hot.\n\nRated 5 out of 5 by kramer from great product This is the second shirt I have purchased in this style. I wear them almost every day and wash them many times. They are very durable.\n\nRated 5 out of 5 by JJG1 from Great shirt Great shirts. Great quality and material. Washes great and hangs out on the line to dry. Great with the Carhartt suspenders.\n\nRated 5 out of 5 by Bihangir from Fantastic! Awesome product, awesome quality. Will buy more for sure. I bought this on sales and even then, it is a good buy.\n\nRated 5 out of 5 by Anthony from A go-to for work, play, anywhere! Great fitting shirt, true to size. A comfortable classic for any occasion. It is heavy weight, but does not feel thick, just a nice fit.\n\nRated 5 out of 5 by Carpenter from Best dress shirts These are great shirts they fit well, they are long enough. I would recommend them to anyone.\n\nRated 5 out of 5 by KLJ93 from Great fit These shirts are great for work or everyday wear. They fit perfectly and are of very high quality.]" time="0.816"><properties><property name="score" value="0.8858896791934967" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[One of my favourite times in the barbershop. I love conversations with guys about life, business, sex, sports and news.\n\n\n\nA friend of mine said, &quot;The best conversations happen over a good cup of coffee&quot;.\n\n\n\nThat\u2019s what I\u2019ve done in my new book... \u2018Good Coffee, Bad Conversations\u2019.\n\n\n\nIf you\u2019ve got a few minutes, I\u2019d like to share a story with you.\n\n\n\nIn the late 90\u2019s, my business partner, Jeff \u2013 a good friend of mine and also a barber \u2013 was having problems. He was having lots of conflict with his wife and daughter and he was angry all the time.\n\n\n\nI asked him if he wanted to talk about it.\n\n\n\nHe said, \u201cNo. That would be good coffee, bad conversation.\u201d\n\n\n\nIt\u2019s true, Jeff was scared to open up and share what was really going on in his life. He was afraid of what he would hear.\n\n\n\nI remember this time because I wanted to help Jeff, but I didn\u2019t know what to do. So I did the only thing I knew how to do.\n\n\n\nI started writing.\n\n\n\nAs I wrote, I realized that coffee was a lot like conversation.\n\n\n\nA cup of coffee was good.\n\n\n\nA cup of coffee was bad.\n\n\n\nThat\u2019s when it hit me... \u2018Good Coffee, Bad Conversations\u2019.\n\n\n\nIt\u2019s about the conversations we have and how they can make or break our lives.\n\n\n\nIn my book, I show you the seven bad conversations you need to stop having. You\u2019ll also learn the seven good conversations you should start having instead.\n\n\n\nAs a bonus, I also share the six good conversations I\u2019ve had with some amazing people including Grant Cardone, Kody Deis, Andy Frisella, Gary Vaynerchuk and others.\n\n\n\nThere\u2019s more, but you\u2019ll have to read it to find out.\n\n\n\nThe book is short and concise. It\u2019]" time="1.039"><properties><property name="score" value="0.9949003458023071" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Magic: the Gathering has launched the Dominaria campaign. Dominaria is a plane steeped in rich lore and ancient magics. The world is home to several important factions, including Benalia, home to the Knights of Benalia, and the nation of Corondor. Corondor is ruled by Baron Sengir and his brother, Baron Sardu.\n\nDominaria is also home to some of Magic\u2019s most iconic characters. The Weatherlight is crewed by a legendary host of characters, including Captain Sisay, Karn, and Teferi. But the Weatherlight\u2019s roster is far from limited to famous crew members.\n\nThis is the history of the Weatherlight\u2019s crew members who weren\u2019t famous enough to get their own card.\n\n(Special thanks to Jason Carl\u2019s Comprehensive Rules FAQ and to YSDC\u2019s flavor text wiki for their incredible flavor text sources.)\n\nVenser, Shaper Savant\n\nVenser is an artificer who discovered his latent Planeswalker abilities while recovering from an injury on the plane of Shandalar. After being recruited to join the crew of the Weatherlight, Venser eventually fell into a coma and later revived himself on the plane of Koilos.\n\nKarn, Silver Golem\n\nKarn is a silver golem who used to live on the plane of Argentum. He came aboard the Weatherlight after his homeworld was destroyed by a time riptide.\n\nAven Windreader\n\nAven Windreader is a scout who used to serve in the Feather clan of Aven, a nomadic race. She is joined on the Weatherlight by two siblings: Aven Brigadier and Aven Trooper.\n\nSisay, Captain\n\nSisay is a captain from the nation of Jamuraa, and is the Captain of the Weatherlight.\n\nSquee, Goblin Nabob\n\nSquee is a goblin on the Weatherlight crew. His fellow goblin cohorts on the crew are Rishadan Cutpurse, Rishadan Footpad, and Greel.\n\nIrina, Great Glass-Spinner\n\nIrina is an illusionist who used to serve the Imperial Court of Otaria. She was exiled after the court was overthrown and joins the Weatherlight in search of answers.\n\nSharuum the Hegemon\n\nSharuum is an artificer who, in her past life, became a planeswalker. Sharuum built a stronghold called Koilos on the plane of Koilos, and served as a guardian to the mysterious stronghold.\n\nEdric, Spymaster of Trest\n\nEdric was a spy who was arrested after his cover was blown. Edric\u2019s home nation of Trest was destroyed, and he was freed from his cell aboard the Weatherlight in exchange for his assistance on the ship\u2019s infiltration of Phyrexia.\n\nSquee, Goblin Nabob\n\nSquee is a goblin who served as the goblin court advisor to the Supreme Arbiter of the Imperial Court of Otaria. He joins the Weatherlight crew after being abandoned in Krov on Dominaria, and joins the crew in order to find a way home.\n\nTahngarth, Talruum Hero\n\nTahngarth is a Tahngarth is a minotaur who hails from the nation of Talruum. Tahngarth was captured by Sarpadian pirates and sold into slavery. He was rescued from his servitude by the crew of the Weatherlight.\n\nKarn, Silver Golem\n\nKarn was a silver golem who lived on the plane of Argentum. He boarded the Weatherlight after his homeworld was destroyed by a time riptide.\n\nKarn, Silver Golem\n\nKarn is a silver golem who used to live on the plane of Argentum. He came aboard the Weatherlight after his homeworld was destroyed by a time riptide.\n\nStarke of Rath\n\nStarke of Rath is a master thief who joins the Weatherlight crew after discovering a conspiracy to overthrow the current government of his native country of Rath.\n\nKarn, Silver Golem\n\nKarn was a silver golem who used to live on the plane of Argentum. He came aboard the Weatherlight after his homeworld was destroyed by a time riptide.\n\nGryff, Knights of Benalia\n\nGryff was a member of the Knights of Benalia who died in a battle against an Onakke Ogre. He was later revived by Tiana, Ship\u2019s Caretaker, who repurposed his damaged body as a messenger for the ship.\n\nKarn, Silver Golem\n\nKarn is a silver golem who used to live on the plane of Argentum. He came aboard the Weatherlight after his homeworld was destroyed by a time riptide.\n\nTiana, Ship\u2019s Caretaker\n\nTiana is a ship\u2019s caretaker who was attached to the Weatherlight before it was even built. She was responsible for the upkeep of the ship, including a number of tiny, flying custodians.\n\nTiana, Ship\u2019s Caretaker\n\nTiana is a ship\u2019s caretaker who was attached to the Weatherlight before it was even built. She was responsible for the upkeep of the ship, including a number of tiny, flying custodians.\n\nAven Envoy\n\nAven Envoy was a captain who used to serve in the Knights of Avon. She lost her entire company during a battle against an Onakke Ogre. Aven Envoy was revived by Tiana, Ship\u2019s Caretaker, and took up a new post as envoy for the Weatherlight.\n\nUrza Planeswalker\n\nUrza was a planeswalker who, after his death, was merged with a powerstone and transformed into an artificial planeswalker. He was the first person to travel to the plane of Phyrexia, and later forged a coalition of planeswalkers against the Phyrexians.\n\nSquee, Goblin Nabob\n\nSquee is a goblin who served as the goblin court advisor to the Supreme Arbiter of the Imperial Court of Otaria. He joins the Weatherlight crew after being abandoned in Krov on Dominaria, and joins the crew in order to find a way home.\n\nIcatian Lieutenant\n\nIcatian Lieutenant is a soldier from the nation of Icatia. She joined the crew of the Weatherlight in order to track down her fellow Icatian soldiers who were captured by Rathi forces.\n\nKarn, Silver Golem\n\nKarn is a silver golem who used to live on the plane of Argentum. He came aboard the Weatherlight after his homeworld was destroyed by a time riptide.\n\nBarren Glory\n\nBarren Glory is a captain who came from an apocalyptic future. He used to captain a ship called the Weatherlight, which is (unsurprisingly) a ship not unlike the Weatherlight.\n\nJhoira, Weatherlight Captain\n\nJhoira is a master artificer who was captain of the Weatherlight for a time.\n\nGryff, Knights of Benalia\n\nGryff was a member of the Knights of Benalia who died in a battle against an Onakke Ogre. He was later revived by Tiana, Ship\u2019s Caretaker, who repurposed his damaged body as a messenger for the ship.\n\nCaptain Lannery Storm\n\nCaptain Lannery Storm was a member of the crew of the Weatherlight who was killed during a mission to Rath. She was later revived by Tiana, Ship\u2019s Caretaker, and now serves on the Weatherlight as a crew member.\n\nNimbus Maze\n\nNimbus Maze was an artificer who served the Knights of Benalia. He was sent to accompany Squee on a diplomatic mission to Icatia, but was killed by a phyrexian vat lord during the visit.\n\nSquee, Goblin Nabob\n\nSquee is a goblin who served as the goblin court advisor to the Supreme Arbiter of the Imperial Court of Otaria. He joins the Weatherlight crew after being abandoned in Krov on Dominaria, and joins the crew in order to find a way home.\n\nAjani, Mentor of Heroes\n\nAjani is a leonin planeswalker who was once the leader of the Capashen clan on the plane of Alara. His pride led to a war between the clans of Capashen and Kumena. After the war, Ajani renounced his leadership and traveled to Dominaria.\n\nSquee, Goblin Nabob\n\nSquee is a goblin who served as the goblin court advisor to the Supreme Arbiter of the Imperial Court of Otaria. He joins the Weatherlight crew after being abandoned in Krov on Dominaria, and joins the crew in order to find a way home.\n\nBarren Glory\n\nBarren Glory is a captain who came from an apocalyptic future. He used to captain a ship called the Weatherlight, which is (unsurprisingly) a ship not unlike the Weatherlight.\n\nSquee, Goblin Nabob\n\nSquee is a goblin who served as the goblin court advisor to the Supreme Arbiter of the Imperial Court of Otaria. He joins the Weatherlight crew after being abandoned in Krov on Dominaria, and joins the crew in order to find a way home.\n\nKarn, Silver Golem\n\nKarn is a silver golem who used to live on the plane of Argentum. He came aboard the Weatherlight after his homeworld was destroyed by a time riptide.\n\nCaptain Lannery Storm\n\nCaptain Lan]" time="4.971"><properties><property name="score" value="0.9387495890259743" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[BestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disincastrare la Verifica Torx\xae Per Tutti i Modelli\n\nBestoBasket Cacciaviti Sfilabile Meccanico per Disinc]" time="5.947"><properties><property name="score" value="0.712774246931076" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[All the new games for the week\n\nOnce upon a time, the first week of the month meant a new Switch eShop release was imminent, but recently, Nintendo has shown an admirable level of restraint. The first week of this month is no different, as there's just a couple of new games on the way to the system's digital store.\n\nComing to North America this week is the sequel to the surprise hit of 2017, Wonder Boy: The Dragon's Trap. The platformer has a rich history on the PC and consoles, but this remake proved to be well worth the time spent by the fans. With a new look and modern gameplay, this is certainly worth your time if you haven't played it yet. The Dragon's Trap releases in North America on February 9.\n\nFor our European readers, this week brings us into the dangerous world of SteamWorld Dig 2. While the first game was critically acclaimed and earned quite a bit of love, this sequel looks to improve on the formula by adding a lot more variety and some online multiplayer options. Dig 2 is a bit more than a week away, and will be available on February 9.\n\nIf that's not enough for you, then you'll be happy to know that the Nintendo Switch's Virtual Console service has something new in store for you. For the week of February 13, the Nintendo 64 classic Donkey Kong 64 will be available on the system's store.\n\nAs always, we'll keep an eye on the eShop and update you on new releases if and when they become available. If you're holding out for some new physical releases, we should have some news later today.\n\nThis article may contain links to online retail stores. If you click on one and buy the product we may receive a small commission. For more information, go here.]" time="1.005"><properties><property name="score" value="0.999822735786438" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[mikeshea.net\n\nScribblechat: Thoughts and Notes on Black Mountain\n\nby Mike Shea on 3 October 2003\n\nThis morning I saw the movie &quot;Black Mountain&quot; in a theatre near me. I had been really looking forward to it since I heard about it months ago. I enjoyed it a lot, but the final scene didn't end with the punch I was expecting. The entire movie was a quiet, intense build-up to that one final scene and, as much as I enjoyed the movie, I found myself disappointed at the end.\n\nThe film is good, I highly recommend it, but be ready for an intensely quiet movie. You might even want to bring a set of earplugs to help block out the noise from the rest of the theatre.\n\nSend comments to mike@mikeshea.net or follow @mshea on Twitter. If you enjoyed this article, please use this link to Amazon.com for your next online purchase.]" time="0.458"><properties><property name="score" value="0.999799907207489" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[View Full Version : How can I identify potential variables?\n\nMikeJackness Hi all,\n\nI'm trying to learn about and identify some potential (input) variables for a simple device.\n\n\n\nI've looked at the classic approach using the Analyser to get a range of DC voltages (input values) and their corresponding output values but that's not what I'm looking for.\n\n\n\nI want to find out what variables may affect the final values I am looking for and why they have an effect on the outputs.\n\n\n\nAn example would be, lets say I am looking for DC values of between 2V and 4V. The device I am testing has a button which, when pressed, will alter the value of something between 1% and 100%. How can I find out what that value is? I know that it's 1% - 100%, but I want to know the value of the device and how it affects the outcome.\n\n\n\nThanks\n\nMike\n\nIan Smith You need to identify what it is you are actually measuring.\n\n\n\nAre you measuring the current through a certain circuit? If so, you need to measure the voltage drop across it (and also the current through the circuit itself)\n\nAre you measuring the temperature of a thermocouple?\n\nAre you measuring the force produced by a tensile tester?\n\netc etc.\n\n\n\nThe only way to find out is to identify what is actually changing (when the button is pressed) and what is being measured.\n\n\n\nHope that helps.\n\nMikeJackness You need to identify what it is you are actually measuring.\n\n\n\nAre you measuring the current through a certain circuit? If so, you need to measure the voltage drop across it (and also the current through the circuit itself)\n\nAre you measuring the temperature of a thermocouple?\n\nAre you measuring the force produced by a tensile tester?\n\netc etc.\n\n\n\nThe only way to find out is to identify what is actually changing (when the button is pressed) and what is being measured.\n\n\n\nHope that helps.\n\n\n\nIt does help a lot, thank you. I'm looking at the values in the actual device rather than the sensor values.\n\n\n\nI'm trying to test a stepper motor, and I want to see if the resistance of the motor changes when it turns.\n\n\n\nI guess I'll have to figure out what the resistance is when the motor is not turning and what it is when it is, then identify the variables that affect it.\n\n\n\nAny more ideas would be great.\n\nIan Smith I'm trying to test a stepper motor, and I want to see if the resistance of the motor changes when it turns.\n\n\n\nAh! Well, you've gone and identified the system (stepper motor) and the variables (resistance). Now all you have to do is find out the actual resistance and how it changes when the motor is turned.\n\n\n\nAlso, have you determined whether the motor is powered from an AC mains supply or from a DC battery (or similar)? You need to know this because AC mains supplies may vary the current they provide, depending on the load (which may include the motor), and some DC supplies may have a current limit (which you may need to bypass if it's too low).\n\n\n\nHope this helps.\n\nMikeJackness Hi all,\n\nI've been doing some reading and thinking about how to approach this, and it looks like my answer is to use two multimeters.\n\n\n\nOne to measure the voltage of the stepper motor (when turned) and the other to measure the current flowing through it (when turned).\n\n\n\nThe questions are;\n\n\n\nWhat kind of multimeter should I use?\n\nShould I use a panel mount multimeter for ease of use, or a DMM?\n\nShould I try to build a sensor for the current measurement?\n\nIan Smith Hi all,\n\nI've been doing some reading and thinking about how to approach this, and it looks like my answer is to use two multimeters.\n\n\n\nOne to measure the voltage of the stepper motor (when turned) and the other to measure the current flowing through it (when turned).\n\n\n\nThe questions are;\n\n\n\nWhat kind of multimeter should I use?\n\nShould I use a panel mount multimeter for ease of use, or a DMM?\n\nShould I try to build a sensor for the current measurement?\n\n\n\n1. Both, to begin with. You may only need one, but you may need two, and it's better to have two of them.\n\n\n\n2. As above. If you can't build the current measurement, then either use one of the basic models that has a current setting, or use the voltage reading on the meter and then calculate the current from that. (e.g. multiply by 0.6 to get approximate amperes)\n\n\n\n3. Again, as above. If you can't build a current measurement, then you can get a panel meter to do the job for you, or you can try one of the fancier meters that have current measurement as one of their functions.\n\nMikeJackness Thanks for the replies. I am currently using two multimeters, a panel meter and a handheld meter.\n\n\n\nI've built an adjustable resistance using one of the 5W, 12V heater elements from a microwave. When I put it in series with the motor it showed a resistance of around 400 Ohms.\n\n\n\nI was then trying to identify how much current it drew when it was switched on, and I couldn't get a good result because the panel meter wasn't reading it properly.\n\n\n\nI think that's where I'm at with it at the moment. I may try using my handheld multimeter to see if it can read the current draw, but that may prove difficult because I'm not sure where the current is going through the stepper.\n\n\n\nThanks again for all the advice.\n\nIan Smith Thanks for the replies. I am currently using two multimeters, a panel meter and a handheld meter.\n\n\n\nI've built an adjustable resistance using one of the 5W, 12V heater elements from a microwave. When I put it in series with the motor it showed a resistance of around 400 Ohms.\n\n\n\nI was then trying to identify how much current it drew when it was switched on, and I couldn't get a good result because the panel meter wasn't reading it properly.\n\n\n\nI think that's where I'm at with it at the moment. I may try using my handheld multimeter to see if it can read the current draw, but that may prove difficult because I'm not sure where the current is going through the stepper.\n\n\n\nThanks again for all the advice.\n\n\n\nI can't help with the current measurement, I don't know much about motors.\n\n\n\nBut, if you]" time="3.396"><properties><property name="score" value="0.9998061458269755" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Now you can print documents, or send a fax, from virtually anywhere. If you have an Internet connection and a phone, you can create, send, and receive faxes from the convenience of your home or office.]" time="0.179"><properties><property name="score" value="0.9986180067062378" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Unbelievably cute.\n\n\n\nThis is a review of the artbook for Lucky Star. I'm not going to write a full review, because you can find that on other sites.\n\n\n\nLucky Star is a 4-panel comedy manga that revolves around the lives of four girls. In this artbook, it also includes background sketches, 3D illustrations of some of the characters, and even some concept art of the outfits.\n\n\n\nThe artwork is cute, and will make you smile, and for me, that's what makes this a good book. It has all the artwork for the series, and all the side-materials as well.]" time="0.284"><properties><property name="score" value="0.99982088804245" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[LINGUIST List 12.2282\n\nThu Apr 26 2001\n\nReview: Discourse Studies\n\nEditor for this issue: Terence Langendoen &lt;terry linguistlist.org&gt;\n\nDirectory\n\nMessage 1: Genre Analysis and Discourse Analysis\n\nKen Hyland, Introducing Discourse Analysis (Blackwell Introductions to Applied Linguistics) , xii+267pp, paperback ISBN 0-631-22512-8, hardback ISBN 0-631-22511-X, $47.95 / ??35.00 The blurb on the back: This textbook introduces students to the main traditions of discourse analysis, from structuralism to modern approaches, and includes many examples of actual texts for study and analysis. Hyland has developed his own approach, based on a number of leading methods, to show the student how to apply them to language use. Divided into four parts, this book explores the key concepts, applications and findings of these methods, in the first section looking at language and discourse as a unit, in the second examining patterns of text structure, in the third explaining how discourse is used to construct particular kinds of meaning, and in the fourth looking at how we study discourse. The first three sections introduce three different approaches to discourse analysis - the macro-structural approaches of Saussure and Gee, the micro-structural approach of Halliday, and the interpersonal approaches of Brown and Hodge. The last section explores ways in which different kinds of discourse are studied, including conversation, literature, power, work and the media. What I liked: This is an excellent textbook. It provides a good survey of a range of discourses analysis approaches and provides clear descriptions of the methodological principles of these approaches. This is an approachable introduction to a complex and diverse field. The topics covered include the analysis of conversation, text and the various modes of literature. Hyland shows how the same principles can be applied to these different types of text. He draws on a wide range of literature to show how discourses can be studied and includes many examples of actual texts. The book is very clearly written, contains many well-chosen examples, and provides a good introduction to a subject which is likely to be of increasing interest to Linguists. What I didn't like: The chapter on the analysis of conversation and the chapter on the analysis of discourse in fictional texts would benefit from a more rigorous discussion of the key concepts used in this type of analysis. For example, the chapter on conversation provides an excellent introduction to this type of discourse, but it would be better if there was a clearer discussion of the concepts of turn-taking and overlap, and a more careful explanation of the use of 'look-back' studies and the main types of transcription. The chapter on discourse in fictional texts provides an introduction to a type of analysis which has not been covered elsewhere in the book, and this is an important addition to the text. However, this chapter does not include the discussion of the key concepts of discourse analysis which is provided in the other chapters. It would have been better if this chapter could have followed the same format as the other chapters, and discussed how a particular approach was developed, what its key concepts are, and how these concepts can be applied. I think that this is an important addition to the book. Another minor point is that there are several typos in the book, but I don't think that this detracts from the overall quality of the book. What is missing: A discussion of the key concepts of discourse analysis which can be applied to all types of text.]" time="1.834"><properties><property name="score" value="0.499798446893692" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Home Articles &amp; Publications\n\nHOME ARTICLES &amp; PUBLICATIONS\n\nFind out more about the Ark Community Foundation and our many achievements here.\n\nIN THE NEWS\n\nWHAT'S AHEAD\n\nSee what exciting events and programs the Ark Community Foundation is involved in this year!]" time="0.171"><properties><property name="score" value="0.9084698557853699" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[More than 600 students in the CSU system are waiting for their transcripts to be cleared.\n\nTranscripts at CSU campuses are frozen for over 600 students who attended the Pueblo campus last semester, and if they don't complete their credits and final exams, they could lose their scholarships and their seats.\n\nAccording to the CSU-Pueblo website, 615 students have not yet completed their spring semester. They will lose their financial aid eligibility if they don't finish.\n\n&quot;It's like they're on a waiting list,&quot; said Dale Lange, who works at the registrar's office. &quot;We're waiting for the teachers to grade their finals and the professor to submit the grades and then they can release the transcript.&quot;\n\nOn March 20, the Pueblo campus was told to cease instruction and close by the chancellor's office.\n\n&quot;It's just horrible,&quot; said CSU-Pueblo student Julea Prentice. &quot;I don't know if I'm graduating or not.&quot;\n\nPrentice said she's upset she's paying for classes that she can't use, but she feels she doesn't have a choice.\n\n&quot;If I want to go to graduate school I have to get a certain GPA, so I can't afford to fail this semester,&quot; said Prentice.\n\nMeanwhile, her son is suffering because he needs a caretaker for his medical condition.\n\n&quot;That's why I'm a mother, that's why I'm a mom, that's why I'm a father,&quot; said Prentice.\n\nPrentice said she feels alone and doesn't know what to do.\n\n&quot;If I don't have an answer, I can't do anything,&quot; said Prentice. &quot;I just have to wait.&quot;\n\nIt's a wait-and-see game for CSU students.\n\n&quot;I just hope we all pass,&quot; said Prentice.\n\nSome students, like Adrienne Longoria, have called out for help.\n\n&quot;Help us,&quot; she said. &quot;Help us students.&quot;\n\n&quot;We would have probably been able to catch up if they had let us finish our classes,&quot; said Longoria. &quot;It's just frustrating that they can't let us finish our last few weeks.&quot;\n\nLongoria and Prentice said they want to go to graduate school, but if they don't pass this semester, they could lose their scholarships.\n\n&quot;I'm trying to be hopeful,&quot; said Prentice.\n\n&quot;It's frustrating, but I don't know what to do,&quot; said Longoria. &quot;I'm hoping we get our money back and I'm hoping we can finish this semester.&quot;\n\nCSU-Pueblo officials said they don't know when transcripts will be released because each student's professor has to grade the final exams and send them to the registrar's office.]" time="1.431"><properties><property name="score" value="0.950812300046285" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Download jyrd.ptc.me Nude Selfies free porn video, hd xxx mobile porn.\n\nTorrentz will always love you. Farewell.\n\nCategory: Misc. Tagged With: hayley kiyoko jyrd nude, hayley kiyoko nude, hayley kiyoko porn.\n\ng l a m is a one stop shop for celebrity gossip news &amp; celebrity entertainment including gossip news on music gossip news on. In an interview with paper magazine on november 1, 2016, she.]" time="0.270"><properties><property name="score" value="0.9965125918388367" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99651259&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99651259
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Justin Fox\n\nIf you are the type of person who needs a car that can turn on a dime in the city but also tow your boat to the lake, the 2014 Lexus ES will probably fit the bill. When Lexus first introduced the current ES in 2006, the formula was right. But by 2012, it had become dated. The original ES was slow, thirsty and not very capable. The second generation added power and speed and was the first ES to tow. With the third generation, Lexus has taken that formula one step further and brought in all-wheel drive and a more refined driving experience. But it also kept the old formula.\n\nStyling\n\nThe 2014 Lexus ES shares its looks with the outgoing model and in a way with the new IS sedans as well. So it is a blend of a wedge shape and something more traditional. From some angles, it is vaguely reminiscent of the LFA. Its grille is pronounced, its side character lines are muscular and its curves on the hood and trunk are attractive and not overdone.\n\nIts front end is more Lexus-like than before and its tail is more sedan-like than before. It has a more modern look without being overly aggressive and its design is not overly bland like some of its competitors. It is a good looking sedan but it will be a little too plain for some people.\n\nTruck Trend's 2014 Best Interiors trophy went to the new Lexus ES. It has an attractive dash that has been borrowed from other Lexus models but it is still nice and functional.\n\nTechnology\n\nI was surprised at how good the 2014 ES was. It was comfortable and quick and the transmission was a work of art. It was easy]" time="0.866"><properties><property name="score" value="0.9996188282966614" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Pastry chef Kayoko Uematsu is a cake virtuoso. She creates works of art from her home-kitchen studio. She shared with us how to make delicious gourmet cakes. Watch and learn!\n\nHow to make gourmet cakes from your home kitchen]" time="0.139"><properties><property name="score" value="0.9840314388275146" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.98403144&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.98403144
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Did you think there\u2019s no such thing as free lunch? Wrong!\n\nI was an old punk. I ate cheaply. I ate unhealthily. I went out for food once a week, never paid more than $8 for a plate of rice and 2 pieces of chicken. Other times I ate popcorn and French Fries with my friends while watching movies in the cinema. And as I became an adult, I ate badly and grew fat.\n\nLast year, I decided to change my life. I started to save money, and from that day, I stopped eating cheaply.\n\nHere are some tips I picked up.\n\n1. Choose the cheaper meal of the day\n\nMcDonalds have only 2 types of meals that are cheaper than the rest. These are the McD Snack Meal and McD Value Meals.\n\nThe McD Snack Meal cost RM5.69, which includes a Chicken McSpicy or Filet-O-Fish, 2 Nuggets and a 375ml soft drink. The McD Value Meal costs RM8.19, which includes a Big Mac, a McSpicy or a Filet-O-Fish, 2 Nuggets and a 375ml soft drink.\n\nBoth meals are cheaper than the other meals that you can buy.\n\nThe cheapest meal of the day in McDonalds is the Cheeseburger meal.\n\n2. Eat at \u2018The\u2019 Restaurant instead\n\nIn Singapore, McDonalds is known as \u2018The\u2019 Restaurant. And you don\u2019t need to ask \u2018Where is the nearest McDonalds?\u2019, because in Singapore, \u2018The\u2019 Restaurant and McDonalds is the same thing.\n\nIt\u2019s a myth that \u2018The\u2019 Restaurant is more expensive than McDonalds. In fact, most of the meals that are available in McDonalds can be found at \u2018The\u2019 Restaurant.\n\nThe most expensive meal in \u2018The\u2019 Restaurant is the Set Dinner (Veg/Chicken). The meal costs $7.20. The most expensive meal in McDonalds is the Cheeseburger meal, which costs $4.29.\n\nAll other meals in \u2018The\u2019 Restaurant are cheaper than the other meals in McDonalds.\n\nFor example, the cheapest meal in \u2018The\u2019 Restaurant is the Curry Puff. The Curry Puff costs $3.10. The cheapest meal in McDonalds is the Cheeseburger meal, which costs $4.29.\n\n3. Eat from 7-Eleven instead\n\nThere are two reasons why I suggest you eat at 7-Eleven instead of McDonalds or \u2018The\u2019 Restaurant.\n\nThe first reason is, the food at 7-Eleven is almost as good as the food in \u2018The\u2019 Restaurant.\n\nThe second reason is, 7-Eleven sell drinks and fast food cheaper than McDonalds and \u2018The\u2019 Restaurant.\n\nFor example, I ordered a portion of Mac and Cheese in 7-Eleven. I ordered the Mac and Cheese as a take away, because I do not have any microwave in my office.\n\nThe cost of the Mac and Cheese is $3.30. And I bought a Pineapple Juice, which costs $1.30.\n\nThe total cost of my meal is $4.60. The same meal costs $4.79 in McDonalds.\n\nIt\u2019s true that the Mac and Cheese in 7-Eleven tastes a bit different from the one that I bought in McDonalds.\n\nIt\u2019s also true that the taste is not as good as the one that I can buy in \u2018The\u2019 Restaurant.\n\nBut it\u2019s true that the cost is almost the same.\n\n4. Avoid eating at the food court\n\nFood courts are expensive. The price of food sold at the food courts are almost the same as the cost of the same meal in a fast food restaurant.\n\nSo I\u2019ll suggest you avoid eating at the food court.\n\n5. Make food at home\n\nI make sure I cook once a day. I try my best to cook healthy meals. And the cost of the food that I make at home is much cheaper than the cost of a fast food meal.\n\n6. Eat with other people\n\nThe total cost of my meal depends on how many people are eating the same meal.\n\nFor example, I shared a Big Mac meal with my friend. The total cost of the meal was $8.20. That\u2019s because the meal cost $4.29 each.\n\nThe cost of the meal depends on how many people are eating the same meal. The cheaper meal is when I eat alone.\n\n7. Share the meal\n\nSometimes I share the meal with my partner. I usually share a meal with my partner once a week. I share the meal because I feel bad that my partner is eating alone while I\u2019m eating with my friends.\n\nIf I share the meal with my partner, the total cost is $4.60.\n\nIf I eat alone, the total cost is $4.79.\n\n8. Always look for a promotion\n\nIn McDonalds, \u2018The\u2019 Restaurant and 7-Eleven, there are always promotions.\n\nFor example, 7-Eleven and \u2018The\u2019 Restaurant sometimes offer discount on their drinks and fast food.\n\nSo I suggest you always look for a promotion before buying a meal.]" time="2.868"><properties><property name="score" value="0.9397337675094605" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[HapWanderlust September 1, 2011, 8:10 pm it's a very interesting product...\n\nAnonymous September 3, 2011, 12:30 am Thanks for the review, very interesting product and its feature seems to be very promising. However, I wonder what the maximum resolution would be at the time when you take picture and share it instantly, like facebook etc? Would it be just at the same resolution as of the regular picture or even better?\n\nGordon394 September 5, 2011, 12:11 pm This is one of the most intriguing products I've seen for a while. Will the quality be comparable to the smartphone quality, or will it be significantly worse? Are you sure you'll be able to get the same picture quality from the handset as with the phone tethered to a DSLR? I have to say I'm excited.\n\nMatthew September 5, 2011, 8:22 pm Thanks for the review, very interesting product. Do you know how much it will be available to the market and how it will be distributed?\n\nHiFiChris September 7, 2011, 3:56 pm Not sure about this. Love the idea but dont think it's ready. Biggest concern is image quality and would be pretty disappointing to take photos that look awful. I'd be worried about the image quality from the Nikon and would expect that some of the 'features' are just Nikon marketing. Probably best to wait until there are some real world reviews before investing.\n\nR.I. September 11, 2011, 1:16 pm This sounds amazing but what about the battery life?\n\nI really like this feature because it would be handy for traveling.\n\nPaul 4 October 25, 2011, 7:33 am would like to buy one but looks too expensive for a one off piece of kit\n\nPaul 4 October 25, 2011, 7:34 am would like to buy one but looks too expensive for a one off piece of kit]" time="0.934"><properties><property name="score" value="0.9941975772380829" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The leadership of the world is not giving a serious thought to the future of the world and they are not ready to save the world. If this condition continues then the future of the world will be very difficult to predict.\n\n\n\n\n\n\n\n\n\n\n\nThe world is facing some serious problems and a few of them are as follows:\n\n\n\n\n\nClimate change\n\nUnemployment\n\nIncome inequality\n\nHunger\n\nResource scarcity\n\nPublic health\n\nMass migration\n\nInternal conflicts\n\nAchievements of humankind\n\nThe most important achievements of humankind have been its ability to observe, to experiment, and to reason. These capabilities have given humans the power to conquer the world and to dominate everything. The presence of powerful armies of humans has protected humans from the evil of other humans.\n\nPeople are now able to live in many different places and these different places have different types of people.\n\n\n\n\n\nAre people ready to save the world?\n\nThe world is facing many problems, however, the leaders of the world are not ready to save the world. If this condition continues then the future of the world will be very difficult to predict. The world is facing many problems, however, the leaders of the world are not ready to save the world. If this condition continues then the future of the world will be very difficult to predict.\n\nThe countries which are rich in wealth and resources will enjoy a good future, but the countries which do not have wealth and resources will face very difficult situations. The climate change has become one of the biggest problems of the world and the government of the world is not taking serious actions to tackle this problem.\n\nIf the governments of the world are unable to deal with this problem then the world will suffer a lot from this problem. The fact is that the world is in very big trouble. If there is no action to save the world then the future of the world will be very difficult to predict.\n\nIn order to deal with the current problems of the world, the governments of the world will have to make some very tough decisions.]" time="0.958"><properties><property name="score" value="0.9997974634170532" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[&quot;It was not easy for me at all. I'm so used to the typical nursing home or hospital setting, it's hard to break out of that. And that's what I was used to. They're a very good hospital. My father is getting good care. They have good, knowledgeable, friendly staff. I really don't have any complaints. It's a little more expensive than some places, but not too bad. They have an excellent chef. He's excellent. The only complaint I have is that they are slow to come and get him. They don't come out and take him for a shower. We have to go get him. But we know they're busy. It's not like they can just send someone.&quot;]" time="0.362"><properties><property name="score" value="0.9998188614845276" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Definitions\n\n1. a planet in a theory of the universe that included the sun and all the other stars.\n\n2. an ancient name for the Great Bear constellation, used in astrology and mythology.]" time="0.164"><properties><property name="score" value="0.5992363691329956" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.59923637&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.59923637
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[A major renovation project at York University in Toronto is expected to run at least $26 million over budget, due to unanticipated problems that are forcing school officials to tear out and rebuild much of the work already completed.\n\nConstruction started in January 2014 on the $136-million project, which is meant to replace the York Arena, which opened in 1971, with a new, modernized facility with more seating and amenities.\n\nWork on the exterior of the arena was completed, but the contractor, Armel Holdings, noticed a problem when it came to the steel supports for the roof in June 2015, York University president Rhonda Lenton said.\n\n&quot;We realized there were some structural concerns around the roof, so we paused the project,&quot; Lenton told CBC Toronto.\n\nSteel reinforcements will be needed to strengthen the roof, which was built with &quot;soft&quot; steel. (York University)\n\nWorkers discovered the roof of the arena had been built with &quot;soft&quot; steel, which is not as strong as the &quot;hard&quot; steel that is typically used for these types of projects, said Lenton.\n\nIn order to make sure the project is safe, crews need to rip out and rebuild about 50 per cent of the roof, at an estimated cost of $16 million.\n\n&quot;We\u2019ve got to take this roof down,&quot; said Lenton.\n\n&quot;We will be replacing the steel \u2014 the beams \u2014 and re-engineering the entire roof structure, as well as all the supporting beams that support the roof.&quot;\n\nSchool officials say they will be applying for additional funding to cover the cost of the roof replacement.\n\n&quot;We will be going to the province for the additional funding that we will need, for sure,&quot; Lenton said.\n\nThey are also hoping to get money from the York University Foundation, which is a non-profit fundraising organization that operates separately from the university, according to its website.\n\nShe said she does not believe taxpayers will have to foot the bill for the additional cost.\n\nUnforeseen problems\n\nYork University president Rhonda Lenton says she does not believe taxpayers will have to cover the $16-million cost of rebuilding the roof. (CBC)\n\nThe problems do not end with the roof.\n\n&quot;We realized that there were some structural issues around the seating, and we\u2019re replacing the seating, the wood seating in the arena,&quot; Lenton said.\n\nIn total, there will be 12 new levels added to the arena, along with a new heating and cooling system.\n\n&quot;I think, because of the complexity of this building, we have to anticipate what other unforeseen problems may come along,&quot; Lenton said.\n\nIn December 2015, school officials sent a memo to university faculty and staff outlining the potential for &quot;additional costs and associated risks&quot; that could arise in the future.\n\n&quot;We will continue to keep all our stakeholders updated as the project progresses and when we know more details about timing and scope of work,&quot; the memo said.\n\n&quot;I\u2019m just waiting for more details,&quot; said Brian Lee, a York University professor and co-chair of the university\u2019s Local 79 union.\n\nLee said he thinks the university should have been more specific about the challenges facing the project.\n\n&quot;There is no need for it to be kept secret,&quot; Lee said.\n\n&quot;There\u2019s nothing secretive about construction projects.&quot;\n\nScheduling setback\n\nThe timeline for the completion of the project has been extended by more than a year, said Lenton.\n\nThe university had initially planned to have the arena ready by the 2017-2018 school year, but that deadline has been pushed back to the 2018-2019 school year.\n\n&quot;We did have an aggressive schedule. We have to replace a roof in a building that is up on a hill,&quot; Lenton said.\n\nLenton also said there will be more parking spaces at the arena once the project is complete. The university had to remove between 300 and 400 parking spaces in order to begin the work.\n\n&quot;I think we will have more spaces than we had before,&quot; Lenton said.]" time="2.128"><properties><property name="score" value="0.9260357320308685" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[October 19, 2015\n\nOnce again we find ourselves reflecting on our own good fortune as we consider the news from Paris. We are not only alive, but many of us have access to a wealth of luxury, both intellectual and material, that few of our ancestors would have dreamed of. And we also have access to media that have transformed the world. One of my sons recently discovered the fascinating series \u201cHow the Universe Works\u201d (available on Netflix). This is a television show, with a high-tech production budget, and it is fascinating. It has some of the best images of the stars and planets ever produced, with an engaging story to go with them. The soundtrack alone is amazing. I highly recommend it. But it only works because of our technology, which has the power to tell stories across the globe.\n\nYou don\u2019t have to leave the planet for science to reach you. It\u2019s right there in your living room.\n\nToday I read a story in the paper about a study that examined the safety of exercise and high protein diets. This study was designed to answer a question that most of us have faced at some point: should I cut back on protein? My story pointed out that the study was very small and inconclusive, but the real story is about how we learn about our world and why this study is inconclusive. It is a story that has been repeated again and again. When new information comes out, we all read it with the same kind of judgment that you would apply to a restaurant: is this place worth my time and money? The more of it you eat, the more of it you want to read. When I read the story in the paper, I didn\u2019t try to understand the process by which the authors made their conclusions. Instead, I simply had a gut reaction: was the story worth the cost of the information it contained?\n\nThat is the only real question that matters when we choose what to read. This question is especially important today because there is so much information out there. When the costs of information are low, it becomes much easier to consume, and with this abundance comes the difficulty of choosing what to read. So we focus on our gut. And this is because we have a brain designed to make these kinds of choices, one that is driven by emotional needs, not rational ones. We are not designed to rationally evaluate the content of each article, but instead we are designed to make emotional decisions about what to consume.\n\nThis point was well illustrated by some research done by scientists at the University of Michigan. They had two groups of subjects: one that was told to read a story that was fun and amusing, and one that was told to read a story that was sad. Then they had them make some choices about the food they would eat for the next week. The group that had been reading the sad story ate 25% fewer calories than the group that had been reading the amusing story. Why? Well, they thought that food was a reward, and since they had been unhappy in the experiment, they cut back on the number of rewards that they would give themselves. This is a very rational process, but the very fact that they were not looking at the process that led them to their choice means that it had to be some form of emotional response.\n\nIn this way, we respond to information much like we respond to food. Food provides us with energy and pleasure, so it is important to choose it well. So we think in terms of gut decisions, not because they are the best ones, but because they help us to cope with our lives. Information is also a source of energy and pleasure, and so we choose it as well.\n\nBut while we are creatures that make emotional choices about our food and information, we also have other needs. These are needs that were emphasized by the process of evolution: needs to know about the world and needs to know about ourselves. So when we read articles, or watch television, or go on social media, we need to know whether they provide this type of knowledge as well.\n\nMost of the time we are looking for a story that is interesting, and that will hold our attention. But we also want a story that tells us something about our world. The better the story is, the more likely we are to learn from it. A story about calories and exercise might not be particularly interesting, but it is important for us to know. A story about the stars and planets might be fascinating, but it doesn\u2019t matter if it is a good story or not.\n\nThis is because we need information about our world to survive and to thrive, and we know this to be true because it is encoded in our DNA. Our brains have evolved to help us process the information that we receive, and so it makes sense that we]" time="2.391"><properties><property name="score" value="0.9998087286949158" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Official Download Source - Manufacturer Website\n\nWindows ME, Windows 98\n\nSoundMAX Integrated Digital Audio Driver\n\nThis package supports the following audio controllers:\n\n- SoundMAX Integrated Digital Audio - Version 6.0.11.0601c - 5.5 MB]" time="0.196"><properties><property name="score" value="0.8287380337715149" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.82873803&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.82873803
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[By \u201cspecial\u201d, I mean, \u201ca man who plays a very particular role\u201d. This man is a type of \u201cperformer\u201d, a type that is often put forward in business culture and in the literature of business leadership, and many men admire it and want to be seen as this type of man. It is the man who is never at a loss for words, and always knows exactly what to say, and is a good listener too. He has an instinctive understanding of people\u2019s moods and emotions. He is the man with the plan. He knows what\u2019s going on. He has a direct influence on others. He is never without a list of bullet-points, even if he\u2019s just going into the shop. He\u2019s good at handling pressure and he\u2019s a team player. He is the man who is looked to for a reaction.\n\nYou can call this type of man many different things. I am going to call him a \u201cbusiness executive\u201d, as it is most widely-used, but it is also a term that is rather different from the role of \u201cbusiness leader\u201d, and which I will also address. I am a business executive, and I have been a business executive for as long as I can remember, and I\u2019m a fairly typical example of a business executive. I think that most people reading this will find that it accurately describes at least a few men that they know.\n\nMost men aspire to this particular role. Some men even feel a burning need to play it. They will not be happy or feel satisfied in life unless they are in this particular role. They often judge themselves by it. They have a high level of need and drive to be a business executive.\n\nBeing a business executive is one of the \u201cgood\u201d ways to be a man, it is one of the ways in which men \u201cmake a difference\u201d and \u201chave an impact\u201d. It is also one of the ways in which men achieve \u201crespect\u201d, \u201crespect\u201d being an essential element in most men\u2019s internal lives. Many men do feel that it is through this role that they will achieve respect, in the same way that they might feel that they achieve respect in the rest of their life. They might not be consciously aware of this, and they may have a lot of other reasons for doing what they do, but respect is there, and is an important part of their motivation.\n\nThis type of role is one of the ways in which men achieve success and status. It is also one of the ways in which they can \u201cgain influence\u201d and \u201cchange things\u201d. It is the way that they \u201chelp people\u201d and \u201cchange things\u201d and \u201cmake a difference\u201d. It is the way that men can \u201chave a great career\u201d, and achieve \u201csuccess\u201d, and \u201cbe respected\u201d, and \u201cmake money\u201d.\n\nIt is not necessary to believe that these are the only or even the most important reasons for having a job. In my experience, most people, male and female, care about these things to some degree or other. I am not attempting to persuade anyone that these are \u201cunimportant\u201d things, I am not setting out to denigrate them. But it is important to understand that these are reasons for working. People, both male and female, need these things, and it is to these needs that they often respond when asked why they work.\n\nBusiness executives often find it easy to get good work. Many people are happy to trust their judgement, because they are good at understanding the people around them, and they can provide answers to difficult questions. They often find it easy to get good work, and they may have a number of people to choose from when looking for a new job.\n\nThey often feel that they are \u201cjustified\u201d and \u201chave a right\u201d to be in the position that they are in. They have earned it.\n\nThis type of role is very similar to that of \u201cbusiness leader\u201d, but these are very different roles. Many people confuse them, and use them as synonyms, and they are often referred to as such in business literature and the popular media. They are not synonyms. They are two different roles, and their differences can be seen in the relationship of the people around them.\n\nThe business executive is a man who is often very popular and well-liked. He is a man who is often admired. He is often seen as the man to turn to for answers. People come to him to find out what is going on, and to find out what to do. He often knows what to say, and he has an instinctive understanding of people\u2019s moods and emotions, and a quick and accurate understanding of the situations that he is faced with. He is a man who is often used by other people as a sounding-board, as a person to talk things over with, as a man to provide feedback and reaction.\n\nBut this is often not the case with the business leader. He often gets into trouble with the people around him, and he often gets an awkward response to his advice. He is not often looked to for feedback or reactions. He often does not feel that he has the understanding that he wants, and that he needs, and that he feels he deserves. He often feels that he is not looked to as much as he would like, and that people would turn to someone else if he were not there.\n\nThis type of role is the most important in a business organisation. It is not a role that most men can aspire to. In fact, I would say that most men do not aspire to it. And for many, if not most, men it is not a role that they can play well.\n\nFor many men, playing this role will lead them into unhappiness. They will feel frustrated and unhappy, and it is often a role that causes men to turn against the organisation that they are working for.\n\nMany men who try to be business executives will often find themselves in a difficult situation. Many men will find that they are working for organisations that do not have a business executive position. They will find that they are attempting to play a role that does not exist in the organisation that they are working for.\n\nMany men are unhappy when they try to be business executives. Many of them will be unhappy if they do not play that role.\n\nThere are, of course, many other types of men, and many other types of roles that they play. But this particular type of man and role is a very special one.\n\nA case study of a \u201cbusiness executive\u201d\n\nThis is a man who is very good at his job, a very clever and able man, a man who has an instinctive understanding of people\u2019s moods and emotions, and who has a natural ability to speak effectively and confidently to other people. He has a good understanding of other people, and he has an ability to work with other people. He understands other people\u2019s motivations and he understands what is important to them. He has a good understanding of what people will do, what they will say, and what they will not say, and he understands what is going to happen and how things will turn out. He understands people\u2019s moods and emotions. He is quick to understand how things are going to turn out, and he is quick to make decisions.\n\nThis man is a good communicator. He is the type of man who people turn to for reactions. He is the man who knows what to say, and who knows how to say it. He is good at dealing with difficult people, and he can handle difficult situations. He is good at explaining things to other people. He is good at handling pressure, he is good at dealing with crises, and he is good at giving people confidence.\n\nThis man has a good understanding of his organisation, and he has a good understanding of his work. He is often the man that people go to when they need information. He is the man who has a good overview of the situation.\n\nThis man has a good understanding of his organisation\u2019s finances. He has a good understanding of his organisation\u2019s future, and he can see things that other people cannot see. He is the man who knows what is going to happen. He is a man who has the answers, and a man who knows what to do. He has the information that other people need. He has an instinctive understanding of his organisation\u2019s situation. He has a good overview of the situation. He is good at taking decisions.\n\nThis man can be effective at many things. He is good at working with other people. He is good at providing feedback. He is good at explaining things to people. He is good at providing support and comfort to people. He is good at listening to people and taking them seriously. He is good at convincing people. He is good at getting people to agree to things. He is good at persuading people. He is good at making plans. He is good at using his initiative. He is good at seeing what is going on. He is good at moving things forward.\n\nThis man is used to being respected, and he is used to people listening to him. He is used to being popular. He is used to having an influence on people. He is used to]" time="5.247"><properties><property name="score" value="0.9997902380095588" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[ABC News Good Morning America ABC Feb 10, 2013 7:00am PST\n\n. he is married to the former martha lee hawkins. he has two children, bill and samantha. &gt;&gt;&gt; it is an amazing story about someone you may have never heard of. he wrote a best-seller and he just turned 13. abc's ginger zee is here with more. good morning, ginger. &gt;&gt; reporter: good morning, robin. allen barra who is not quite yet a teenager, he is a writer and he has a deal with harpercollins. &gt;&gt; i have been interested in writing for a very long time. i was always trying to write different kinds of books. &gt;&gt; when he was 10, he started a diary. he had it bound and published. &gt;&gt; it was for a school project but the book was an instant hit. &quot;i wrote the diary of a wimpy kid&quot; was one of the top selling books of 2010. the second book, &quot;the third wheel&quot; was a number one best seller. he and his sister had started the website, kidwrite.com to promote their stories. now at 13, he has a deal for a third book. he plans to use the money to pay for college. &gt;&gt; i'm hoping to go to brown and to study international relations and history. &gt;&gt; his publisher is thrilled with his success. &gt;&gt; he has a rare combination of]" time="0.736"><properties><property name="score" value="0.9283396601676941" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[This article is about the combat ability. For the status effect, see Poison\n\nFor the Hunter creature, see King black dragon\n\nVenom is a poisonous status effect inflicted by many creatures, and is inflicted by the dragon dagger special move. It deals low damage over time, reducing the player's defence level, and decreasing the effectiveness of protection prayers and similar effects. There is no venom available as a spell in RuneScape.\n\nWhen venom is applied, a timer of 16 seconds begins and the afflicted player receives the message &quot;You are beginning to feel venomous!&quot; This timer can be frozen using either an antipoison, or an antifire potion if the dragon dagger special attack is used against the player.\n\nContents show]\n\nSigns and symptoms Edit\n\nIf you get hit by poison, you may notice the following symptoms:\n\nInability to eat while venom is active (unless the player is using food that heals life points only)\n\nInability to cure poison with antipoison or heal poison using poison cures whilst the venom is active (unless the player is using antipoisons or super antipoisons)\n\nA decrease in defence level\n\nVenom damage Edit\n\nPoison has a poison damage value of 10 + 4 * n {\\displaystyle 10+4*n} , where n {\\displaystyle n} is the number of hits received. This means that the average damage of a venom hit is 22.5% higher than the hit's base damage. The maximum poison damage is 120 + 10 * n {\\displaystyle 120+10*n} , which occurs when the player is hit 16 times by venom.\n\nAs with other damage over time effects, poison damage is not affected by damage modifiers or protection prayers.\n\nResistance Edit\n\nThe player's resistance to venom is related to the player's Defence level, and reduces the damage done by venom accordingly. A player with level 90 Defence or higher will be almost immune to the venom of most dragons, while those with a lower Defence level will take considerable damage from even low-level dragons. The dragon dagger's special attack provides temporary immunity to poison from venom, however this immunity is lost when the dagger's special attack is used up, or if the player unequips the dagger.\n\nTrivia Edit\n\nEven though the Venom poison effect was not in RuneScape Classic, a poisoned player would say the same thing: &quot;I feel a little woozy...&quot;]" time="1.226"><properties><property name="score" value="0.46894729137420654" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.46894729&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.46894729
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Wednesday:\n\nBrought two back, used one, but my power supply has burned out on the other one. Had to bring it home today and I will exchange it tomorrow.\n\nUsed two dead batteries on the other camera tonight for a special event. Took many pictures. I will go home and try to fix it before we go out again tonight.\n\n\n\nThursday:\n\nToday is a bad day, my power supply for my new camera has burned out. I had to send my new camera home with my wife. I now have no camera to use for tomorrow. I am hoping that I can have my son pick up a battery for me and return my wife's camera to me tomorrow.\n\n\n\nI am going to return the camera tomorrow and if it isn't under warranty and it doesn't fix my power supply, I am getting a new one. This is not a good way to start off the new camera.]" time="0.468"><properties><property name="score" value="0.9997995495796204" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Description\n\nOn 2 June 1945 in the Tiergarten of Berlin, hours after Hitler had committed suicide, Major-General Hans Krebs (with his right hand raised) signed the German instrument of surrender. The capitulation of all German forces to the Allies was unconditional.The Reich Chancellery in Berlin was chosen for the signing. Soviet leader Joseph Stalin had insisted on a place in Berlin, his army had reached the city ahead of the others, and he had expressed a wish to receive the capitulation in his capital. The same location had been used for the signing of the surrender of Germany by the Russian army in November 1918. The others present were the Soviet military commander Marshal Georgi Zhukov, U.S. General of the Army Dwight D. Eisenhower, and British Field Marshal Sir Bernard Law Montgomery, whose signatures appear at the bottom. This moment is often said to be the end of World War II, but is not the date used for official purposes. The terms of the surrender were finally worked out in Esen, on the river Mulde, between 24 and 27 April 1945, and approved by Eisenhower and Zhukov on 4 May. The official document (the text of which is in the final picture) was signed in Berlin in the evening of 8 May. It was ratified by the Soviet government on 1 September 1945 and by the other allies later.]" time="0.661"><properties><property name="score" value="0.8157024681568146" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.81570247&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.81570247
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[About \u201c[Go]\u201d\n\nDoo Wop\u2013esque song by The Weeknd.\n\nIts sound is very similar to the 50s and 60s; therefore it is likely that this song would have been the introduction to The Weeknd\u2019s debut mixtape, \u201cHouse of Balloons.\u201d\n\nThe song was written and produced by The Weeknd, with co-production by Doc McKinney and Illangelo.\n\nThe Weeknd has stated in several interviews that the song was written about one of his ex-girlfriends.]" time="0.300"><properties><property name="score" value="0.9997685551643372" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[\n\nSenator KROGER (Minister for Population and Immigration) (10:46 AM) \u2014I seek leave to make a short statement.\n\n\n\nThe PRESIDENT \u2014Leave is granted for one minute.\n\n\n\nSenator KROGER \u2014The Leader of the Opposition indicated yesterday that he would be referring to the legal framework in relation to the privatisation of Telstra. This is not surprising given that the issue of Telstra was discussed at the last election and he has been a long-time supporter of privatisation.\n\nIt is interesting to note, though, that it is the opposition who have failed to produce their privatisation plan\u2014not that they have one. And they have failed to produce it for one simple reason: it will have to meet with the ACCC guidelines in relation to Telstra. I was very happy to provide those guidelines to the opposition. I can provide them again this morning if they want.\n\n\n\nSenator Campbell \u2014We know you are the legal supremo.\n\n\n\nThe PRESIDENT \u2014Senator Krueger, on your own account, to get a bit of latitude\u2014I think you have five minutes now.\n\n\n\nSenator KROGER \u2014Thank you, Mr President. What I can do is provide them again this morning, but if the opposition has the same rhetoric it has used at the last election\u2014and I know it is used at every election by the leader of the opposition\u2014then it will fail to have its proposal examined in any great detail by the ACCC. It is not because we have an axe to grind, it is because the opposition have an axe to grind. But it does not pass the ACCC test, and it is interesting that they do not want to have that detailed examination of their proposal.\n\nLet us have a look at what Telstra said yesterday. The headline, of course, in the Australian Financial Review is \u2018Telstra asks regulator to speed up broadband overhaul\u2019. You would be very happy about that, Senator, wouldn\u2019t you?\n\n\n\nSenator Campbell \u2014We are.\n\n\n\nSenator KROGER \u2014The point is that you have not said what the rates are. But if you had done your job in the last election and had produced your plan, then the ACCC would have had the option to have you examine your plan and to determine what the rates should be and whether they could be extended. What we have done is ensure that the interest of consumers and business in this country are protected. We have done it by putting in a position where the ACCC will have the opportunity to examine any privatisation proposal that the opposition has.\n\nAnd, Mr President, there are a number of aspects of Telstra that have been raised by the opposition. They are in the policy documents that the opposition are putting around. There is no doubt that those aspects of Telstra will be examined by the ACCC, as will be anything else they may have up their sleeve.]" time="1.484"><properties><property name="score" value="0.4426837960879008" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[This day started like any other. I was getting ready for school. I went through my routine; shower, shave, towel dry, and run a comb through my hair. As I ran the comb through my hair, I noticed a small bump on my head]" time="0.210"><properties><property name="score" value="0.582988977432251" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The government has set up a panel to study the RBI's recommendation on high-denomination currency notes, Economic Affairs Secretary Shaktikanta Das said on Monday.\n\n\n\nThe high-powered committee will take a decision after it gives its recommendations, he said.\n\n\n\nThe Reserve Bank had last week recommended withdrawal of Rs 500 and Rs 1,000 currency notes to curb funding of terrorism and tackle the problem of counterfeit notes.\n\n\n\nThe decision to set up the committee was taken at a high-level meeting chaired by Finance Minister Arun Jaitley and attended by top officials from the Finance Ministry and the Reserve Bank.\n\n\n\nDas said the government would take a decision on the recommendations of the RBI only after the panel submits its report.\n\n\n\n&quot;The committee has been set up, which will look into the matter and come up with its recommendations. The government will take a call only after the recommendations are submitted,&quot; Das said.\n\n\n\nAsked if the panel would be headed by a member of the Prime Minister's Office (PMO), he said: &quot;No, it is headed by a senior Finance Ministry official and members of the RBI.&quot;\n\n\n\nThe panel is likely to comprise senior officers from the Finance Ministry and RBI.\n\n\n\nThe panel is likely to meet on Tuesday, sources said.\n\n\n\nOn a query on the option to withdraw the new series of Rs 500 and Rs 2,000 notes, Das said: &quot;It is a hypothetical question.\n\n\n\n&quot;The government will take a call on the RBI's recommendations only after the panel submits its recommendations.&quot;\n\n\n\nIn a bid to curb funding of terror, the RBI had recommended withdrawal of the bank notes of Rs 500 and Rs 1,000 denominations.]" time="0.749"><properties><property name="score" value="0.9998033046722412" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Social media company Twitter Inc said it suspended about 70 million accounts in the past quarter to curb abusive behaviour, but its active user base grew about one per cent in the same period, disappointing investors.\n\n[SAN FRANCISCO] Social media company Twitter Inc said it suspended about 70 million accounts in the past quarter to curb abusive behaviour, but its active user base grew about one per cent in the same period, disappointing investors.\n\nThe company said it has seen continued &quot;abuse, harassment, and other kinds of behaviours that distort and distract from the public conversation.&quot;\n\n&quot;The best defence against these challenges is a combination of technology and people,&quot; Twitter chief executive Jack Dorsey said in a statement.\n\nThe company said it also saw some growth in its advertising base, adding that it expected to generate US$650 million in revenue in the fourth quarter, topping market expectations.\n\nsentifi.com Market voices on:\n\nTwitter also announced it would begin notifying users of covert political influence efforts, after having blocked such suspicious activity in the past.\n\nThe changes came as the company reported a profit of US$91 million for the third quarter, which was in line with market expectations.\n\nTwitter shares dipped slightly in after-hours trade, but gained 2.6 per cent over the last 12 months.\n\nTwitter said it saw a 1 per cent user growth in the third quarter, adding three million monthly active users.\n\nThe company's daily active users grew 12 per cent year-on-year, a modest increase as it continues to grapple with efforts to combat abusive behavior and harassment.\n\nThe company also said it would start to notify users of &quot;malicious automated accounts&quot; in a move that comes after Facebook and Google have taken similar steps.\n\nIn June, Twitter updated its terms of service and privacy policy to clarify its rules on the use of bots, which are a favourite of fraudsters and malicious actors, such as Russian propagandists.\n\nTwitter said it &quot;dramatically reduced&quot; the number of suspicious log-ins that could have led to accounts being compromised.\n\nWhile the company continues to suspend millions of accounts, critics have complained that the company has not done enough to police activity on the platform.\n\nOn Thursday, Mr Dorsey said Twitter would &quot;double-down on the most important factor of our success: You.&quot;\n\nWhile he said the company did not expect to be free of abuse or misuse, it wanted to ensure the conversation on Twitter remained healthy.\n\nTwitter said the problem was particularly severe in private, one-to-one conversations, where the rate of harassment was five times higher than the regular feed.\n\n&quot;We want to increase the collective health, openness, and civility of public conversation, and to hold ourselves accountable towards progress,&quot; said Twitter's chief executive.\n\n&quot;We are making progress as we go.&quot;\n\nAFP]" time="1.439"><properties><property name="score" value="0.9974818825721741" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Culture of Empathy: Bats and Bells\n\nCulture of Empathy: Bats and Bells\n\nYou are likely familiar with the cultural myths surrounding the bat, often involving vampires and bloodsuckers, evil demons or even winged creatures that steal one\u2019s soul. In the Western world, the bat has for centuries, and for many people even today, been associated with negativity and danger.\n\nThe bat, on the other hand, is a highly empathetic creature. Bats have amazing social connections and perform complex rituals that ensure the welfare of their group members. They are amazing hunters of both insects and fruit, thanks to their very good eyesight and a large brain for processing data. Because of their social abilities and prey-catching skills, the bat has been hunted and killed.\n\nAccording to the Bats of Trinidad and Tobago website, \u201cSince bats are prey for many animals, the loss of bat populations due to predation is widespread. Predatory birds and snakes, for example, may take a large number of bats from a colony.\u201d\n\nI often wish for such care, consideration and connection when I\u2019m interacting with other people. I often wonder what it would be like to have relationships built on respect, reciprocity and trust. My hope is that someday all the people in the world will connect with their hearts and brains, and practice empathetic responses to others.\n\nThough a long time ago, I have even heard the birds and bats say\u2026\n\nCome on baby, light my fire.\n\nI wanna be burned, and baby, then you\u2019ll desire.\n\nTo my readers, I\u2019d like to introduce a new piece of music that I recorded on a ukulele. I\u2019m not sure if I\u2019ve mentioned this before, but my instrument of choice is the ukulele. I\u2019ve been studying classical guitar for a few years and I\u2019m not so sure I\u2019ll ever play anything other than the uke.\n\nThe song I am introducing is a tune that came to me as I was studying bats for my Culture of Empathy book. I was curious to learn more about the bat\u2019s in Trinidad and Tobago. Their indigenous names include:\n\nBakabak (Carib)\n\nAlik-Bakabak (Bhojpuri)\n\nBakabak-a (Bhojpuri)\n\nBakabak-fefa (Creole)\n\nBakabak-fura (Creole)\n\nBakabak-ki-bot (Creole)\n\nBakabak-ra (Creole)\n\nBakabak-rah (Creole)\n\nBakabak-rah-alik-bakabak (Creole)\n\nBakabak-rah-mich (Creole)\n\nBakabak-rah-te (Creole)\n\nBakabak-ra-te (Creole)\n\nBakabak-ra-te-bakabak-ra (Creole)\n\nBakabak-ra-te-pumpu-bakabak-ra (Creole)\n\nBakabak-ra-te-pumpu-alik-bakabak-ra (Creole)\n\nBakabak-ra-te-pumpu-bakabak-ra-alik-bakabak-ra (Creole)\n\nBakabak-ra-te-pumpu-mich-alik-bakabak-ra (Creole)\n\nBakabak-rah-te-pumpu-mich-alik-bakabak-ra (Creole)\n\nBakabak-rah-te-te-mich-alik-bakabak-ra (Creole)\n\nBakabak-tite (Creole)\n\nKaiyot (Indian)\n\nMalik-kabaak (Indian)\n\nMakai-kuar (Indian)\n\nMas (Indian)\n\nPumpu (Indian)\n\nUkuar (Indian)\n\nXalk-ra-te (Creole)\n\nXalk-rah-te (Creole)\n\nZokak-ra-te (Creole)\n\nZokak-rah-te (Creole)\n\nZokak-rah-te-pumpu-alik-bakabak-ra (Creole)\n\nZokak-rah-te-pumpu-bakabak-ra-alik-bakabak-ra (Creole)\n\nZokak-rah-te-pumpu-bakabak-ra-te-pumpu-alik-bakabak-ra (Creole)\n\nZokak-rah-te-pumpu-bakabak-ra-te-pump]" time="1.269"><failure message="RuntimeError: The expanded size of the tensor (547) must match the existing size (514) at non-singleton dimension 1.  Target sizes: [1, 547].  Tensor sizes: [1, 514]">E   RuntimeError: The expanded size of the tensor (547) must match the existing size (514) at non-singleton dimension 1.  Target sizes: [1, 547].  Tensor sizes: [1, 514]</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Shopper They came so fast, the colour is more bright than I expected but the quality is nice, my size is M and it was very good\n\nD***a I ordered the size M, took the XXL. Perfectly fit the body and did not stretch. Came to the moscow region for a month. I like the color. On the parameters og 84-61-86-88-chest 107.\n\nE***a Delivery is fast. The size is chosen by the table.\n\nG***o I haven't received the product yet, but the customer service of the seller is really good.\n\nA***n It's for the first time that I ordered something from this seller. The order was done at 11:40, and I already received it at 18:08, the parcel came from China to Romania in 4 days! The products are high quality, and the sizes are pretty well chosen. I recommend the store and will buy from it again!\n\nE***a Very cool dress, ordered as a gift, but not what i expected. Ordered and delivered quickly, but i liked more with a tie on the neck.\n\nM***s Perfect. I am a little large and ordered a size bigger so it will be nice and loose\n\nO***o I ordered two sizes bigger. Came up perfectly. I recommend to order more size for a comfortable fit.\n\nM***i Very good quality, it's a little big on the shoulders but it's ok :)\n\nA***a Quality is excellent. Everything is neatly stitched, i advise you to buy!\n\nG***g I am satisfied, beautiful fabric, high quality, very comfortable, thanks for the gift.\n\nO***e Same as the picture. Very cute and perfect fit! Recommend\n\nT***o So beautiful and cool i love it\n\nA***z I recommend the product. Come nicely wrapped and the material is good. Very nice\n\nD***s Perfect same as the picture, it arrived very fast to spain\n\nS***e The dress is perfect, the colour is exactly as the picture. It's also nice and soft. It was a gift for my mom and she is delighted!\n\nC***o The product is perfect. I am super happy with it!\n\nT***t My daughter is happy with the product, arrived in spain in 30 days, thanks!\n\nD***a Looks like the picture. Very good product, very comfortable\n\nG***e super, ty\u0142ek takiej wyjdzie dobrze umiem zaryzykowa\u0107 ;)\n\nM***a Perfect. This is the second time i bought this product, and i am still very happy. Thank you\n\nT***t This dress was the same as the photo. It was a gift for my mother and she was super happy.\n\nG***n I liked the product and it arrived in perfect condition.\n\nL***i Beautiful. This is the second time i buy it, very beautiful.\n\nC***t Love it!!!\n\nS***n The order was received very fast. The product is exactly the same as the description. It's very beautiful, quality, nice, soft and comfortable. Thank you very much!\n\nA***n Ordering more colors\n\nB***t Very good fabric\n\nD***o I got the dress on time, fits good, the colour is accurate. Thanks\n\nB***r Great quality and looks exactly like the picture.\n\nA***h I am very satisfied with the product, the only thing is that the photo in the arm is white. I do not think it is white in reality. Everything else is very good.\n\nM***y Very cool dress! Looks very good on. Good fabric! Very fast delivery!\n\nD***a Fast shipping and a good product! I'm satisfied. Thank you very much.\n\nShopper Very good quality. Thank you seller, the product is the same as the photo.\n\nN***n The order arrived on time. The dress is good. The material is pleasant. I liked it. I recommend the product. Thank you.\n\nShopper The product arrived very fast, the product is exactly like the photo, i recommend the product, it came with a good packaging.\n\nL***y Very good quality, the fabric is pleasant, thank you!\n\nShopper Product matches the description, quality at altitude\n\nG***a Fast shipping. Very good quality. Very comfortable.]" time="2.318"><properties><property name="score" value="0.49383312463760376" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Washington (CNN) On his way to a diplomatic crisis with Qatar, President Donald Trump spent part of the day Tuesday praising a trio of Middle East leaders who are strong allies in the fight against ISIS and terrorism, as well as helping to crack down on Iran.\n\nIn the course of a 50-minute speech in Saudi Arabia to a group of leaders of more than 50 Muslim countries, Trump mentioned the names of three leaders: King Salman of Saudi Arabia, King Faisal of Iraq and President Abdel Fattah el-Sisi of Egypt. He described them as &quot;friends&quot; who had &quot;brought calm&quot; to the region.\n\n&quot;This is a battle between barbaric criminals who seek to obliterate human life, and decent people of all religions who seek to protect it. This is a battle between good and evil,&quot; Trump said, according to a White House pool report. &quot;If we do not stand in uniform condemnation of this killing, then not only will we be judged by our people, not only will we be judged by history, but we will be judged by God. This is not a battle between different faiths, different sects or different civilizations. This is a battle between barbaric criminals who seek to obliterate human life, and decent people of all religions who seek to protect it. This is a battle between good and evil.&quot;\n\nIt's an evolution for Trump, who had attacked the trio of Middle East leaders during the campaign and, after he won, avoided being too complimentary about them.\n\nAnd in the days before his speech, Trump had been on the attack again against Qatar. He had bashed the Gulf country in a series of tweets for alleged &quot;funding of terrorism,&quot; which triggered the diplomatic rift. Trump went to Saudi Arabia on his first trip abroad as President.\n\nTrump has backed the Saudi-led boycott of Qatar and threatened &quot;severe punishment&quot; against Doha for its alleged &quot;sponsorship of terrorism,&quot; according to a Qatari official who spoke with CNN.\n\nTrump's comments about Salman and Faisal came just after he spoke about the Iran nuclear deal. Trump said it was an &quot;embarrassment to the United States&quot; and called it &quot;one of the worst and most one-sided transactions the United States has ever entered into.&quot;\n\nIt is also not the first time Trump has praised King Salman. After his inauguration, Trump made a phone call to the king. He called him a &quot;very wise person who wants to see things get much better rapidly&quot; and called their relationship &quot;extremely good.&quot;\n\nKing Salman was the first world leader to speak to Trump after he was inaugurated, and Salman joined Trump in a visit to a home in Riyadh that was built for families of the victims of the 9/11 terrorist attacks in the US.\n\nSaudi Arabia's King Faisal was assassinated in 1975.\n\nBut despite that history, Trump has only been complimentary about the Saudis, but not in the way some have accused him of being too cozy with Russian President Vladimir Putin. The Saudis helped him create the first page in his Middle East policy.\n\nHe praised the Saudis after he won the election, saying, &quot;Saudi Arabia and I get along great.&quot; He had spoken about it as a campaign issue, saying in March 2016 that he &quot;would like to protect Saudi Arabia&quot; but that the country is &quot;going to have to help us economically&quot; in the fight against ISIS.\n\nTrump had also said in December 2015 that he was against the &quot;concept of bombing&quot; and was for a strong intelligence-gathering effort instead. He said then, &quot;It is a mistake to bomb the oil fields. You bomb the oil fields, you kill civilians, you destroy the oil fields. They are going to get into something else. They are going to produce and sell it someplace else. It will not be a pretty picture.&quot;\n\nAt the time, Trump was being accused of being naive about the region and for giving a nod to terrorists when he questioned whether the US should be spending money and lives to intervene in Syria.\n\nTrump has been criticized for appearing to blame Qatar for the current state of affairs in the region. Some, like the U.S. ambassador to Qatar, have said Trump is trying to improve ties with Qatar in order to fight terrorism.\n\nTrump also spoke during his visit to Saudi Arabia about his support for the monarchy.\n\n&quot;We are not here to lecture. We are not here to tell other people how to live, what to do, who to be or how to worship. Instead, we are here to offer partnership, based on shared interests and values,&quot; Trump said. &quot;We are not here to tell other people how to live, what to do, who to be or how to worship. Instead, we are here to offer partnership, based on shared interests and values.&quot;]" time="2.293"><properties><property name="score" value="0.9982662320137023" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Photo: Courtney Marsh\n\nThe line to get into Drai\u2019s Beach Club on Monday night stretched halfway around the block, which was a good thing.\n\nI don\u2019t mind waiting in line for stuff. But it\u2019s better if it\u2019s worth it.\n\nInside Drai\u2019s, which is located in The Cromwell on the rooftop of that casino, there was enough space to be lost in, and I used it to hide for a bit to observe the crowds and]" time="0.278"><properties><property name="score" value="0.8747095465660095" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Dei aperti oggi dalla Regione Piemonte gli uffici per il trasferimento delle attivit\xe0 dalla Sic - Societ\xe0 Industria Costruzioni - al costituendo cda di Sistri - Sistema Territoriale della ricostruzione e riqualificazione idrogeologica, in sintesi una sorta di Anas dei Comuni. Le sezioni di Pinerolo, Chieri, Torino e Moncalieri sono state create per i cittadini in transito, anche se in questo momento non ci sono pendenze che li richiedono. La sede di Pinerolo \xe8 presso la sede di Regione Piemonte in via Monsignor Forno, 2, ma l'appuntamento \xe8 dalle ore 10 alle ore 13, dal luned\xec al venerd\xec. Le altre saranno gestite da Bologna, dove in via Crispi al n.8 e via Mascagni n.6. Lo staff incaricato per la giornata di oggi prender\xe0 i certificati in essere in Sic e, qualora non si tratti di ricostruzione, provveder\xe0 a stamperli in copia fotostatica, inserire nella pratica un modulo che costituir\xe0 richiesta di trasferimento e conservare la copia nei propri archivi. I cittadini in possesso del certificato in essere, o del sostitutivo del titolo abilitativo, dovranno recarsi nell'ufficio e presentare la pratica con copia dei documenti. Si potr\xe0 quindi ritirare l'atto compilato ed essere abilitati all'uso della nuova regolazione. Se la richiesta non \xe8 espressa per opere di ricostruzione, gli uffici provvederanno al trasferimento alla nuova amministrazione. Le varie competenze dei Comuni (ad esempio, le pavimentazioni, che hanno ottenuto una nuova regolazione della Regione) dovranno essere portate in una seconda fase, per i successivi trasferimenti. Si tratta, in tutti i casi, di interventi di adeguamento alla Legge Madia (230/2010), che non ha ancora recepito la normativa regionale del 2014, sulla riordino degli uffici comunali. Con le nuove modalit\xe0, non ci saranno pi\xf9 sezioni di competenza territoriale, ma saranno dei centralini di ambito regionale. In Regione ci saranno anche uffici &quot;complementari&quot;, dove sar\xe0 possibile fare le pratiche per il reddito di cittadinanza, che dovranno poi essere trasferite all'ufficio dell'Inps.]" time="1.796"><properties><property name="score" value="0.994960755109787" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99496076&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99496076
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Anyone ever had a knot in their throat for 10 days straight? Like, they could feel it and it hurt? That\u2019s me right now, and I think it\u2019s really starting to interfere with my relationship with food.\n\nContinue reading \u201cThe Last 10 Days\u201d]" time="0.172"><properties><property name="score" value="0.9188169836997986" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Efren Castillo on Boxing\n\nWith the recent announcement of Canelo Alvarez vs. Gennady \u201cGGG\u201d Golovkin and Vasyl Lomachenko vs. Nicholas Walters for May 2, the sport of boxing is picking up some good steam. On Saturday, Danny Garcia and Robert Guerrero will be fighting in the co-main event of the Alvarez-Kovalev undercard, and they are both in the top five of my pound-for-pound list. The winner of this fight will be the mandatory challenger to Keith Thurman for the WBC welterweight title.\n\nThis fight is just one example of how competitive boxing has been in recent years. There are so many great fighters that we could have one of the best years in boxing in recent memory. But before we go any further, let\u2019s take a look at the four fights that were announced this past week.\n\nFor Alvarez-Kovalev, it was the best fight that could have been made at this time in boxing. In the beginning of 2015, the year started off great with Miguel Cotto vs. Canelo Alvarez, which was supposed to be one of the best fights of the year. But instead, it was a non-event and was a flop. Both fighters looked bad, and the fans were disappointed. That being said, the promoters realized that Cotto vs. Alvarez II would not do great numbers, so a rematch was out of the question.\n\nWith Cotto leaving Top Rank, there was a gaping hole in the middleweight division. Alvarez wanted to stay at 154 pounds, but the only option for him was to go up to 160 to fight the middleweight champion Gennady Golovkin. But with Golovkin only having one fight in 2016, and the third bout against Martin Murray in March, the timing could not have been any better for this fight.\n\nWhile the fight is taking place in Las Vegas, it should not hurt the Mexican fighter. The best fights involving Mexican fighters in Las Vegas have always been ones where the opponent is big in name and popularity. Alvarez has already faced Floyd Mayweather, Erislandy Lara, Alfredo Angulo, James Kirkland and Miguel Cotto. He has had two pay-per-view fights in Las Vegas, and he has never been hurt by the size of the city or venue.\n\nAlvarez is favored to win this fight, but that could all change if he loses to Golovkin. For those that don\u2019t know, boxing fans can be ruthless, and one loss can change everything. But in the end, the fans and the boxing industry will be the biggest winners in this fight.\n\nThen we move on to Lomachenko vs. Walters. These two fighters have two of the most awkward and beautiful styles in boxing. With Lomachenko being at 122, and Walters being at 130, this is a fight that could not be missed. Walters will be giving up eight pounds to the Ukrainian, but that is not a huge factor in this fight. Lomachenko is only about 1.5 inches shorter than Walters, and he is naturally taller.\n\nThe big difference between these two fighters is the lack of power by Walters. He has knocked out 34 out of his 35 opponents, but his most recent fight was against IBF super featherweight champion Miguel Berchelt. Walters could not KO Berchelt, and he lost a 12-round decision. Walters does not have a huge knockout percentage, and Lomachenko does not get knocked out. So it will be interesting to see how this fight plays out.\n\nThe fight that I am looking forward to the most is Garcia vs. Guerrero. These two fighters are very competitive in their division, and they are both in their prime. Garcia has already had two losses this year, but they were to Shawn Porter and Keith Thurman. Those two fights were competitive and competitive in style. But Garcia showed heart and talent in both of those fights.\n\nGuerrero is coming off of a very exciting win over Andre Berto in November. He was hurt several times in the fight, but he pulled out a split decision. Garcia is coming off of a draw with Porter in March, and he is currently ranked No. 3 by the WBC. While Guerrero is ranked No. 4, he has the ability to beat Garcia. Both fighters are fairly similar in style, and if Garcia is smart, he will win this fight.\n\nThe last fight is a big one: Nicholas Walters vs. Vasyl Lomachenko. I am hoping that this fight does not happen until 2018. The reason is because Lomachenko has a fight with Guillermo Rigondeaux on Dec. 9, and Walters is fighting in March. I want to see both fighters in top condition in this fight. With a possible date of April 28, 2018, it will give the fighters plenty of time to prepare.\n\nI believe that this fight could possibly happen before the end of the year. With Walters fighting on March 4 against Jason Sosa, and Lomachenko fighting on Dec. 9, it could set up this fight for the end of the year. Both fighters are on HBO, and the boxing industry will be happy if this fight takes place. Walters has already fought Lomachenko once in 2013, and Lomachenko is undefeated since that fight.\n\nThe only fight that I am looking forward to more than this one is GGG vs. Alvarez. But this is the second most anticipated fight of the year for me. Walters has proven that he can fight great fighters, but he has not faced anyone that could compare to Lomachenko. Walters was supposed to fight Miguel Marriaga on Nov. 26, but it was cancelled due to his weight.\n\nFor Lomachenko, he has had two tough fights against Guillermo Rigondeaux and Jason Sosa, and they have both been challenging fights for him. Lomachenko has fought great fighters such as Nicholas Walters, Gary Russell Jr., and Orlando Salido]" time="3.157"><properties><property name="score" value="0.9973254799842834" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Self Portrait was a virtual representation of one of Ratonhnhak\xe9:ton's genetic memories, relived by Desmond Miles in 2012 through the Animus.\n\nContents show]\n\nDescription Edit\n\nRatonhnhak\xe9:ton was getting dressed when he heard the sound of barking. Worried that the guards were alerted to his presence, he exited the room, only to find a domesticated wolfdog.\n\nDialogue Edit\n\nRatonhnhak\xe9:ton: (frightened) Who's there?\n\nThe wolfdog growled.\n\nRatonhnhak\xe9:ton: Shush!\n\nRatonhnhak\xe9:ton: (calming) Hush!\n\nRatonhnhak\xe9:ton: Good boy.\n\nRatonhnhak\xe9:ton: Now, where'd you come from?\n\nRatonhnhak\xe9:ton: What a curious creature.\n\nRatonhnhak\xe9:ton: You are one of the villagers' pets, no?\n\nRatonhnhak\xe9:ton: The closest village is east of here.\n\nRatonhnhak\xe9:ton: And they keep you locked up, like an object?\n\nRatonhnhak\xe9:ton: That is not right.\n\nRatonhnhak\xe9:ton: I will take you where you can run free.\n\nThe wolfdog barked and followed Ratonhnhak\xe9:ton.\n\nRatonhnhak\xe9:ton: There, now.\n\nRatonhnhak\xe9:ton: Out of here, boy.\n\nRatonhnhak\xe9:ton: We need to be quick.\n\nRatonhnhak\xe9:ton: Come on!\n\nRatonhnhak\xe9:ton: I have a plan, but I cannot do this alone.\n\nRatonhnhak\xe9:ton: You will help me, won't you?\n\nRatonhnhak\xe9:ton: I have a job for you.\n\nRatonhnhak\xe9:ton: Good boy.\n\nRatonhnhak\xe9:ton: I am going to distract the guards.\n\nRatonhnhak\xe9:ton: You must go out the back door, and fetch help.\n\nRatonhnhak\xe9:ton: There are warriors who are eager to prove themselves.\n\nRatonhnhak\xe9:ton: They are in the forest to the south.\n\nRatonhnhak\xe9:ton: Go now!\n\nRatonhnhak\xe9:ton: I'll distract them!\n\nRatonhnhak\xe9:ton: There's a way out this way!\n\nRatonhnhak\xe9:ton: What?\n\nThe wolfdog growled and rushed past Ratonhnhak\xe9:ton.\n\nRatonhnhak\xe9:ton: Oh, well.\n\nRatonhnhak\xe9:ton: Guess I'm not alone after all.\n\nRatonhnhak\xe9:ton: But we are two, and they are many!\n\nRatonhnhak\xe9:ton: If you'll give us some cover, we can...\n\nRatonhnhak\xe9:ton: (sighing) It was worth a try.\n\nRatonhnhak\xe9:ton: I'll hold them off, but you need to fetch help.\n\nRatonhnhak\xe9:ton: You remember the warriors in the forest, yes?\n\nRatonhnhak\xe9:ton: Go!\n\nRatonhnhak\xe9:ton: No, you're not leaving me here!\n\nRatonhnhak\xe9:ton: You can come with me, or fetch help, but you are not leaving me here.\n\nRatonhnhak\xe9:ton: I have an escape route.\n\nRatonhnhak\xe9:ton: Follow me!\n\nRatonhnhak\xe9:ton: Good dog!\n\nRatonhnhak\xe9:ton: The steps are there, come!\n\nRatonhnhak\xe9:ton: Come on!\n\nRatonhnhak\xe9:ton: What are you waiting for?\n\nRatonhnhak\xe9:ton: Good dog.\n\nRatonhnhak\xe9:ton: A woman!\n\nRatonhnhak\xe9:ton: How did she get in here?\n\nRatonhnhak\xe9:ton: Who is she?\n\nRatonhnhak\xe9:ton: Can she help us escape?\n\nRatonhnhak\xe9:ton: (startled) Is that you?\n\nRatonhnhak\xe9:ton: Are you a wolf?\n\nRatonhnhak\xe9:ton: What are you doing here?\n\nRatonhnhak\xe9:ton: If you want to leave with me, we need to move!\n\nRatonhnhak\xe9:ton: We need to be quick!\n\nRatonhnhak\xe9:ton: (giggling) You'll protect me, yes?\n\nRatonhnhak\xe9:ton: Yes, yes.\n\nRatonhnhak\xe9:ton: Oh, it's so dark!\n\nRatonhnhak\xe9:ton: There must be a way out!\n\nRatonhnhak\xe9:ton: There are, but I cannot see where they lead!\n\nRatonhnhak\xe9:ton: You could sniff them out.\n\nRatonhnhak\xe9:ton: Good boy!\n\nRatonhnhak\xe9:ton: Oh, no!\n\nRatonhnhak\xe9:ton: He's blocking our escape!\n\nRatonhnhak\xe9:ton: There has to be a way around!\n\nRatonhnhak\xe9:ton: Perhaps if we avoid him, we can escape!\n\nRatonhnhak\xe9:ton: No!\n\nRatonhnhak\xe9:ton: Wait!\n\nRatonhnhak\xe9:ton: (sighing) They are everywhere!\n\nRatonhnhak\xe9:ton: I don't know if we can get out of here.\n\nRatonhnhak\xe9:ton: (startled) There are more of them!\n\nRatonhnhak\xe9:ton: (sighing) I do not think I can defeat them alone.\n\nRatonhnhak\xe9:ton: You have the smell of the forest on you.\n\nRatonhnhak\xe9:ton: You must lead the others to us!\n\nRatonhnhak\xe9:ton: Find the warriors and bring them to us!\n\nRatonhnhak\xe9:ton: I will do everything in my power to keep them distracted.\n\nRatonhnhak\xe9:ton: You know the way.\n\nRatonhnhak\xe9:ton: Good dog!\n\nRatonhnhak\xe9:ton: (yelling) Come on, we're waiting!\n\nRatonhnhak\xe9:ton: Where are the warriors?\n\nRatonhnhak\xe9:ton: (growling) What is that sound?\n\nRatonhnhak\xe9:ton: It is coming from above!\n\nRatonhnhak\xe9:ton: Is this the help you've brought?\n\nRatonhnhak\xe9:ton: (growling) No, it cannot be!\n\nRatonhnhak\xe9:ton: Good dog!\n\nRatonhnhak\xe9:ton: (yelling) That way!\n\nRatonhnhak\xe9:ton: I think the warriors have arrived!\n\nRatonhnhak\xe9:ton: (yelling) We need your help!\n\nRatonhnhak\xe9:ton: Come on!\n\nRatonhnhak\xe9:ton: No, I said that way!\n\nRatonhnhak\xe9:ton: (yelling) Where are you going?\n\nRatonhnhak\xe9:ton: Come back!\n\nRatonhnhak\xe9:ton: There is no need to be so rude!\n\nRatonhnhak\xe9:ton: Who are you?\n\nRatonhnhak\xe9:ton: (startled) A bear?\n\nRatonhnhak\xe9:ton: I had a wolf for a companion.\n\nRatonhnhak\xe9:ton: You can see him outside.\n\nRatonhnhak\xe9:ton: (surprised) Another dog?\n\nRatonhnhak\xe9:ton: That one is not as friendly as yours.\n\nRatonhnhak\xe9:ton: What are you doing here?\n\nRatonhnhak\xe9:ton: How can you be here?\n\nRatonhnhak\xe9:ton: They are ready to attack, you must distract them!\n\nRatonhnhak\xe9:ton: Good dog!\n\nRatonhnhak\xe9:ton: He can help me, I just need to get him outside!\n\nRatonhnhak\xe9:ton: I cannot leave without the other!\n\nRatonhnhak\xe9:ton: Do not be a fool!\n\nRatonhnhak\xe9:ton: He's too heavy!\n\nRatonhnhak\xe9:ton: You have to help me.\n\nRatonhnhak\xe9:ton: (yelling) Leave him and go!\n\nRatonhnhak\xe9:ton: It is too late!\n\nRatonhnhak\xe9:ton: We are doomed!\n\nRatonhnhak\xe9:ton: They're moving too fast!\n\nRatonhnhak\xe9:ton: How can you be so calm?\n\nRatonhnhak\xe9:ton: They will kill us!]" time="5.024"><properties><property name="score" value="0.7137717306613922" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[I remember the first time I saw Dazed and Confused. I didn't know what to expect, but it blew me away. As I was preparing to sit down with Richard Linklater to discuss his new film, I had that same feeling. How do you get an audience to empathize with characters they would normally write off as deplorable, if not outright despicable? We talked about this and more, along with the film's complex themes, which Linklater and his actors spent years developing. This interview has been edited and condensed.\n\nHugh: So what I'm really curious about, it's a little bit about tone. The last scene is going to be one of the most talked about, but there are lots of moments in the film where I don't think it's going to be easy to laugh. We're going to laugh, but it's going to be a complicated laugh. I want to know how you feel about that scene, and the importance of it being there, and what else is going to be the most talked about scene for people?\n\nRichard: You know what, I don't have a clue. I had a couple of moments. You know, I don't think it's anything like the rape scene. You know, that was one that everybody seemed to feel like the first time they saw it, at least in the audience screenings, there was a palpable kind of, &quot;Oh my god.&quot; But the bowling alley thing, we went through a couple of different iterations of it, and I thought the kids were just going to keep on throwing stuff at him, and he'd get up and leave, but no, that was the one that made people in the audience just go, &quot;I can't.&quot; I mean, I remember, I just think of, I have to confess that I haven't seen the movie in a while, but I remember some of the testing, because the movie tested a lot. And they always give you these cards, like &quot;what scene did you love, what scene did you hate,&quot; or whatever, and I was shocked by some of the things that people found totally inconsequential. You know, the whole detention room scene was almost universally praised. They loved that, they thought that was great, and the end of it, not so much.\n\nHugh: Yeah. The thing that's different about this film, obviously, is it's not only looking at people and their humanness. It's also looking at their flaws. I mean, you could write these guys off as total scumbags, but it's also very clear that they're young, and they're just trying to find their way. And that's what makes this so powerful.\n\nRichard: That was very intentional, and the fact that I had kids in the film, that was a choice that I made very consciously, because I wanted to get the perspective of the younger people who are looking at the world for the first time. And, you know, they don't have the same values that we do. They're different, and they don't realize it, but they are. They haven't been tested yet. And they're also kind of looking at, you know, maybe the judgmental generation that they have to become. So it was very conscious.\n\nHugh: So you're interested in their stories. What about the parents? I mean, you're going to be a parent soon, how has that changed the way you look at this film and the stories?\n\nRichard: I think it's made me even more interested in getting the stories right of the younger people, and being as honest as I can with them. I mean, I've thought about, I want to do a prequel, you know, like set in the '70s. I've thought about doing one that goes into the '90s, and then just getting closer and closer and closer to what it's like today, but really starting in the '70s and the '80s.\n\nHugh: It's a very complicated thing, and I know you've talked about this a lot, I mean, this idea of nostalgia. But I feel like there's a particular version of nostalgia that goes with this film. You know, you look at the sort of deep Americana in the film. I mean, how do you feel about that?\n\nRichard: Well, I guess I don't really know what you mean by that.\n\nHugh: It's this kind of very rich Americana that really looks at what life was like before people moved to the suburbs.\n\nRichard: Well, that was intentional, that was sort of the style of the film, but there were a lot of choices in the film, you know, that were very deliberate, and one was to make it as realistic as possible. You know, there was no comedy, there was no satire. I mean, it was really the intention, and one of the ideas that we were talking about was, how can you really make people identify with these people, if they're like, I mean, just people would say, &quot;Oh, they're just total scumbags,&quot; and that's not the intention. You know, it's like, how can you make people feel like this is a reality, and it's not a joke? And I guess that was one of the choices, you know, like that was just something that was a choice that we were very intentional about.\n\nHugh: What about you? How much do you want this film to be like the experience you remember, and how much do you want to add your own personal story into it?\n\nRichard: You know, I guess, I can remember the moment, it's hard to put into words, but it was, it was a really, it was a weird thing, I was in high school, and we were supposed to be staying at this place, and I was going to walk there. And I was, it was very late, it was like, after 10:00 at night, and I started walking, and I started walking, and I started walking, and I started walking, and I was like, you know, what the hell, you know? And I started walking, and I started walking, and I ended up in my bed, but I just had this sense of loneliness and alienation that I can't describe, but it was a very, it's kind of hard to put into words.\n\nHugh: Yeah.\n\nRichard: So I think that's one of the things that was in the film, that's in all my films, but I think that was something that maybe people don't talk about. But it's, you know, the great teenage angst, that's like a rite of passage, you know, and I guess that was one of the ideas. I think that's what you mean by kind of a different nostalgia.\n\nHugh: Yeah, but I also wonder if it's almost more... It's like an explanation of the parents, of why they are who they are. You know, because I feel like, I mean, you're an independent filmmaker. You're not making a studio movie, but a lot of these people are making that decision because it's a career, but it's also that they're really starting to question their parents.\n\nRichard: That's a good point. Yeah, I mean, I guess that's probably true. But I'm not sure that the same characters would make the same choices. You know, it's a little bit different.\n\nHugh: That's true.\n\nRichard: So I mean, I guess there's probably a different way of looking at it, but I'm not sure.\n\nHugh: Yeah.\n\nRichard: But that's probably true.\n\nHugh: One of the other things that I think is so powerful about the movie, is you can't really tell that they're all the same age.\n\nRichard: Right.\n\nHugh: You know, I think a lot of the time, especially in independent movies, you really can't, especially with characters who are supposed to be that age. But you really do a great job of putting that in there.\n\nRichard: You know, that was a real choice. The only choice that we did was when the girls are taking their bikini tops off, we did make them a little bit older than the guys, because the guys, they were mostly 18, and the girls, you know, they were 21, and they wanted to look that way, you know.\n\nHugh: Yeah.\n\nRichard: I didn't want to deal with, you know, the issue of kids playing these kinds of roles, you know.\n\nHugh: Yeah.\n\nRichard: So, you know, that was one of the choices that we made, but other than that, you know, that was a choice that we made.\n\nHugh: You know, I can't help but think of the movies that inspired this. You know, I'm wondering if there are any that you feel like really influenced you, or you thought about as you were working on this.\n\nRichard: There were lots of movies, but I can't think of any specifically. But, you know, it was really one of the, like I said, it was really one of the influences was, you know, when I was a kid, and I watched all those movies, and I thought, &quot;Wow, I want to make movies that have this feel.&quot; And that's what I wanted to do, and when I say, &quot;this feel,&quot; I mean like, you know, I guess it's sort of the feel of early '70s films, but the cool thing about those films, they have their own feel, you know. I mean, they were different, and I]" time="5.012"><properties><property name="score" value="0.92497761381997" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Dead Space is a popular sci-fi horror game franchise from Visceral Games and Electronic Arts. In addition to two sequels and several handheld spin-offs, the series also has an animated movie called Dead Space: Aftermath which was released in 2011. The movie is set between the events of Dead Space and Dead Space 2.\n\nThis week, we're looking at the movie's trailer and comparing it to its source material. I've played through the game and watched the movie and will be judging the trailer based on what was included in the movie. I'll only be judging the first 30 seconds of the trailer to get an idea of what to expect from the whole movie.\n\nThe trailer starts with a shot of the U.S.G Ishimura, the ship from Dead Space. In the game, the ship is overrun by the necromorphs and infested with scary creatures. The Ishimura's distress beacon is flashing, indicating it's under attack. However, the ship looks too clean to be infested. It's not enough to see the ship itself, you would have to see the dead bodies and monsters crawling around to get the idea.\n\nNext, we see a number of humans fleeing from something off-screen. In the game, they would be running from the necromorphs that have overrun the Ishimura. Unfortunately, they're all human beings in a space suit. You'd have to be pretty stupid to mistake that for the necromorphs. Plus, there's no one attacking the humans. In the game, there would be an endless amount of monsters chasing after the players. There wouldn't be a calm look to everything.\n\nThe next shot shows an exterior shot of the Ishimura while the camera rotates around it. Again, it's too clean to be infested.\n\nNext, we see a shot of some characters, all looking shocked. Again, there's no reason for this if it's not an infestation. It's a bit like seeing a normal-looking family in a horror movie. No one is scared. In fact, there's no real reason for them to be on the Ishimura, which is supposed to be an extremely dangerous place.\n\nThe next shot shows the captain of the Ishimura, Gibb, running up a long ramp. In the game, the Ishimura is more like a battle station than a ship, so the captain would be in the control room. We then see someone telling Gibb to &quot;get everyone out of here.&quot; This person has to be an important character from the game or a secondary character. But I'm guessing the person is Gibb's secretary. Again, there's no one yelling or crying. In fact, it seems like the only reason for them to be running around is because it's time to leave.\n\nThe last shot shows the Ishimura firing on something. It looks like they're firing at something, but it's a small target in the distance. The characters aren't running around like it's an emergency. It doesn't look like they're fighting a desperate battle. In fact, the trailer is too clean. All of the shots look like they were taken in a clean studio, not the remains of a space station.]" time="1.570"><properties><property name="score" value="0.9997942646344503" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[It is possible that you may not want your users to have the ability to make certain changes to the data that is being tracked. For example, you may want to track a list of servers, but you don't want your users to be able to add servers, remove servers, or make any other changes to the server list.\n\nThe answer to this problem is to use triggers. You can create triggers that run when a particular type of change is made to the data, and it is possible to make these triggers prevent the data from being changed.\n\nTo create a trigger, you can use the command-line function sqlt. For example, you could write a trigger that prevents a user from removing a particular field from the database.\n\nHere is an example of how you might create a trigger that prevents a user from removing the server_name field from the list of servers:\n\n[mars@luna src]$ sql --user=alena --password=huh --default-access=READ --tables=ListOfServers,ListOfIPAddresses --echo CREATE TABLE ListOfServers(name TEXT); CREATE TABLE ListOfIPAddresses(ip_address TEXT); CREATE TRIGGER server_name_deletion_not_allowed ON ListOfServers FOR DELETE AS BEGIN RAISE INFO 'You may not delete the field server_name'; END; SELECT * FROM ListOfServers; name ---------- nameserver1 nameserver2 name ---------- nameserver1 name ---------- nameserver2\n\nThe SQL statement you would use to drop the trigger would look like this:\n\n[mars@luna src]$ sql --user=alena --password=huh --default-access=READ --tables=ListOfServers,ListOfIPAddresses --echo DROP TRIGGER server_name_deletion_not_allowed ON ListOfServers; DROP TRIGGER server_name_deletion_not_allowed ON ListOfServers;\n\nNotice that we used the same SQL statement to create and drop the trigger. This is a powerful feature of SQL, which allows you to control exactly when a particular change will occur.\n\nFor example, if you want to remove the server_name field from the database, but you want to retain the data that has already been entered, you could write the trigger as follows:\n\nCREATE TRIGGER server_name_deletion_not_allowed ON ListOfServers FOR DELETE AS BEGIN RAISE INFO 'You may not delete the field server_name'; IF (@old.server_name &lt;&gt; @new.server_name) BEGIN INSERT INTO ListOfIPAddresses VALUES (@old.server_name, NULL); UPDATE ListOfServers SET server_name = NULL WHERE server_name = @old.server_name; END; END;\n\nYou would then delete the field, as follows:\n\n[mars@luna src]$ sql --user=alena --password=huh --default-access=READ --tables=ListOfServers,ListOfIPAddresses --echo DELETE server_name FROM ListOfServers; DELETE server_name FROM ListOfServers;\n\nThis example uses the triggers to prevent data from being deleted, but you can use triggers to make other changes as well. You can use the following table to help you determine which commands should be used to insert and delete information:\n\nSQL Command Insert Insert into table Insert into table columns Update Update table Update table columns Delete Delete table Delete table columns]" time="1.771"><properties><property name="score" value="0.9947806596755981" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[http://www.stangtv.com/shopping/listing.aspx?itemid=5806&amp;catid=1314\n\n\n\nMy brother just bought this last night and said it is very nice. My brother just bought this last night and said it is very nice.]" time="0.203"><properties><property name="score" value="0.9643375277519226" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Perfect! Thank you very much for thess beautiful items!! fast and great quality Top seller, nice one, very fast delivery only took 2 days too come from Germany too the UK, The record was in mint condition!,,,, I Would Recommend Highly quick delivery, record in perfect condition - couldn't be better. thank you very good. Thanks.Great deal. Thanks a lot ! All good, recommended seller !!! Great item, great seller! Superb all round. Professional packing, conservative grading, fast. Thoroughly recommended! Great. thanks! top Very fast shipment, highly recommended, AAAAA+++++++++++++++ top seller, thanks alles super!!! immer wieder gern!!! Perfect !!! very fast Shipping !! Friendly Contact !! A+++++++++ , Perfekt !! Super schneller Versand !! Freundlicher Kontakt !!! 1++++++++++ perfect! Excellent service, very professional. Very fast, good quality top Thanks! Thank you for your quick resoponse!! Quick delivery.No problems.beste Ware=bestenDank No problems! Perfect - thank you so much! Great service. Thank you. Very nice. Thank you! Brilliant seller, arrived very fast for an international delivery and the condition was perfect. I highly recommend this seller and would happily purchase from them again. Vielen Dank! Great seller. Items as described. Deal with confidence. A+ Alles bestens! Gerne wieder! Great service Thank you arrived Mon 23/3/2015 great record, fast shipping, thanks! excellent record, quick shipment, recommended seller! Absolutely perfect in every way: Mint (M) Excellent! Thanks a lot! Good international seller. Great packaging! THANKS!! Very Good, thanks!!!! perfect thanks Awesome disc!!! Perfect transaction!!! PERFECT!!! record arrived teh other day, much better condition as stated, TOP!!! Perfect!!! Thank You!!! Everything perfect! Thank you Awesome!! Great packaging. Very happy!! Thank you for your quick resoponse!!]" time="0.911"><properties><property name="score" value="0.9938312768936157" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99383128&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99383128
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Just when you thought the internet couldn't get any stranger, along comes this clip of a goat using a GoPro camera to take selfies with his friends.\n\nFor this clip, viral video company Barely Political collaborated with the animals at Old MacDonald's Farm Sanctuary in California, who treated the GoPro to an adventure, doing a spot of fishing, swimming and lying on the grass.\n\nIt is unclear exactly what the point of the video is, except to showcase goats that have some level of intelligence.]" time="0.279"><properties><property name="score" value="0.9758872985839844" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[We offer you opportunity to buy an affordable video surveillance system for your home or office. The system is packed with multiple features that allow you to receive secure home security, even if you are at a different location. The system allows you to control your environment from remote locations via a computer or mobile phone. The system can detect sounds and motion that are not normally audible or visible to the human eye. The system includes multiple camera views. If you have animals that might go missing, this is an ideal way to monitor your house.]" time="0.286"><properties><property name="score" value="0.9998194575309753" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Gonzalo \u2018ZeRo\u2019 Barrios, the best Super Smash Bros. for Wii U player in the world, had been competing in Apex for years, and for the most part his reputation remained untarnished by upsets. But this year, things were different. In Apex 2017, he was defeated for the first time since 2015, eliminated in pools by a man that he once conquered: Samuel \u2018Dabuz\u2019 Buzby.\n\nThe question on the minds of many Smash fans was: Why did Dabuz defeat ZeRo, the player who has never been beaten in this game in tournament?\n\nWell, according to Dabuz himself, it was a relatively simple answer.\n\n\u201cI can speak for myself, but one thing that I have definitely noticed is I\u2019ve been playing with a controller for longer,\u201d Dabuz said. \u201cI was playing a lot with the Gamecube controller in 2012 and 2013, and he was mainly playing with the Wii U controller. That\u2019s a big thing, and I think that in general I know how to read his tendencies a little bit better.\u201d\n\nThe controller is a big factor because of the way that Smash for Wii U is played. Because the game was released on the Wii U, it is not played with a controller that has buttons like the PS4 controller, but rather a gamepad with a touchscreen.\n\nZeRo plays with a Gamecube controller with his right hand on the directional pad, and his left hand on the A and B buttons. Because of this, he has to learn the way he plays on that controller and how to switch to it from a standard Wii U gamepad. Dabuz, on the other hand, is more used to the standard controller, which requires more precision and speed to accomplish the same things.\n\nDabuz and ZeRo also play in different regions, which are played on different stages. For example, Dabuz plays on the East Coast, which has the Dream Land stage. This stage is the most neutral stage, and a player who has grown up on that stage would have an edge on a player who was more familiar with the other stages in the game.\n\nZeRo, on the other hand, plays on the West Coast, where the counterpick stages are often Dream Land 64 and Smashville. These stages can change the way a player plays, especially the super defensive style that ZeRo often employs.\n\nThere are also a few other things that can throw a player off their game, including the noises of the crowd and the pressure of being at the top. In a tournament where every person is trying to eliminate you, it can cause even the best to have a momentary lapse in focus and lose a match they should have won.\n\nThese are all things that have affected ZeRo, but they are things that he can get past. Because he is the best, the biggest target and a huge crowd favorite, it is unlikely that the crowd will stop cheering for him any time soon. And when he has a momentary lapse in focus, his playstyle is based around having more than one good option, and his tech skill is such that he can often do something even when he isn\u2019t at his best.\n\nZeRo will be at CEO next week, and this is his chance to reclaim his title as best in the world.\n\nCover photo by Robert Paul via Twitter]" time="1.729"><properties><property name="score" value="0.9997847278912863" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[by Vincent\n\nOnePlus has officially released the Open Beta version of Android 9.0 Pie for the OnePlus 6. OnePlus, being the company they are, have managed to get the Android 9.0 Pie out to their devices at a very fast pace and they are the only company, at the time of writing, who have managed to do so on a flagship device. The Open Beta 3 update is based on Android 9.0 Pie and will get your OnePlus 6 up to date with the rest of the Android devices. OnePlus have also released their Beta application for Android, the app is currently available for Android devices and will help users test out the new Beta updates before they are released.\n\nThis update brings a number of new features to the OnePlus 6, some of which include the new gestures, including a navigation bar gesture, swiping gestures and double-tapping the display. You can now double-tap to wake your device or tap the power button to go to the lock screen. Also in the Open Beta 3 is a new feature known as Gaming Mode 3.0. The new gaming mode has been redesigned and users will be able to choose which apps they want to use the mode with, so for example, if you do not play any games, you can still use the mode with other apps.\n\nOther new features in the Android 9.0 Pie update include adaptive brightness, adaptive battery and Adaptive Brightness in which the brightness will automatically adjust based on the environment. There is also a new design and layout for the entire Android 9.0 Pie update. Some of the features also include the Smart Display feature which is only found on Android Pie devices. The Smart Display feature includes the redesigned Do Not Disturb, new time-to-leave feature, new quick settings and new power management settings.\n\nNow onto the update, the update is currently available to download via the Open Beta website, for more information on how to install the Open Beta updates, check out this guide. Once the update is downloaded, you can follow our guide to install the update via the recovery. The full changelog for the update can be seen below:\n\nSystem Updated system to Android 9.0 Pie\n\nBrand new UI for Android Pie\n\nNew navigation gestures (Available for devices with pop-up camera, Full Screen display)\n\nUpdated Android security patch to 2018.12 General New features for setting app background and text colors\n\nOptimization for Do Not Disturb\n\nOptimization for haptic feedback\n\nOptimization for Volume panel\n\nOptimization for accidental touch\n\nNow able to switch to last app via 3 dots button in the navbar Phone New UI for speed dial and call history screen\n\nWhite UI for dial pad\n\nTransfer support for dual front camera\n\nOptimization for receiver\u2019s screen\n\nNew quick settings design (text color and background color of each tile can be adjusted)\n\nNew features for notification dots (Notification dots are enabled by default. Tap and hold a notification to turn it off)\n\nOnePlus Community app icon is changed Navigation Updated UI for home screen\n\nUpdated Android security patch to 2018.12 Gaming Mode 3.0 Added new feature to adjust display temperature\n\nAdded frame rate benchmark for better game experience. The benchmark runs after exiting a game.\n\nAdded battery level option for battery saver\n\nNew design for power saving mode\n\nOnePlus Launcher Updated icon for Launcher app\n\nUpdated Android security patch to 2018.12\n\nThere are some known issues in the update which OnePlus have said they are working on fixing. The following issues will be fixed in the next update:\n\nCertain third party apps may fail to connect to the internet\n\nWhen using Bluetooth audio devices, the audio may occasionally cut out after a few minutes\n\nAlert slider settings may disappear after the device reboots\n\nAs you can see, OnePlus are still very busy fixing up the Android 9.0 Pie update for the OnePlus 6. You can also follow the Open Beta program, which allows you to see when a new update is available to download and helps you to try out beta updates before they are released.]" time="2.151"><properties><property name="score" value="0.5668750405311584" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Don't let this be you\n\nShutterstock\n\nMake a clean getaway with this simple tips to stay safe on your travels.\n\nShutterstock\n\nMake a clean getaway with this simple tips to stay safe on your travels. (Shutterstock)]" time="0.162"><properties><property name="score" value="0.9816303849220276" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Original Problem\n\nTwo integers x and y are given. The sum of the two is even. What is the largest possible value of x-y?\n\nExample\n\nGiven x = 1 and y = 2 we have x + y = 3. So the answer is x - y = -1.\n\nNote: This problem has appeared at several sites. The only known correct solution is in the book How to Solve It, by Polya. The problem is there posed as one of 1000 math problems, and I\u2019m not sure about the original source of the problem.\n\nPrevious Solution\n\nJavaScript\n\nfunction largest(x, y) { if (x &lt; 0 &amp;&amp; y &lt; 0) { throw new Error('x or y is negative'); } else if (x &lt; y) { x = x - y; } else { y = y - x; } if (x == 0 || y == 0) { return x; } else { return largest(x, y - 1); } } largest(1, 2); // = -1 largest(-1, 1); // = -1 largest(1, -2); // = 1 largest(-1, -2); // = 1 largest(0, 0); // = 0 largest(2, -1); // = 2 largest(1, 1); // = 1\n\nCode\n\n/** * Given two integers x and y, compute the largest possible value of x - y * * @param {number} x The larger of]" time="0.658"><properties><property name="score" value="0.9993372559547424" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99933726&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99933726
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[[In many cases, this book is available for &quot;slightly used&quot; prices at Amazon and the rest of the internet. I also offer this advice: if you find yourself using this book often, buy the paperback, not the hardback.]\n\n\n\nFor a one volume handbook of natural remedies and herbal medicine, I think this book offers quite a bit. The different herbal sections are separated out by area of the body with the body part followed by the particular ailment, i.e., heart disease is followed by the subhead heart disease and some of the suggestions are alfalfa, astragalus, and hawthorne. It also gives you the Latin name of the herb (or tree, or flower, etc) which is followed by a very detailed description of how the particular herb can be used in a variety of ways. You can also find the name in parentheses to the right. In this way, you can find an herb quickly in this volume that you may have heard of, but don't know exactly what it is used for or how to use it.\n\n\n\nI feel like the author offers some very useful information here, but unfortunately the descriptions are so brief that they can be confusing. It's hard to glean all of the information in the little boxes. I've come to rely on it as a resource when I don't have anything else to go on, but I also realize that it is a one-volume reference book and cannot be expected to cover everything. I like to have other books around for that purpose. This book does give a good foundation in herbal medicine, but I don't like using it as a primary resource in itself.]" time="0.858"><properties><property name="score" value="0.9987396597862244" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[L\u2019Aquedotto\n\nL\u2019Aquedotto del vescovo Sisto di Heidenheim\n\nIl primo lavoro che merita di essere considerato \xe8 l\u2019Aquedotto di Heidenheim. Anche se in questo caso l\u2019opera non \xe8 frutto di uno sforzo volontario e non aveva lo scopo di fornire acqua, ma ha come scopo principale la delimitazione del terreno (leggi il resto della notizia).\n\nL\u2019Aquedotto del vescovo Sisto di Heidenheim (XI secolo) \xe8 una delle pi\xf9 importanti e affascinanti opere dell\u2019architettura dell\u2019epoca carolingia.\n\nLa costruzione dell\u2019Aquedotto del vescovo Sisto di Heidenheim, cio\xe8 l\u2019opera a cui si \xe8 qui riferito, avvenne durante il dominio dei vescovi e di cui il famoso Sisto di Heidenheim fu vescovo. Egli viene da una nobile famiglia e si laurea in teologia presso l\u2019universit\xe0 di Parigi, dove acquisisce una grande dottrina.\n\nIn seguito egli torna alla sua citt\xe0 natale per occuparsi dei suoi affari, ma anche per prendersi cura dei suoi parrocchiani. Viene chiamato a prendere le redini del suo vescovado, allorch\xe9 il precedente vescovo, eletto da Carlo Magno, viene accusato di simonia.\n\nIl vescovo Sisto diviene il successore del suo predecessore dopo esser stato eletto nel 1074. Sotto la sua guida il vescovado cresce e prospera notevolmente.\n\nLa dimostrazione pi\xf9 evidente di questo successo \xe8 la costruzione dell\u2019Aquedotto del vescovo Sisto di Heidenheim, la cui costruzione ha inizio nel 1079, quando Sisto \xe8 vescovo, e termina nel 1092, quando cessa le sue funzioni come vescovo.\n\nAncora oggi si pu\xf2 ammirare questo famoso edificio lungo la strada che va da Heidenheim a N\xf6rdlingen. La sua lunghezza \xe8 di una trentina di chilometri, e il suo diametro di cinque metri.\n\nGli spunti di tale opera li ebbe il vescovo Sisto da tre guerre che furono combattute durante i suoi anni di vescovado, e che hanno luogo, la prima, nell\u2019anno 1077 contro i Tedeschi, la seconda, contro i Bavari nel 1085, e la terza, ancora contro i Bavari, ma questa volta per la conquista della citt\xe0 di Bisanz nel 1092.\n\nIn conseguenza di questi attacchi, il vescovo Sisto fu costretto a rispettare un\u2019astuta tattica militare e ad impiegarla nel campo della politica. Egli fu sempre sul chi vive per quanto riguarda la difesa dei confini del suo territorio.\n\nPer ottenere questo scopo, egli, appoggiato dal vescovo di Bamberga, costru\xec il pi\xf9 grande Aquedotto del suo tempo.\n\nSotto la guida del suo vescovo, il vescovo di Heidenheim, il vescovo di Bamberga ed il vescovo di W\xfcrzburg, si \xe8 provveduto ad avanzare verso i confini nordoccidentali, e i confini dei territori confinanti al nord e ad ovest.\n\nI lavori furono attuati nel campo militare, cos\xec come pure nell\u2019ambito delle arti e delle scienze, perch\xe9 il vescovo Sisto cre\xf2 nel 1078 un\u2019Accademia, a Heidenheim, in cui si studiavano le matematiche e l\u2019astronomia.\n\nIl vescovo Sisto riusc\xec cos\xec a mettere fine alla sua missione e a trasformare Heidenheim in un centro per la conservazione della cultura.\n\nLa costruzione dell\u2019Aquedotto del vescovo Sisto, avvenne attraverso un processo molto interessante. Anche se il vescovo Sisto non costru\xec l\u2019Aquedotto, ma invece lo compr\xf2 e lo espropri\xf2, egli, comunque, ne diresse la costruzione e con questo oper\xec una grande opera.\n\nI lavori dell\u2019Aquedotto del vescovo Sisto avvennero in due diverse fasi. La prima delle quali si ebbe tra l\u2019anno 1079 e l\u2019anno 1084, la seconda, tra l\u2019anno 1085 e il 1092.\n\nI primi lavori iniziarono quando Sisto compr\xf2 e prese in esproprio una quantit\xe0 considerevole di terreno, tra il suo vescovado, e il territorio che allora apparteneva a W\xfcrzburg. In conseguenza di questo acquisto, l\u2019Aquedotto pass\xf2 inoltre da W\xfcrzburg a Heidenheim.\n\nLe trattative che precedettero l\u2019acquisto dell\u2019Aquedotto non furono particolarmente facili, e Sisto si trov\xf2 nelle condizioni di dover impegnare molto denaro e cercare di ottenere il necessario appoggio anche nel campo militare.\n\nSisto, comunque, ebbe la capacit\xe0 di perseguire il suo scopo, e di ottenere il suo Aquedotto, pur non potendo completare l\u2019intero progetto, di cui egli, tuttavia, lavor\xf2 ad un livello rilevante.\n\nL\u2019Aquedotto di Heidenheim \xe8 un monumento per la cultura, e una vera e propria meraviglia, cos\xec come lo fu, quindi, per il suo tempo. Oggi l\u2019Aquedotto \xe8 ancora l\u2019opera principale di un moderno vescovo, che di continuo, invece, esalta la sua coscienza della cultura di tale Aquedotto.]" time="4.424"><properties><property name="score" value="0.9988245725631714" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99882457&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99882457
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Level Poll Past Versions\n\nLevel 624 is the 14th level in Rattle Dunes. To pass this level, you must wake up 8 fireflies and score at least 500 points in 24 moves or fewer.\n\nContents show]\n\nDifficulty Edit\n\nThis level is intense.\n\nStars Edit\n\nPoints 500 5,000 10,000 Difficulty Easy Hard Intense\n\nStrategy Edit\n\n1st board: Clear the blocks normally.\n\n2nd board: Clear the blocks normally. Use a color burst to exit.\n\n3rd board: Clear the blocks normally. Use a color burst to exit.\n\n4th board: Clear the blocks normally. Use a color burst to exit.\n\nTrivia Edit\n\nWalkthrough Edit]" time="0.336"><properties><property name="score" value="0.9229748249053955" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.92297482&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.92297482
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Chrysanthemum is a non-flowering perennial plant belonging to the Asteraceae family. The asteraceae family is composed of around 24,000 species. The genus Chrysanthemum is a composite group and consists of approximately 200 species. In Hindi, the plant is known as Karanji, Kuntimuk, Kwatamal. It is a good source of vitamin C and is used as a food and a medicinal plant. The leaves are considered to be of some value as a pot-herb. The entire plant is highly toxic. This plant is often confused with one belonging to the Rosaceae family, known as Golden Shower (Cassia fistula).\n\nChrysanthemum is widely used in Ayurveda, Siddha, Unani, Homoeopathy and allopathy. The plant is used as an appetizer, digestive, vermifuge, for treating jaundice, anthelmintic, skin infections, piles, and disorders of uterus and vagina. The plant is reported to have antiviral activity.\n\nThe extract of Chrysanthemum spp. is used in form of drops, ointments, nasal sprays, aerosols, gargles, and nasal drops in Western medicine. It is a non-opiate analgesic. It is an antipyretic, vasodilator, hypotensive, and anti-inflammatory agent. It is used for treating dental infections, painful cold sores, and sinusitis.\n\nIt is a powerful disinfectant used in hospitals and clinics as a wound antiseptic. It is an effective treatment for infections of the skin, boils, and other skin lesions. The preparation is a thick solution of manganese salts in a 2% solution of hydrogen peroxide and contains 4% to 10% of this chemical, which acts as an antiseptic. It is also used to treat malaria.\n\nChrysanthemum oil is considered as a good antimicrobial agent. It is used in the form of inhalation for the treatment of sinusitis. It is effective against fungi, yeast, bacteria, and viruses. This oil is also used as an antiviral agent for flu, chickenpox, herpes, and herpes zoster.\n\nThe extract is used in the treatment of infectious diarrhoea, hepatitis, and amoebic dysentery. It is also used to reduce irritation and promote healing of the urethra. The extract is used in the treatment of oral thrush and vaginal candidiasis.\n\nThe plant is used for treating leucorrhoea and vaginitis in traditional medicine. It is used for treating burning and pain in the urethra. The plant is used as an antiseptic for the treatment of mouth sores. It is used for treating boils, blisters, insect bites, burns, and wounds. It is also used for treating skin conditions like eczema, ringworm, and hives.\n\nSome of the names that this plant is known by include Chrysanthemum morifolium, Cassia fistula, Chrysanthemum indicum, Chrysanthemum chamissonis, Homonoia chrysanthemi, and Xiong-hua.\n\nChrysanthemum is considered to be of great value for medicinal purposes. The plant is not only used as an appetizer and digestive, but also for treating the disorders of uterus and vagina. The herb is considered as a good antiseptic and as a healing agent. It is useful for healing wounds, ringworm, and jaundice. It is also considered to be good for treating tuberculosis and reducing phlegm.]" time="1.761"><properties><property name="score" value="0.9943230152130127" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[A Jersey City firefighter, one of the department\u2019s most seasoned, has been stripped of his rank and placed on administrative leave as a result of alleged sexual harassment and stalking.\n\nThe city took the action against James Fitzpatrick, a 13-year veteran, last month, according to police and fire officials.\n\nFitzpatrick, who has worked at Engine 6, told The Jersey Journal he is appealing his demotion and administrative leave and does not admit to the accusations.\n\n\u201cThey\u2019re saying this is harassment and stalking, but I don\u2019t know what else to do,\u201d he said. \u201cI just want to talk to her.\u201d\n\nCity spokeswoman Jennifer Morrill said Jersey City law prohibits her from discussing the matter because it is a personnel issue.\n\nThe Journal reported in October that city police had been investigating Fitzpatrick on a complaint that he sent hundreds of harassing and harassing and stalking text messages to a firefighter who works at a different firehouse.\n\nCity officials said at the time that they knew of at least 200 messages. The city launched an investigation and suspended Fitzpatrick with pay in November, and he was demoted on Jan. 8.\n\nThe suspended firefighter and police said they were not aware of any criminal charges against Fitzpatrick.\n\n\u201cWe are investigating it and reviewing it internally,\u201d Morrill said last month. \u201cThere is no further information at this time.\u201d\n\nFitzpatrick, a Marine Corps veteran who has been with the Jersey City Fire Department for 13 years, said he is appealing his demotion and administrative leave.\n\nHe said he received a letter from the city in December saying his leave was the result of a \u201cpersonnel matter,\u201d but he said he has not been able to see the letter\u2019s contents because it is sealed by the city\u2019s law department.\n\n\u201cThey\u2019re saying this is harassment and stalking, but I don\u2019t know what else to do,\u201d he said. \u201cI just want to talk to her.\u201d\n\nFitzpatrick said he met the firefighter while they were both in college, and the two became good friends. Fitzpatrick said he started sending her text messages in September because she had not returned his calls. He said he sent her messages out of concern that she was missing work because she was depressed.\n\nThe messages were more than once a day, he said, and he said he never threatened the woman.\n\nThe woman\u2019s lawyer, Susan Ross, said last month that Fitzpatrick sent \u201cmore than 200 harassing and stalking text messages.\u201d\n\n\u201cThe harassment has included threats to my client's life,\u201d she said.\n\nRoss did not respond to a request for comment this week.\n\nFitzpatrick\u2019s wife, Cheryl Fitzpatrick, said the situation has been difficult for their two young daughters, who know nothing about the accusations.\n\n\u201cMy husband is an outstanding fireman,\u201d she said. \u201cIt has been very difficult on my children.\u201d\n\nTerrence T. McDonald may be reached at tmcdonald@jjournal.com. Follow him on Twitter @terrencemcd. Find The Jersey Journal on Facebook.]" time="1.564"><properties><property name="score" value="0.9998195767402649" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Sen. Joe Donnelly Joseph (Joe) Simon DonnellyDems pressure GOP to take legal action supporting pre-existing conditions Senate Dems build huge cash edge in battlegrounds Fed chief lays out risks of trade war MORE (D-Ind.) is touting his support for President Trump Donald John TrumpCorker: US must determine responsibility in Saudi journalist's death Five takeaways from testy Heller-Rosen debate in Nevada Dem senator calls for US action after 'preposterous' Saudi explanation MORE\u2019s first Supreme Court nominee, Judge Neil Gorsuch, to make the case to skeptical progressives in Indiana that he can still win their votes in a tough reelection race.\n\nDonnelly has made his vote for Gorsuch, who was confirmed in April, a central part of his reelection pitch.\n\n\u201cMy decision to support Judge Neil Gorsuch was based on an overwhelming need to fill the Supreme Court vacancy,\u201d Donnelly wrote in an op-ed in The Indy Star this week.\n\nADVERTISEMENT\n\n\u201cI strongly believe Judge Gorsuch is a qualified, mainstream jurist who will base his decisions on his understanding of the law and is well-respected among his peers. Judge Gorsuch\u2019s extensive experience and respect in the legal community is important to the integrity of the Supreme Court,\u201d he added.\n\nDonnelly, who is facing a reelection battle against Republican businessman Mike Braun in November, is trying to win over progressives and not alienate them in a state that went for Trump in 2016.\n\nA poll from last month showed that 49 percent of Hoosiers say that they would prefer to vote for someone other than Donnelly, compared to 42 percent who say they will vote for the incumbent senator.\n\nDonnelly told The Hill last month that he is focusing on making his case to voters and he doesn\u2019t pay attention to polls.\n\nBut while Donnelly touts his vote for Gorsuch, he\u2019s been careful to show he won\u2019t be someone who will blindly vote for the president\u2019s agenda.\n\n\u201cI won\u2019t support a nominee who doesn\u2019t respect the rights of Ho]" time="1.037"><properties><property name="score" value="0.9097586870193481" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Highland League Founded: 1975 Country: Scotland Level: 5 Number of teams: 17 most based in: Scotland (4) Region(s): Scotland Sponsor(s): TBD Website: TBD\n\nThe Highland Football League is a senior non-league association football competition in the north of Scotland.\n\nContents show]\n\nPremier League Edit\n\nThe league consists of 17 clubs, playing each other home and away.\n\nSeason Premier Division Relegated from 1. 2016\u201317 (A) (A) Oban Camanachd\n\n(B) Kyles Athletic 2. 2015\u201316 (A) (A) (A) Lochaber\n\n(B) Skye Camanachd 3. 2014\u201315 (A) (A) Lochaber\n\n(B) Kyles Athletic 4. 2013\u201314 (A) (A) Lochaber\n\n(B) Strathspey Camanachd 5. 2012\u201313 (A) (A) (A) Skye Camanachd\n\n(B) Lochaber 6. 2011\u201312 (A) (A) (A) Skye Camanachd\n\n(B) Kyles Athletic 7. 2010\u201311 (A) (A) (A) Lochaber\n\n(B) Kyles Athletic 8. 2009\u201310 (A) (A) (A) Skye Camanachd\n\n(B) Strathspey Camanachd 9. 2008\u201309 (A) (A) (A) Skye Camanachd\n\n(B) Strathspey Camanachd 10. 2007\u201308 (A) (A) (A) Skye Camanachd\n\n(B) Lochaber 11. 2006\u201307 (A) (A) (A) Kyles Athletic\n\n(B) Strathspey Camanachd 12. 2005\u201306 (A) (A) (A) Lochaber\n\n(B) Skye Camanachd 13. 2004\u201305 (A) (A) (A) Skye Camanachd\n\n(B) Kyles Athletic 14. 2003\u201304 (A) (A) (A) Skye Camanachd\n\n(B) Lochaber 15. 2002\u201303 (A) (A) (A) Strathspey Camanachd\n\n(B) Lochaber 16. 2001\u201302 (A) (A) (A) Skye Camanachd\n\n(B) Kyles Athletic 17. 2000\u201301 (A) (A) (A) Lochaber\n\n(B) Kyles Athletic 18. 1999\u201300 (A) (A) (A) Lochaber\n\n(B) Kyles Athletic\n\n1. In the initial year of the league, reserve teams of the Highland Football League clubs were eligible to take part, although the games were not counted as official League matches.\n\n2. Only one team can be promoted to the league, although in the first year only two teams took part, and in the second year four teams took part.\n\n3. Both the first and second placed teams in the reserve league took part.\n\n4. The winners of the league and the runners-up compete in a playoff for the right to play in the Scottish Cup, and if the runners-up win they enter the competition in the first round, the winners in the fourth.\n\nLeague History Edit\n\nLeague History Season Champions Relegated 1. 2008\u201309 Strathspey Camanachd Kyles Athletic 2. 2007\u201308 Lochaber Skye Camanachd 3. 2006\u201307 Lochaber Kyles Athletic 4. 2005\u201306 Skye Camanachd Lochaber 5. 2004\u201305 Lochaber Kyles Athletic 6. 2003\u201304 Skye Camanachd Lochaber 7. 2002\u201303 Strathspey Camanachd Lochaber 8. 2001\u201302 Skye Camanachd Kyles Athletic 9. 2000\u201301 Lochaber Kyles Athletic 10. 1999\u201300 Skye Camanachd Kyles Athletic\n\nTitles by club Edit\n\nAberdour - 1\n\n2000\u201301\n\n\n\nSkye Camanachd - 8\n\n2000\u201301, 2003\u201304, 2005\u201306, 2008\u201309, 2011\u201312, 2013\u201314, 2014\u201315, 2015\u201316, 2017\u201318\n\n\n\nLochaber - 5\n\n1999\u201300, 2003\u201304, 2008\u201309, 2010\u201311, 2011\u201312\n\n\n\nKyles Athletic - 4\n\n1996\u201397, 1997\u201398, 2008\u201309, 2011\u201312\n\n\n\nLochside Rovers - 3\n\n1989\u201390, 1992\u201393, 1998\u201399\n\n\n\nStrathspey Camanachd - 3\n\n1997\u201398, 1998\u201399, 2009\u201310\n\n\n\nBoleskine Rovers - 1\n\n1993\u201394\n\n\n\nInvergordon - 1\n\n1991\u201392\n\n\n\nKinlochshiel - 1\n\n1991\u201392\n\n\n\nFort William - 1\n\n1991\u201392\n\n\n\nArisaig - 1\n\n1990\u201391\n\n\n\nCaberfeidh - 1\n\n1987\u201388\n\n\n\nAberdeen University - 1\n\n1985\u201386\n\n\n\nFormartine United - 1\n\n1982\u201383\n\n\n\nStrathspey Thistle - 1\n\n1979\u201380\n\n\n\nFort William Rovers - 1\n\n1947\u201348\n\n\n\nRoss County - 1\n\n1946\u201347\n\n\n\nFortrose - 1\n\n1945\u201346\n\n\n\nNairn Thistle - 1\n\n1942\u201343\n\n\n\nInverness Thistle - 1\n\n1941\u201342\n\n\n\nStrathspey Thistle - 1\n\n1940\u201341\n\n\n\nFort William Thistle - 1\n\n1939\u201340\n\n\n\nCambridge University - 1\n\n1935\u201336\n\n\n\nDeveronvale - 1\n\n1935\u201336\n\n\n\nCaberfeidh - 1\n\n1934\u201335\n\n\n\nInverness High School FP - 1\n\n1933\u201334\n\n\n\nMurdo Macleod - 1\n\n1933\u201334\n\n\n\nIona - 1\n\n1932\u201333\n\n\n\nFasnakyle - 1\n\n1931\u201332\n\n\n\nCaberfeidh - 1\n\n1930\u201331\n\n\n\nRoyal HSFP - 1\n\n1929\u201330\n\n\n\nRovers - 1\n\n1928\u201329\n\n\n\nAberdour - 1\n\n1927\u201328\n\n\n\nDeveronvale - 1\n\n1926\u201327\n\n\n\nInverness Thistle - 1\n\n1925\u201326\n\n\n\nForres Thistle - 1\n\n1924\u201325\n\n\n\nDeveronvale - 1\n\n1923\u201324\n\n\n\nFortrose - 1\n\n1922\u201323\n\n\n\nInverness Thistle - 1\n\n1921\u201322\n\n\n\nForres Thistle - 1\n\n1918\u201319\n\n\n\nInverness Thistle - 1\n\n1917\u201318\n\n\n\nFortrose - 1\n\n1916\u201317\n\n\n\nInverness Caledonian - 1\n\n1915\u201316\n\n\n\nDeveronvale - 1\n\n1914\u201315\n\n\n\nFortrose - 1\n\n1913\u201314\n\n\n\nDeveronvale - 1\n\n1912\u201313\n\n\n\nRoyal HSFP - 1\n\n1911\u201312\n\n\n\nFortrose - 1\n\n1910\u201311\n\n\n\nInverness Thistle - 1\n\n1909\u201310\n\n\n\nDeveronvale - 1\n\n1908\u201309\n\n\n\nDeveronvale - 1\n\n1907\u201308\n\n\n\nInverness Thistle - 1\n\n1906\u201307\n\n\n\nHighland Regt. - 1\n\n1905\u201306\n\n\n\nRoyal HSFP - 1\n\n1904\u201305\n\n\n\nFortrose - 1\n\n1903\u201304\n\n\n\nHighland Regt. - 1\n\n1902\u201303\n\n\n\nRoyal HSFP - 1\n\n1901\u201302\n\n\n\nDeveronvale - 1\n\n1900\u201301\n\n\n\nDeveronvale - 1\n\n1899\u201300\n\n\n\nHighland Regt. - 1\n\n1898\u201399\n\n\n\nInverness Caledonian - 1\n\n1897\u201398\n\n\n\nDeveronvale - 1\n\n1896\u201397\n\n\n\nDeveronvale - 1\n\n1895\u201396\n\n\n\nFortrose - 1\n\n1894\u201395\n\n\n\nDeveronvale - 1\n\n1893\u201394\n\n\n\nHighland Regt. - 1\n\n1892\u201393\n\n\n\nHighland Regt. - 1\n\n1891\u201392\n\n\n\nFortrose - 1\n\n1890\u201391\n\n\n\nDeveronvale - 1\n\n1889\u201390\n\n\n\nDeveronvale - 1\n\n1888\u201389\n\n\n\nInverness Thistle - 1\n\n1887\u201388\n\n\n\nDeveronvale - 1\n\n1886\u201387\n\n\n\nFortrose - 1\n\n1885\u201386\n\n\n\nFortrose - 1\n\n1884\u201385\n\n\n\n]" time="4.877"><properties><property name="score" value="0.16702309846878052" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[STOCKHOLM \u2014 In a Scandinavian twist to the final race of the series, American David Hargitt was awarded the North American Pro Kart Challenge championship over Canadian Fredrik Ekblom on Saturday night at Karlskoga Arena in Sweden.\n\nHargitt had an on-track confrontation with Ekblom earlier this season at Spa-Francorchamps, Belgium, during which Hargitt, now 28, ran into Ekblom, a now 18-year-old Canadian, with his kart, sending both into the barriers.\n\nBoth drivers were knocked out of the race.\n\nBut Hargitt ended the six-race season as the champion, his third overall series title and first since 2004. Hargitt and Ekblom finished with 107 and 96 points, respectively.\n\nHargitt finished first and second in the last two races to hold onto his lead. Ekblom won the first race of the year and finished third in the final race.\n\nDavid Hargitt, in a Pirelli factory-backed kart, and Ekblom, a Tony Kart factory driver, finished second and third behind series champion Thomas Preining.]" time="0.669"><properties><property name="score" value="0.9997918009757996" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.9997918&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.9997918
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[A man died Sunday in China\u2019s western Xinjiang region after setting himself on fire outside a police station, a regional government spokesman said.\n\nThe man set himself ablaze after fighting with local police over a traffic fine, Nur Bekri, head of the regional government, told a press conference.\n\nPolice in the northwestern region were carrying out \u201cstability maintenance work\u201d at the time of the incident, he said.\n\n\u201cHe self-immolated after a quarrel over a traffic violation and the police station,\u201d Bekri said, adding that the police handled the case correctly.\n\n\u201cThe government urges people to resolve their disputes through legal means,\u201d he said.\n\nIn recent months, dozens of people have set themselves on fire in Xinjiang to protest what they say are human rights abuses, while scores more have resorted to violence or armed assaults against government officials or police stations.\n\nWhile official figures show that most of those who have set themselves on fire are Uyghur, a Turkic-speaking, largely Muslim ethnic group, a majority of those who carried out attacks were Han Chinese, according to reports.\n\nAn estimated 200 people died as a result of the self-immolations or violence last year.\n\n(AFP)\n\n\u7f51\u53cb\u8bc4\u8bba\uff1a\n\n\u7f51\u6613\u8fbd\u5b81\u7701\u961c\u65b0\u5e02\u624b\u673a\u7f51\u53cb\uff1a\n\nThis is nothing. In the most difficult time of the Cultural Revolution, more than a thousand people self-immolated themselves to death, over five million were violently killed, and an entire country was sacrificed for one man. The self-immolators are committing suicide for what? All of this is being sacrificed for a man.\n\n\u7f51\u6613\u6cb3\u5317\u7701\u5eca\u574a\u5e02\u624b\u673a\u7f51\u53cb\uff1a\n\nSending military troops into Urumqi, letting people starve, shoot, slaughter, etc., the government has long since abandoned any kind of \u201cmaintenance of stability.\u201d But can they face the fact that a few hundred Uyghurs have gone to extremes? This is the embarrassment of the government, a proof that the lives of the Uyghur are worth even less than the government thinks. I\u2019m just hoping that the Uyghur people will gain their own freedom one day.\n\n\u7f51\u6613\u5e7f\u4e1c\u7701\u4f5b\u5c71\u5e02\u624b\u673a\u7f51\u53cb\uff1a\n\nTerrorism and arson.\n\nOnly a thief is afraid of fire.\n\nThe history of the burning of books in the East is the same as that of the burning of witches in the West.\n\nThe responsibility of the Uyghurs is in the hands of the government and the people of China.\n\nIf the people of China can be unified, if the government can be clear, it will not be as it is now, nor will it have the effects of spreading instability throughout the world.\n\n\u7f51\u6613\u7f8e\u56fd\u624b\u673a\u7f51\u53cb\uff1a\n\nThe most important thing in the Uyghurs\u2019 lives is the love for their children and themselves, as the life of one\u2019s family is the foundation of all things. This is how the Uyghurs are, and this is the reason why they fight back, and this is why they do what they do.\n\nFor example, the government forcibly inserts women into the Uyghur\u2019s households and forces them to wear head scarves, it forcibly strips the children from their families and rears them in \u201corphanages,\u201d forcibly rips young girls from their parents\u2019 arms to be sent to Tibet, forcibly controls the number of children Uyghurs are allowed to have, forcibly forces Uyghurs to go to Han villages, forcibly encourages them to go to work in the cities, forcibly puts Han businesses into Uyghur homes, forcibly buys out their livestock and agricultural fields to build industrial parks, forcibly forces Uyghur young people to go to school to study and the rest are to go to do factory work, forcibly builds supermarkets, shops, hotels, etc., which gives them no chance to get employment, forcibly makes it so that Uyghur young people can\u2019t find a spouse to get married and have children, forcibly demolishes their houses and forces them to move to another area, forcibly blocks them from having religious activities, etc., etc., the Uyghurs are continuously being forced to endure.\n\nIf you look at the Uyghur protests from a perspective of someone observing them, you can understand how it all began and why it has continued to this day.\n\nThe government has no choice but to take all this, because the Uyghurs are a minority group with nothing to back them up and their lot is really not the greatest.\n\nOn the other hand, if the government wants to go easy on the Uyghurs, this might also lead to the Uyghurs feeling discontent with the government, so it has to be a careful balancing act.\n\nYou know, because of this constant bickering, the government can no longer help the Uyghurs, because it\u2019s already using all its energy to help the Han, which is what it cares about most.\n\n\u7f51\u6613\u6e56\u5317\u7701\u6b66\u6c49\u5e02\u624b\u673a\u7f51\u53cb\uff1a\n\nIf the man really did self-immolate because of the fines, isn\u2019t that a little too extreme? If he had used this method before, it could be understood. But he\u2019s done this only now, this means that he\u2019s only using this method for his own ends. What does this mean? How can people kill themselves and turn themselves into such an extreme protest against injustice? Who are the ones to blame?\n\n\u7f51\u6613\u5317\u4eac\u5e02\u624b\u673a\u7f51\u53cb\uff1a\n\nI am just trying to understand, he set himself on fire to fight over a traffic violation?\n\n\u7f51\u6613\u7f8e\u56fd\u624b\u673a\u7f51\u53cb\uff1a\n\nWhen the time comes, there will only be Uyghur resistance and nothing else.\n\n\u7f51\u6613\u6e56\u5357\u7701\u624b\u673a\u7f51\u53cb\uff1a\n\nJust look at this, everyone. The Uyghurs have their own methods of protest and methods of expressing their wishes and opinions, which have nothing to do with us.\n\nSo can we stop always picking on them and accusing them of terrorist actions? What they are doing is simply that they are using extreme methods to express their unhappiness and discontent, but I hope that they will consider the extreme actions of their people and act in accordance with the situation in China.\n\nI would like to ask everyone, what do you think?\n\nI believe the true thinking of the masses should be understood.\n\n\u7f51\u6613\u9655\u897f\u7701\u624b\u673a\u7f51\u53cb\uff1a\n\nWhat a shame, they have this kind of personality and they\u2019re so different from us. But since they are living under our government, and their lives are being constantly destroyed, I understand the psychology of people like this.]" time="4.075"><properties><property name="score" value="0.41171595454216003" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Text &amp; images Copyright (C) 2002 Jakub Maly. Any reproduction or distribution outside of the FamiLAB is prohibited.\n\nBackground\n\nRecently I built a decent MIDI controller that I was using with Reaktor 5 to control some of my modular synth presets, and later used it for recording. Since my computer is very powerful, I wanted to see how far I could go with the processing power, and if I could make it work with something faster than Reaktor.\n\nWhat's a VST?\n\nVirtual Studio Technology (VST) is a plugin system designed by Steinberg for use in their DAW Cubase. Since it's free to use, it's been picked up and used in a variety of other DAWs, the most popular being FL Studio, but also in other apps such as Sony Sound Forge. The current version is VST 2.x, and the most important feature of the system is that it makes a virtual]" time="0.453"><properties><property name="score" value="0.9997969269752502" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Relive the nostalgic days of yesteryear with this piece of Americana. This set of 10 postcards contains vintage snapshots of travel, sightseeing, and sports around Chicago.\n\n- Set of 10 postcards\n\n- Each postcard measures 4.25&quot; x 6.25&quot;\n\n- Postcards are blank on the back\n\n- Packaged in a tube]" time="0.236"><properties><property name="score" value="0.9980502128601074" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[La tenue de T.I.M.E. a toujours \xe9t\xe9 respect\xe9e. Malgr\xe9 cela, le bruit de la machine, ajout\xe9 aux bruits de la ville et de l\u2019arm\xe9e, peut perturber et la m\xe9ditation est devenue difficile. La vision d\u2019un grand arbre \xe0 c\xf4t\xe9 de la machine \xe9tait tr\xe8s attendue.\n\nEnsuite, nous avons b\xe9n\xe9fici\xe9 d\u2019un super beau soleil.\n\nNous avons d\xfb travailler un peu pour rassurer les villageois qui ont appris \xe0 avoir peur du robot. Nous avons d\xfb expliquer que nous ne pouvions pas d\xe9sactiver la machine tous les soirs. Mais, les enfants ont \xe9t\xe9 ravis de nous montrer leur dessin.\n\nEt voil\xe0 l\u2019arbre en vrai, la terre le recouvre doucement et en plus il fait bon.]" time="0.544"><properties><property name="score" value="0.9975637197494507" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99756372&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99756372
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[We the people of India are collectively called as Hindus.\n\nThe Quran says that we are better than other nations.\n\nIn the Quran it is written as\n\n\u201cThe most honorable of you in the sight of God is the most pious of you.\u201d (49:13)\n\nIt also says that Allah is closer to the believers than their jugular vein. (Quran 50:15)\n\nHindus consider the cow to be sacred. The Quran teaches us that it is not what enters the stomach that is impure but what comes out of the mouth.\n\nIn the Quran it is written as\n\n\u201cGod does not hold you responsible for your unintentional oaths. And He forgives you for any oaths you take to intentionally falsify something. (Quran-66:2)\n\nIt says that we should have absolute faith in God.\n\nIt says that the scholars are our friends and neighbors, that is why they must be obeyed and respected. (Quran- 4:71)\n\nOur God is one.\n\nIt is written in the Quran as\n\n\u201cThose who believe, and those who are Jewish, and the Christians, and the Sabians \u2013 whoever believes in God and the Last Day and does righteous deeds \u2013 will have their reward with their Lord. (Quran 2:62)\n\nIt is also written in the Quran as\n\n\u201cIndeed, the religion in the sight of God is Islam. (Quran 3:19)\n\nA Hindu must strive to become a better person and have patience, and trust in God.\n\nThe QURAN says as\n\n\u201cAnd it is He who sent down to you the Book. In it are verses that are entirely clear, they are the foundation of the Book; and others not entirely clear. As for those in whose hearts there is a deviation (from the truth) they follow that which is not entirely clear thereof, seeking discord and seeking an interpretation (suitable to them). And no one knows its [true] interpretation except God. But those firm in knowledge say, \u201cWe believe in it. All [of it] is from our Lord.\u201d And no one will be reminded except those of understanding. (Quran 3:7)\n\nWe must follow what is in the Quran and what was said by the Prophet (PBUH).\n\nWe can learn from all great men.\n\nIt is written in the Quran as\n\n\u201cTo you we sent the scripture in truth, confirming the scripture that came before it, and guarding it in safety: so judge between them by what God has revealed, and do not follow their low desires [to turn away] from the truth that has come to you. For each of you We have appointed a law and a way. If God had so willed, He would have made you a single community, but [He intended] to test you in what He has given you; so strive as in a race in all virtues. The goal of you all is to God; it is He that will show you the truth of the matters in which you dispute.\u201d (Quran 5:48)\n\nThe Quran says as\n\n\u201cThis day, I have perfected your religion for you, completed My favor upon you, and have chosen for you Islam as your religion.\u201d (Quran 5:3)\n\nHindus must believe in God and His Prophets.\n\nThe Quran says as\n\n\u201cThe similitude of Jesus before God is as that of Adam; He created him from dust, then said to him: \u201cBe\u201d. And he was. The truth is from your Lord, so be not of those who doubt.\u201d (Quran 3:59)\n\nHindus must believe in the Quran and what it says.\n\nThe Quran says as\n\n\u201cAnd We have sent down to you the book explaining all things, a guide, mercy, and glad tidings for those who have submitted.\u201d (Quran 16:89)\n\nThe Quran says as\n\n\u201cIf only they had stood fast by the Law, the Gospel, and all the revelation that was sent to them from their Lord, they would have enjoyed happiness from every side. There is from among them a party on the right course; but many of them follow a course that is evil.\u201d (Quran 5:66)\n\nHindus should follow their great men like Krishna and Mahavir.\n\nThe Quran says as\n\n\u201cHe has ordained for you of religion what He enjoined upon Noah and that which We have revealed to you, [O Muhammad], and what We enjoined upon Abraham and Moses and Jesus \u2013 to establish the religion and not be divided therein. Difficult for those who associate others with God is that to which you invite them. God chooses for Himself whom He wills and guides to Himself whoever turns back [to Him].\u201d (Quran 42:13)\n\nThe Quran says as\n\n\u201cYou [true believers in Islamic Monotheism, and real followers of Prophet Muhammad and his Sunnah (legal ways, etc.)] are the best nation produced [as an example] for mankind. You enjoin what is right and forbid what is wrong and believe in God.\u201d (Quran 3:110)\n\nMuslims are to call people to follow the path of the right.\n\nThe Quran says as\n\n\u201cWe have sent you forth as a witness, as a bringer of good tidings and a warner, as well as to give mercy.\u201d (Quran 34:28)\n\nMuslims must follow the truth.\n\nIt is written in the Quran as\n\n\u201cYou shall be on the right course, as long as you are firm and mindful of God.\u201d (Quran 2:110)\n\nThe Quran says as\n\n\u201cYou [true believers in Islamic Monotheism, and real followers of Prophet Muhammad and his Sunnah (legal ways, etc.)] are the best of peoples ever raised up for mankind; you enjoin what is right and forbid what is wrong and believe in God.\u201d (Quran 3:110)\n\nThe Hindus must follow the truth and must not take the path of falsehood.\n\nThe Quran says as\n\n\u201cYou will certainly be tested in your possessions and in yourselves. And you will certainly hear from those who were given the Scripture before you and from those who associate others with God much abuse. But if you are patient and fear God \u2013 indeed, that is of the matters [worthy] of determination.\u201d (Quran 3:186)\n\nThe Hindu should seek the guidance of God.\n\nThe Quran says as\n\n\u201cIs it not to God that sincere devotion is due? But those who take for protectors other than God (say): \u201cWe only worship them that they may bring us nearer to God.\u201d Truly God will judge between them in that wherein they differ. But God guides not such as are false and ungrateful.\u201d (Quran 16:72)\n\nThe Hindu should follow the teachings of the Quran.\n\nIt is written in the Quran as\n\n\u201cLet there be no compulsion in religion: Truth stands out clear from Error: whoever rejects evil and believes in God hath grasped the most trustworthy hand-hold, that never breaks. And God heareth and knoweth all things.\u201d (Quran 2:256)\n\nHindus should also keep away from alcohol and gambling.\n\nIt is written in the Quran as\n\n\u201cO you who believe! Intoxicants (all kinds of alcoholic drinks), and gambling, and Al-Ansab, and Al-Azlam (arrows for seeking luck or decision) are an abomination of Shaitan\u2019s (Satan) handiwork. So avoid (strictly all) that (abomination) in order that you may be successful.\u201d (Quran 5:90)\n\nHindus must avoid sex outside marriage.\n\nThe Quran says as\n\n\u201cAnd do not go]" time="4.183"><properties><property name="score" value="0.4390040210315159" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Format: Mass Market Paperback\n\nI am a big fan of Julie Garwood. I have read and enjoyed all of her books. Most of her works have been historical romances. She is the queen of the historical romance. In her &quot;Garwoodesque&quot; style, she takes on a contemporary setting, but her attention to detail and superb research remains. If you've read her works in the past, you will find this new one to be in keeping with her other works.\n\n\n\nThe story takes place in modern day Scotland. Lucinda Ashton's family has had it rough for the past two years. Her father has been suffering from an incurable illness, her mother's nerves are beginning to show, and her twin brothers are both leaving for college and working part time jobs to help with the family expenses. But, it is her brother Duncan who she is most worried about. Duncan has always been her hero. The oldest and most handsome of the Ashton boys, she has watched him for years make the right choices and resist temptation. Lucinda wants to be a lawyer. But her parents are barely able to feed and cloth their five children, so Lucinda has had to abandon her dreams of higher education. Lucinda is determined to help her parents and her siblings. She will sell her beautiful pearl necklace so she can pay for her brothers' educations and alleviate some of her parents' financial burden.\n\n\n\nDuncan is thrilled to be going to university in Edinburgh. He is determined to study hard and make something of himself. When he discovers that his sister has sold her necklace, he is furious. But, Lucinda has already made the deal and there is nothing Duncan can do to stop it. Lucinda is devastated by Duncan's reaction to her choice. She is so hurt by his harshness that she stops speaking to him. Duncan is ashamed of his anger, and he is sorry for the pain he has caused his sister. His mother, a loving and kind woman, tells him to go to Lucinda and make things right.]" time="1.000"><properties><property name="score" value="0.9998039603233337" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[di Mara Polli\n\n\xc8 sceso sottotraccia, il Professore, perch\xe9 ne aveva piene le scatole. E perch\xe9 sostenere che la lezione di merda di quel giorno era solo la punta di un iceberg, un ritaglio di una realt\xe0 molto pi\xf9 vasta. In quel preciso momento, aveva preso coraggio, lo stesso coraggio che aveva avuto vent\u2019anni prima, quando s\u2019era inventato lo sciopero a sorpresa. Al posto del tran tran, era passato agli scioperi. Il primo, nel \u201977, aveva avuto un riscontro difficile: un solo docente aveva appoggiato la sua iniziativa, ma la classe era nel complesso solidale con l\u2019insegnante (sbagliato). Il secondo, invece, aveva attirato l\u2019attenzione delle maestre. Quelle della prima e della seconda. E s\xec che aveva ragione: quello, per esempio, era il momento pi\xf9 duro, non solo per lui, ma anche per tutte le donne: l\u20198 marzo, avevano dovuto copiare i fogli per fare i mazzi di fiori (d\u2019autunno, nel caso dei professori). Copiare i fogli, e lasciare l\u2019edificio in ginocchio, a sguazzare nella scienza delle macchine. Ma il professore non sapeva come fosse l\u20198 marzo per le maestre. La maestra che la giornata l\u2019aveva passata con la matita in bocca, che faceva finta di avere i minatori tra i denti. I minatori di turno. Quelli della prima e della seconda.\n\nLui, il professore, era sempre in bilico, ma in un modo o nell\u2019altro ce l\u2019aveva fatta, e questo gli bastava. Non avrebbe mai detto che tutte le donne svolgono un lavoro, alcune stanno al cesso, altre a cucire, altre stanno in sala operatoria, altre ancora passano a ritirare il pane o a farsi servire a tavola. Di tutte queste attivit\xe0, lui avrebbe fatto a meno. Di tutte, e del lavoro in generale. Ma a volte \xe8 dura essere supereroi. Anche se solo per un giorno.\n\nAnnunci]" time="1.591"><properties><property name="score" value="0.9990901947021484" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99909019&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99909019
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Would you like to know what to expect on your vacation in Hotel Kristina? Then have a look at the reviews of Hotel Kristina in Kranjska Gora, written by other travelers.\n\nscores a 8.0 out of 10 , based on 1 reviews]" time="0.204"><properties><property name="score" value="0.9950673580169678" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99506736&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99506736
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The murder of the investigative journalist and two-time Academy Award winning filmmaker who was critical of Russian president Vladimir Putin, points to the fact that there is no independent journalism in Russia, Prof. Mark Galeotti, Senior Researcher at the Institute of International Relations Prague (Czech Republic) and an expert on Russian security services, has told Sputnik.\n\n\xa9 AP Photo / Evgeny Feldman Oligarchs Pushed Away From the Media, Facing Pressure From the Russian Authorities\n\nThe state-run Russian Investigative Committee has opened a criminal investigation into the murder of Boris Nemtsov , a Russian opposition politician, on charges of illegal arms trade.\n\nOn Sunday, February 28, Nemtsov was shot dead while walking across a bridge near the Kremlin in Moscow, on his way home.\n\n\u201cPutin has silenced the few remaining independent newspapers and there is no independent television. If it weren\u2019t for the internet, people would not have any information about what is going on,\u201d Mark Galeotti, Senior Researcher at the Institute of International Relations Prague (Czech Republic) and an expert on Russian security services, told Sputnik.\n\nGaleotti pointed out that as Russia\u2019s opposition leaders are being arrested or killed, the country\u2019s media are totally \u201cunder the control of the Russian authorities.\u201d\n\n\u201cIf you are critical of Putin, you are likely to disappear or be killed. There is no independent journalism,\u201d he stressed.\n\nAt the same time, the academic stressed that \u201cmurder\u201d should not be used in this case, as it is actually \u201ca pre-meditated killing.\u201d\n\n\u201cIt is a message to anybody else who might want to be critical, that there is no point in being outspoken,\u201d Galeotti noted.\n\nSpeaking about the possible motives for Nemtsov\u2019s murder, the academic suggested that there might have been a \u201cspat\u201d between the victim and the Russian authorities over the situation in Ukraine.\n\n\u201cI can imagine that there was some sort of disagreement,\u201d Galeotti said. \u201cNemtsov was an avowed liberal, who was pro-Western. He said that Putin\u2019s policies in Ukraine were wrong.\u201d\n\nThe Russian academic noted that Nemtsov could have been \u201cbrought into the public eye by some of the authorities who are opposed to Putin\u2019s policies.\u201d\n\n\u201cI am not saying that Putin ordered the killing, but if you look at who is opposed to him, there are different people. Nemtsov was a figure of importance in the 1990s and in the early 2000s. So if you are trying to build up a coalition, you would bring Nemtsov back,\u201d Galeotti noted.\n\nThe killing of Boris Nemtsov is \u201chighly significant,\u201d Galeotti said, adding that \u201cnot everybody is killed.\u201d\n\n\u201cThe message is that there is no point in being outspoken. It is an effective silencer of anyone else who might want to be critical of the regime,\u201d the academic concluded.]" time="1.549"><properties><property name="score" value="0.9745194911956787" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Product Description\n\nThis packet contains a variety of great math stations for your students to practice addition and subtraction skills. The packet contains 28 stations and two activity pages for the students to work on independently.]" time="0.162"><properties><property name="score" value="0.9768099188804626" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.97680992&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.97680992
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[I used some of the great pears that I had in the refrigerator for this hearty dish. It was an outstanding choice, in my opinion. However, I am certain it would be just as good using apples, peaches or apricots.\n\n* There are some glazes that might cause a reaction if you are allergic to nuts.]" time="0.203"><properties><property name="score" value="0.952393651008606" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The Radio Tuner is designed to make it easy to listen to different radio stations using standard FM/AM radio receivers.\n\n\n\nThe app includes many different radio stations from different countries around the world.\n\n\n\nThe Radio Tuner is free and without advertising.\n\n\n\nRecent changes:\n\n- Bugfix in app list\n\nThe Radio Tuner is designed to make it easy to listen to different radio stations using standard FM/AM radio receivers.\n\n\n\nThe app includes many different radio stations from different countries around the world.\n\n\n\nThe Radio Tuner is free and without advertising.\n\n\n]" time="0.277"><properties><property name="score" value="0.9996191263198853" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The legacy of past security problems with GM vehicles has forced GM to recall more than 6 million vehicles over the past decade, according to the National Highway Traffic Safety Administration.\n\nGM made most of the recent recalls, including those for ignition switch and power steering problems.\n\nThe automaker will appear before Congress this week to answer questions about its handling of the ignition switch problem. It faces lawsuits from owners of vehicles that were part of the recall, and also faces a criminal investigation.\n\nRep. Diana DeGette (D-Colo.) released a list of recalls from the NHTSA. DeGette has proposed legislation that would require the NHTSA to assess how automakers handled vehicle defects and how their actions impacted safety.\n\nThe data below is taken directly from the NHTSA. The list does not include recalls from the agency's other administrative offices, including the Offices of Defects Investigation.\n\nAlso, the NHTSA does not keep data for recalls issued in the 1950s.\n\n2013: 2.3 million\n\n2012: 1.3 million\n\n2011: 1.3 million\n\n2010: 1.1 million\n\n2009: 710,000\n\n2008: 780,000\n\n2007: 650,000\n\n2006: 480,000\n\n2005: 360,000\n\n2004: 290,000\n\n2003: 290,000\n\n2002: 270,000\n\n2001: 300,000\n\n2000: 310,000\n\n1999: 260,000\n\n1998: 230,000\n\n1997: 260,000\n\n1996: 250,000\n\n1995: 230,000\n\n1994: 240,000\n\n1993: 200,000\n\n1992: 210,000\n\n1991: 220,000\n\n1990: 190,000\n\n1989: 190,000\n\n1988: 200,000\n\n1987: 180,000\n\n1986: 150,000\n\n1985: 140,000\n\n1984: 160,000\n\n1983: 140,000\n\n1982: 120,000\n\n1981: 100,000\n\n1980: 90,000\n\n1979: 90,000\n\n1978: 80,000\n\n1977: 70,000\n\n1976: 60,000\n\n1975: 60,000\n\n1974: 50,000\n\n1973: 50,000\n\n1972: 40,000\n\n1971: 40,000\n\n1970: 30,000\n\n1969: 20,000\n\n1968: 20,000\n\n1967: 10,000\n\n1966: 0\n\n1965: 0\n\n1964: 0\n\n1963: 0\n\n1962: 0\n\n1961: 0\n\n1960: 0\n\n1959: 0\n\n1958: 0\n\n1957: 0\n\n1956: 0\n\n1955: 0\n\n1954: 0\n\n1953: 0\n\n1952: 0\n\n1951: 0\n\n1950: 0\n\n1949: 0\n\n1948: 0\n\n1947: 0\n\n1946: 0\n\n1945: 0\n\n1944: 0\n\n1943: 0\n\n1942: 0\n\n1941: 0\n\n1940: 0\n\n1939: 0\n\n1938: 0\n\n1937: 0\n\n1936: 0\n\n1935: 0\n\n1934: 0\n\n1933: 0\n\n1932: 0\n\n1931: 0\n\n1930: 0\n\n1929: 0\n\n1928: 0\n\n1927: 0\n\n1926: 0\n\n1925: 0\n\n1924: 0\n\n1923: 0\n\n1922: 0\n\n1921: 0\n\n1920: 0\n\n1919: 0\n\n1918: 0\n\n1917: 0\n\n1916: 0\n\n1915: 0\n\n1914: 0\n\n1913: 0\n\n1912: 0\n\n1911: 0\n\n1910: 0\n\n1909: 0\n\n1908: 0\n\n1907: 0\n\n1906: 0\n\n1905: 0\n\n1904: 0\n\n1903: 0\n\n1902: 0\n\n1901: 0\n\n1900: 0\n\n1899: 0\n\n1898: 0\n\n1897: 0\n\n1896: 0\n\n1895: 0\n\n1894: 0\n\n1893: 0\n\n1892: 0\n\n1891: 0\n\n1890: 0\n\n1889: 0\n\n1888: 0\n\n1887: 0\n\n1886: 0\n\n1885: 0\n\n1884: 0\n\n1883: 0\n\n1882: 0\n\n1881: 0\n\n1880: 0\n\n1879: 0\n\n1878: 0\n\n1877: 0\n\n1876: 0\n\n1875: 0\n\n1874: 0\n\n1873: 0\n\n1872: 0\n\n1871: 0\n\n1870: 0\n\n1869: 0\n\n1868: 0\n\n1867: 0\n\n1866: 0\n\n1865: 0\n\n1864: 0\n\n1863: 0\n\n1862: 0\n\n1861: 0\n\n1860: 0\n\n1859: 0\n\n1858: 0\n\n1857: 0\n\n1856: 0\n\n1855: 0\n\n1854: 0\n\n1853: 0\n\n1852: 0\n\n1851: 0\n\n1850: 0\n\n1849: 0\n\n1848: 0\n\n1847: 0\n\n1846: 0\n\n1845: 0\n\n1844: 0\n\n1843: 0\n\n1842: 0\n\n1841: 0\n\n1840: 0\n\n1839: 0\n\n1838: 0\n\n1837: 0\n\n1836: 0\n\n1835: 0\n\n1834: 0\n\n1833: 0\n\n1832: 0\n\n1831: 0\n\n1830: 0\n\n1829: 0\n\n1828: 0\n\n1827: 0\n\n1826: 0\n\n1825: 0\n\n1824: 0\n\n1823: 0\n\n1822: 0\n\n1821: 0\n\n1820: 0\n\n1819: 0\n\n1818: 0\n\n1817: 0\n\n1816: 0\n\n1815: 0\n\n1814: 0\n\n1813: 0\n\n1812: 0\n\n1811: 0\n\n1810: 0\n\n1809: 0\n\n1808: 0\n\n1807: 0\n\n1806: 0\n\n1805: 0\n\n1804: 0\n\n1803: 0\n\n1802: 0\n\n1801: 0\n\n1800: 0\n\n1799: 0\n\n1798: 0\n\n1797: 0\n\n1796: 0\n\n1795: 0\n\n1794: 0\n\n1793: 0\n\n1792: 0\n\n1791: 0\n\n1790: 0\n\n1789: 0\n\n1788: 0\n\n1787: 0\n\n1786: 0\n\n1785: 0\n\n1784: 0\n\n1783: 0\n\n1782: 0\n\n1781: 0\n\n1780: 0\n\n1779: 0\n\n1778: 0\n\n1777: 0\n\n1776: 0\n\n1775: 0\n\n1774: 0\n\n1773: 0\n\n1772: 0\n\n1771: 0\n\n1770: 0\n\n1769: 0\n\n1768: 0\n\n1767: 0\n\n1766: 0\n\n1765: 0\n\n1764: 0\n\n1763: 0\n\n1762: 0\n\n1761: 0\n\n1760: 0\n\n1759: 0\n\n1758: 0\n\n1757: 0\n\n1756: 0\n\n1755: 0\n\n1754: 0\n\n1753: 0\n\n1752: 0\n\n1751: 0\n\n1750: 0\n\n1749: 0\n\n1748: 0\n\n1747: 0\n\n1746: 0\n\n1745: 0\n\n1744: 0\n\n1743: 0\n\n1742: 0\n\n1741: 0\n\n1740: 0\n\n1739: 0\n\n1738: 0\n\n1737: 0\n\n1736: 0\n\n1735: 0\n\n1734: 0\n\n1733: 0\n\n1732: 0\n\n1731: 0\n\n1730: 0\n\n1729: 0\n\n1728: 0\n\n1727: 0\n\n1726: 0\n\n1725: 0\n\n1724: 0\n\n1723: 0\n\n1722: 0\n\n1721: 0\n\n1720: 0\n\n1719: 0\n\n1718: 0\n\n1717: 0\n\n1716: 0\n\n1715: 0\n\n1714: 0]" time="0.481"><failure message="IndexError: index out of range in self">E   IndexError: index out of range in self</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Hamish Morrison\n\nHamish Morrison (born 8 September 1997) is an Australian professional rugby union player who plays for the in the Super Rugby competition. His position of choice is fullback.\n]" time="0.158"><properties><property name="score" value="0.9777672290802002" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.97776723&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.97776723
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Tom Brown's Schooldays (film)\n\nTom Brown's Schooldays is a 1921 British silent drama film directed by H. B. Parkinson and starring Milton Rosmer, Eva Moore and C. M. Hallard. It is an adaptation of Thomas Hughes' novel &quot;Tom Brown's School Days&quot;.\n\n\n]" time="0.201"><properties><property name="score" value="0.9570002555847168" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.95700026&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.95700026
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Rabi Kaul\n\nRabi Kaul (born 1951) is a social activist, writer, human rights activist, former Director of Amnesty International India, and former Chairperson of Alternative Law Forum, Bangalore, India. He is a graduate of the University of Delhi.\n\nHis human rights work has been a major part of his life since 1981 when he began his work in civil liberties defence and the protection of human rights in general, and of Dalit rights in particular. In the 1980s, he was active in the movement against the Emergency. He was in prison for a few months during that time. He is widely respected as a human rights activist, and one of the most important advocates for social justice in India.\n\nKaul served as the Executive Director of the International Humanist and Ethical Union, and as the International Secretary of the International Humanist and Ethical Union, and is a past Chair of the IHEU Coordinating Committee.\n\nHe has worked with Amnesty International since 1989. During his tenure with Amnesty International, Kaul was Director of the organization's Indian chapter, which he helped to build into one of the largest branches in the world. He served as the Regional Director of Amnesty International for South Asia. In December 2012 he was replaced by Raviya Ismail.\n\nKaul also serves as an adviser to the South Asia Human Rights Documentation Centre. He is a recipient of the UN Human Rights Award, the Yash Bharti Award and the \nPeople's Union for Civil Liberties (PUCL) Award, amongst other recognitions. He is also a recipient of the Gandhi Peace Prize (2011).\n\nKaul has written extensively on human rights issues, social justice, and democracy. He has published a number of books, including &quot;Democracy, Minority Rights and Violence&quot; (2008), &quot;Religion, Violence and Non-Violence&quot; (2005) and &quot;Whither Secularism&quot; (1999).\n\nHis publications include:\n\n]" time="1.026"><properties><property name="score" value="0.9712980389595032" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Erskine Hamilton Childers\n\nErskine Hamilton Childers (; 17 June 1870 \u2013 20 November 1922) was an Irish politician who served as the fourth President of Ireland, from 25 June 1921 to his death on 20 November 1922. He was a member of the Irish Republican Brotherhood and Sinn F\xe9in.\n\nErskine Hamilton Childers was born in England in 1870, the eldest son of Robert Erskine Childers, a British Army officer and later a writer and political theorist.\n\nIn the course of his lifetime he spelled his surname in a number of different ways, including Childers and Childs.\n\nChilders attended school at Stubbington House, Fareham. At the age of 16 he became a clerk in a Southampton shipping office. He joined the Liberal Party and became a political speaker on its behalf. In 1899 he was elected as a councillor to Southampton County Borough Council.\n\nHe married Molly Zichy-Woinarski (sister of Francis &quot;Frank&quot; Zichy-Woinarski) in 1908, but had no children.\n\nChilders was a keen yachtsman, and in 1906, with Sir Thomas Lipton, he set up the Shamrock IV syndicate to challenge for the Americas Cup.\n\nHe participated in the unsuccessful Irish Convention of 1917\u20131918, a body established by British Prime Minister David Lloyd George which negotiated an agreement with Irish representatives in an attempt to end the period of Irish nationalist insurrection. The Convention, which met at Dublin Castle, made a number of suggestions regarding the governance of Ireland, which were partially incorporated into the Government of Ireland Act 1920. It was Childers, who, in September 1917, raised the need for a &quot;conversation&quot; between Irish and British leaders.\n\nAfter the Easter Rising of 1916, and his arrest for involvement, Childers helped reorganise the Irish Republican Brotherhood (IRB) in an effort to steer it towards constitutional activity and away from militancy.\n\nIn the 1918 general election, Childers stood as a Sinn F\xe9in candidate in Wicklow. Although he did not win the seat, he polled well and the Sinn F\xe9in party decided to nominate him as one of its members of the Irish delegation to the Paris Peace Conference. It was at the Paris Peace Conference that he met Arthur Griffith and became a convinced Irish republican.\n\nHe became a committed supporter of Arthur Griffith's Sinn F\xe9in party. In 1919, he was arrested and charged with sedition, because of his support of the Irish Republican Army (IRA).\n\nHe was elected unopposed as a Sinn F\xe9in Teachta D\xe1la (TD) to the 2nd D\xe1il at the 1921 elections for the Dublin Mid constituency. He supported the Anglo-Irish Treaty of 1921, and became a member of the Irish Free State Seanad in 1922.\n\nHe became the fourth President of the Executive Council of the Irish Free State (i.e. prime minister) in June 1921, and on 26 June 1921 he was appointed to the office of President of the Republic, (which was to become that of President of the Executive Council on the enactment of the 1922 Constitution in August 1922).\n\nIn 1921, the Government of Ireland Act 1920 split the island into two autonomous territories within the United Kingdom: the six north-eastern counties became Northern Ireland and the rest of Ireland became Southern Ireland, governed by the Irish Free State government in Dublin. However, the Parliament of Northern Ireland chose to opt out of the Free State in 1922. Under the terms of the Anglo-Irish Treaty, this meant that Northern Ireland would become part of the Irish Free State, but the British refused to accept this, saying it would lead to the &quot;break-up of the United Kingdom&quot;. The Irish Taoiseach and President of the Executive Council (prime minister), W. T. Cosgrave, insisted that the British had breached the Treaty and that the Free State had the right to &quot;withdraw from the British Commonwealth of Nations&quot;, but he chose not to use this option. Instead, he demanded that the Northern Ireland government do so, and thus abolish itself.\n\nChilders was shot dead in his home on Inishmore, Aran Islands, County Galway, in November 1922, two months after becoming President. The following January, the &quot;Irish Independent&quot; reported that Michael Noyk, the Russian Consul in Galway, and Joseph Murray, a member of the Irish Republican Brotherhood, were both implicated in his murder. The paper stated that Murray was sent by James O'Connor to warn Childers to stop political work in the Four Courts or else he would be killed, while Noyk knew of the plot and failed to notify the police. A police officer told the paper that he believed it was an ordinary robbery that &quot;went wrong&quot; as &quot;Childers was too tough a man to be robbed&quot;.\n\n]" time="2.581"><properties><property name="score" value="0.05516829490661621" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05516829&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05516829
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Sarah Williams (social worker)\n\nSarah &quot;Sally&quot; Williams (1880 \u2013 April 23, 1951) was an American social worker. Williams established herself as an activist, fighting for women's suffrage, juvenile delinquency, and public health. Williams was well known for her innovative and groundbreaking approaches to improving the quality of life for children, the public, and the underprivileged. Williams began working for the New York City Health Department in 1918, and in 1925 she became the director of the Bureau of Child Guidance. Williams was the only woman in the country with such a position at the time, and held the position until her death in 1951.\n\nSarah Williams was born in 1880 in Philadelphia, Pennsylvania, the second child of James Williams, a grocer and later a dentist, and Ella Burt Williams. At age 10, she attended the Powelton Avenue School in Philadelphia, and by age 12 she had already decided that she wanted to be a teacher. At age 13, Williams attended the Girls' High School and excelled academically. She took part in school plays and oratorical contests, and upon graduation, Williams attended Swarthmore College in 1896. Williams majored in Greek and Latin and continued to study oratorical contests and plays. After graduating in 1900, Williams continued her education at Cornell University, where she studied American History and Economics. While studying at Cornell, Williams became a member of the college women's debating team, which would later become a hobby for her. Upon graduation, Williams pursued graduate studies at Columbia University, where she received her M.A. degree in Social Science in 1907. At Columbia, Williams received one of her most important honors, the Demorest Fellowship. After graduation, Williams decided to enter the field of social work.\n\nAfter receiving her Master's Degree, Williams began to work as a social worker in Philadelphia. While working in Philadelphia, Williams became interested in the idea of juvenile courts, and worked to improve them. Williams began traveling to study different models of juvenile courts and while in London, she visited the Infant Welfare and the Juvenile Probation Departments. While visiting a British Probation Home, Williams became interested in the training of child welfare workers. She returned to the United States and started a training program for juvenile probation officers in Philadelphia.\n\nIn 1911, Williams accepted a job as the chief social worker of the Infant Welfare Association of Philadelphia. In 1915, Williams moved to New York City and accepted a position as the executive secretary of the Child Labor Committee. Williams continued her work with juvenile delinqu]" time="1.180"><properties><property name="score" value="0.9996837377548218" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Tips for Buying Brooklyn End Tables By Bradington-Young\n\nHow to purchase a Brooklyn End Tables By Bradington-Young\n\nTop Tips To Help You Buy And Care For Your Furnishings It doesn't require much time to study this informative article, nonetheless the rewards are fantastic. You is going to be shocked whenever you realize how these guidelines alter your shopping behaviors. When it comes to furniture, there is little if any difference between affordable and top quality. You will get everything you buy and don't be concerned about inferior quality. If you adored this information and you would certainly such as to receive even more facts concerning best camping tents reviews i implore you to check out our own page. Solid wood furniture is perhaps the most suitable purchase. It lasts lengthier and requires less maintenance than different\n\nWhat you should know before buying end side tables .\n\nSmart Tips For Purchasing Quality Affordable Furniture It 's what holds your meal since you eat. What keeps one's body aloft when you sleep. What contains your best valuables to guard them. It holds your lamp, television and books. Furniture is available to us, within our office, at home and elsewhere. Learn how to be a smarter end side tables shopper by reading the below information. If you've got youngsters, you must avoid purchasing end side tables with sharp edges. Little ones have a tendency to trip and fall often which can create a visit to the emergency room if the end side tables has sharp edges. When buying end side tables for the family room, seek out simple styles. You may well be totally into that black leather couch, but a couch adorned with a bunch of floral designs is not something that you are going to be thankful for. You and your children can be happy with contemporary end side tables. Go shopping for end side tables with your hues as your guide. You might love a certain piece, but when you go it with the room, it may not hold the existing decor. Avoid this from happening. It's easy to by\n\nThis bed was as described, very easy to assemble, and had good instructions. Very nice and firm bed. Assembly went well. This bed was as described, very easy to assemble, and had good instructions. Very nice and firm bed. Assembly went well.\n\nBrooklyn End Tables By Bradington-Young...\n\n\n\nSee Products Descriptions &amp; Reviews!!]" time="1.229"><properties><property name="score" value="0.898092786471049" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.89809279&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.89809279
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[&quot;This is the work of idiots who do not know what they are doing. They have no idea about cars, engines or aerodynamics, and in the end they will pay for it,&quot; team principal Peter Sauber said in a statement.\n\nF1 racing, one of the world's most lucrative sports, has for years been plagued by small teams on tight budgets. The sport's commercial supremo Bernie Ecclestone is trying to improve the situation and has created a fund, estimated to be worth up to $40 million, to support teams in dire straits.\n\n&quot;The idea of taking money from one and giving it to the other is not very positive,&quot; Sauber said. &quot;It creates the possibility of a new form of slavery. I do not think that anyone who worked in the factories during the war thought about whether a tank would end up being driven by a German or an Englishman. He was just happy to be making money. But people are not stupid, and there are some consequences that need to be considered. This is just crazy. It will never work.&quot;\n\nFerrari, BMW-Sauber, McLaren and Renault, four of the sport's top teams, have signed up to the initiative but Ferrari president Luca di Montezemolo said Sauber was missing the point.\n\n&quot;If the results of the championship are being hampered by an imbalanced distribution of resources, then there is a reason for the redistribution,&quot; he said.\n\nSauber's team have won 10 world championships but are among a number of outfits on the verge of going under as the sport's top teams spend millions of dollars in search of an elusive tenth title.]" time="0.768"><properties><property name="score" value="0.7671092450618744" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Paladin\n\nA paladin is the sworn defender of a group of creatures, a good-aligned organization, or a person. Most are deeply religious, but the few who aren't still serve their groups or causes with fanatical zeal. The paladin is essentially a warrior of the gods, and most spend their lives wandering the world of Oerth, aiding those they deem good or just. Although a paladin will fight to slay a horde of monsters or a mad king, he will not do so out of a lust for battle or destruction, but out of a need to keep the forces of darkness at bay.\n\nPaladins are renowned for their bravery, their unyielding dedication to the principles of good, and their mastery of martial arms. It is rare to find an evil paladin.\n\nAbilities: Strength and Charisma are both prime requisites for a paladin. Strength is his prime requisite, followed by Charisma. Strength governs how hard a paladin can hit, how effectively he can wear armor, and his carrying capacity. Charisma governs the strength of his spells and the DC's of his class skills. A paladin must be lawful good.\n\nHit Die: d10\n\nStarting Age: Complex.\n\nStarting Gold: Complex.\n\nClass Skills\n\nThe paladin's class skills are Concentration (Con), Craft (Int), Diplomacy (Cha), Handle Animal (Cha), Heal (Wis), Intimidate (Cha), Knowledge (nobility and royalty) (Int), Knowledge (religion) (Int), Profession (Wis), Ride (Dex), and Sense Motive (Wis).\n\nSkill Points at First Level: (2 + Int modifier) x 4.\n\nSkill Points at Each Additional Level: 2 + Int modifier.\n\nClass Features\n\nAll of the following are class features of the paladin.\n\nWeapon and Armor Proficiency: Paladins are proficient with all simple and martial weapons, with all types of armor (heavy, medium, and light), and with shields (except tower shields).\n\nAura of Good (Ex): The power of a paladin's aura of good (see the detect good spell) is equal to her paladin level.\n\nDetect Evil (Sp): At will, a paladin can use detect evil, as the spell.\n\nSmite Evil (Su): Once per day, a paladin can call out to the powers of good to aid her in her struggle against evil. As a swift action, the paladin chooses one target within sight to smite. If this target is evil, the paladin adds her Charisma bonus (if any) to her attack rolls and adds her paladin level to all damage rolls made against the target of her smite. If the target of smite evil is an outsider with the evil subtype, an evil-aligned dragon, or an undead creature, the bonus to damage on the first successful attack increases to 2 points of damage per level the paladin possesses. Regardless of the target, smite evil attacks automatically bypass any DR the creature might possess.\n\nIn addition, while smite evil is in effect, the paladin gains a deflection bonus equal to her Charisma modifier (if any) to her AC against attacks made by the target of the smite. If the paladin targets a creature that is not evil, the smite is wasted with no effect.\n\nThe smite evil effect remains until the target of the smite is dead or the next time the paladin rests and regains her uses of this ability. At 4th level, and at every three levels thereafter, the paladin may smite evil one additional time per day, as indicated on Table: Paladin, to a maximum of seven times per day at 19th level.\n\nDivine Grace (Su): At 2nd level, a paladin gains a bonus equal to her Charisma bonus (if any) on all saving throws.\n\nLay on Hands (Su): Beginning at 2nd level, a paladin can heal wounds (her own or those of others) by touch. Each day she can use this ability a number of times equal to 1/2 her paladin level plus her Charisma modifier. With one use of this ability, a paladin can heal 1d6 hit points of damage for every two paladin levels she possesses. Using this ability is a standard action, unless the paladin targets herself, in which case it is a swift action. Despite the name of this ability, a paladin only needs one free hand to use this ability.\n\nAlternatively, a paladin can use this healing power to deal damage to undead creatures, dealing 1d6 points of damage for every two levels the paladin possesses. Using lay on hands in this way requires a successful melee touch attack and doesn't provoke an attack of opportunity. Undead do not receive a saving throw against this damage.\n\nAura of Courage (Su): Beginning at 3rd level, a paladin is immune to fear (magical or otherwise). Each ally within 10 feet of her gains a +4 morale bonus on saving throws against fear effects. This ability functions while the paladin is conscious, but not if she is unconscious or dead.\n\nDivine Health (Ex): At 3rd level, a paladin gains immunity to all diseases, including supernatural and magical diseases.\n\nDivine Bond (Sp): Upon reaching 5th level, a paladin forms a bond with her god. This bond can take one of two forms. Once the form is chosen, it cannot be changed.\n\nThe first type of bond allows the paladin to enhance her weapon as a standard action by calling upon the aid of a celestial spirit for 1 minute per paladin level. When called, the spirit causes the weapon to shed light as a torch. At 5th level]" time="2.953"><properties><property name="score" value="0.2689039587974548" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[John George Evans\n\nJohn George Evans (April 2, 1809 \u2013 April 18, 1878) was an English-born educator, newspaper editor and political figure in Nova Scotia, Canada. He represented Yarmouth in the Nova Scotia House of Assembly from 1878 to 1882 as a Liberal-Conservative member.\n\nHe was born in Shrewsbury, Shropshire, the son of George Evans and Mary Ann Pritchard, and came to Nova Scotia with his parents in 1816. In 1833, Evans married Ann Squire. He taught school in Digby. Evans was a supporter of the Halifax Conservative Party. He served as deputy collector of customs at Yarmouth from 1843 to 1867 and in 1868 was elected to the provincial assembly. Evans was editor and publisher of the &quot;Yarmouth Herald&quot; from 1859 to 1870. He was elected speaker for the provincial assembly in 1882 but did not serve because he was ill.\n\nHis son George George Evans also served in the provincial assembly.\n]" time="0.521"><properties><property name="score" value="0.9040738344192505" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Flavobacterium gilvum\n\nFlavobacterium gilvum is a Gram-negative, strictly aerobic, rod-shaped and non-spore-forming bacterium from the genus of &quot;Flavobacterium&quot; which has been isolated from sandy soil.\n\n \n]" time="0.139"><properties><property name="score" value="0.9946221113204956" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Ariel Hernandez\n\nAriel Andr\xe9s Hern\xe1ndez Callejas (born 7 August 1995) is a Colombian professional footballer who plays as a midfielder for Deportivo Cali.\n\nHernandez's career began in the youth ranks of Envigado, prior to a move to Deportivo Cali. He played in the club's youth system and featured regularly for the reserves in the Categor\xeda Primera B. Hernandez made his first team debut in a 3\u20130 victory over Alianza Petrolera in July 2014. He appeared in twelve further matches during 2014, including a 2\u20131 defeat to La Equidad and a 1\u20130 loss to Atl\xe9tico Nacional. Hernandez scored his first senior goal on 4 August 2015 in a 2\u20131 victory over C\xfacuta Deportivo. The strike was his sixth of the season and proved to be the winner.\n]" time="0.431"><properties><property name="score" value="0.9960430860519409" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Suspension (mechanical)\n\nIn mechanical engineering, suspension refers to a system of shock absorbers and springs designed to improve the ride quality of a vehicle or other types of machinery. It is an essential component in wheeled vehicles, providing a convenient way to deal with irregularities in the road surface.\n\nMost modern vehicles are fitted with some sort of suspension. It is the primary means of isolating the passengers from the road, as well as absorbing the bumps and vibrations from the road.\n\nThe suspension system on a vehicle provides the following functions:\n\nThere are many systems available that provide all or most of these functions, but most can be categorized by their design and motion characteristics.\n\nThe simplest suspension systems are rigid, and often called &quot;hard suspensions&quot;, because they don't use any fluid damping. Instead, they rely entirely on the elasticity of the metal or the springiness of the suspension elements to soak up the bumps. Some early vehicles used torsion springs as suspension, but modern systems almost all use either coil or leaf springs, or shock absorbers.\n\nThe first and simplest suspension system was used by the French inventor of the wheel, Leonardo da Vinci. It consisted of four flexible limbs connected to the carriage. These limbs served as both the suspension and steering system.\n\nIn addition, Leonardo designed a system that included advanced features such as brake linings, bearings and a double-action brake system. A suspension system for cars was introduced by M. Thon in France in 1898. It consisted of a network of rubber bands that took the place of conventional springs. The rubber bands were enclosed in a conventional coil-spring housing. An advantage of this system was its light weight. Thon's design was introduced to the United States by John M. Curtis in 1904.\n\nBy the 1920s, many independent suspension manufacturers were using the pneumatic springing system as the basis of their systems, as well as using rubber or steel cords for springing. The principal disadvantage of the pneumatic system is that the spring-pressure increases with the compression of the suspension. In addition, a system that is not correctly adjusted will not ride well.\n\nThe third kind of suspension system was developed by the Morgan Motor Company, based in the UK. In this system, the spring is attached to the frame, and the shock absorbers are attached to a special hub, which attaches to the front and rear axles. This system is still sometimes seen in modern cars, since it is less expensive than the other systems and in most cases provides an adequate ride. This system has not been popular in Europe and North America since the 1930s.\n\nThe most common suspension system in modern cars, used in everything from Ford Model Ts to the Chevrolet Corvette, is the &quot;independent front suspension&quot;, or &quot;live axle&quot;. In this system, the suspension is a more or less conventional system of springs and shock absorbers, with the additional feature of a locating link (the king pin in an axle-based suspension). This allows the axle to move up and down relative to the chassis. The locating link ends are attached to the car body, which provides the rolling element to the system.\n\nThe other feature of this system is the axle, which provides the pivoting element of the suspension system. The wheel hub is attached to the ends of the axle, and the wheel itself is attached to the hub. The spring system, located in the body of the vehicle, is attached to the axle via the locating link. The body of the car, therefore, provides the &quot;pivot&quot; of the suspension.\n\nA number of variations on this design have been used, most notably independent rear suspension, in which the rear wheels are located by a trailing arm (a form of wishbone) instead of the traditional transverse leaf spring. Another design commonly found on trucks is the &quot;live rear axle&quot;. In this system, the axle is located by an arm from the leaf spring. The rear wheels can move up and down relative to the chassis without any location from the body.\n\nA design commonly used on many racing cars and on some passenger cars is the DeDion or De Dion suspension, named after its inventors, the De Dion brothers. In this system, the two wheels are located by a triangular or &quot;semi-trapezoidal&quot; bar. The pivot is provided by a bearing block mounted to the car body, while the pivoting point for the wheels is a ball joint on the axle.\n\nThe DeDion design is an expensive design, due to the cost of the wheels, brakes and other parts. However, it does provide excellent wheel location, as the wheels can be moved by several inches in any direction without affecting the steering.\n\nRacing cars generally use the DeDion system, as the location of the wheels is the primary function of the suspension system. However, it is also used on a number of sports cars, including the Jaguar XK120, the Maserati 3500 GT and the Lamborghini Gallardo.\n\nOne of the main disadvantages of the DeDion suspension is that the front wheels are not located directly above the chassis, and there is significant scrub radius. For example, a wheelbase of 86\xa0inches (218\xa0cm) on a vehicle with a wheel diameter of 18\xa0inches (46\xa0cm) would produce a scrub radius of around 8\xa0inches (20\xa0cm). This can cause the car to steer and corner differently depending on the tire wear, since the scrub radius is at its largest when the tires are new.\n\nWhile an advantage of the DeDion system is that it is lighter than a comparable design, this is not a major consideration for racing cars. However, since racing cars are generally not driven at high speeds for long distances, the &quot;chassis bounce&quot; caused by the scrub radius is not an issue.\n\nOne of the first types of suspension system was introduced by Earle Crooker and was called &quot;Independent Coil Spring Suspension&quot;. It was a brand name that existed only in the late 1920s. The name was used by C. G. Ames Co.\n\nIn this system, the springs are attached directly to the frame. In this system, the wheel is attached to the axle, which is supported by a bearing attached to the frame, not the car body.\n\nThe primary advantage of this system is the ease with which the wheelbase can be changed. In most cars, the wheels are located by a link that is attached to the axle and the frame, and the wheels can move relative to the car body. In this system, there is a much smaller length change, as the wheels move up and down relative to the axle.\n\nThe disadvantage of this system is that the body is no longer a &quot;pivot point&quot; for the suspension. The body is now part of the suspension system. This makes it very difficult to change the camber of the wheels (i.e., the angle of the wheel relative to the car body). In addition, the wheel can only move in one plane, as the upper control arm and the lower control arm limit the movement of the wheel. The car may need to be aligned more often, as the geometry of the suspension can change more than a suspension system where the wheels move relative to the body.\n\nThis system is commonly found on American cars built in the 1940s and 1950s, such as the Nash Ambassador, Hudson, and other brands. It is less commonly found on European cars, although it was used on some Mercedes-Benz models]" time="3.777"><properties><property name="score" value="0.9658008590340614" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[L\u2019avvento di Zuckerberg ha definitivamente maturato il passaggio dalla televisione tradizionale a quella digitale. Il successo del prodotto Facebook per Smartphone ha determinato la consacrazione di un nuovo media che non si muove pi\xf9 solo tra i PC ma che, a partire da un apparecchio mobile, si espande a livello planetario. In un solo anno la piattaforma del social network di Menlo Park ha infatti scalato le vette dell\u2019ecosistema tecnologico con un balzo da 5,8 milioni di utenti nel 2010 a quasi 60 milioni di abbonati nel 2011.\n\nSe da un lato le prestazioni del prodotto Facebook per Smartphone hanno assicurato un rapido avanzamento delle quotazioni del titolo in borsa, dall\u2019altro l\u2019entrata in azione del colosso californiano \xe8 stata uno dei fattori principali per l\u2019incremento dei dati di traffico degli oltre 90 milioni di italiani che utilizzano ogni giorno Internet dal proprio smartphone.\n\nMentre negli Stati Uniti la situazione non \xe8 ancora perfetta ma le promesse sono grandi, in Italia il Facebook per Smartphone ha assicurato a Facebook il consolidamento di una base di utenti in crescita. Con le recenti performance che sta raggiungendo il social network, sempre pi\xf9 persone decidono di interagire dal loro smartphone con i contatti personali e gli amici di Facebook, grazie alle nuove funzionalit\xe0 dell\u2019applicazione che ha recepito le nuove esigenze degli utenti. Il fenomeno sta infatti dilagando in tutto il mondo, come dimostra il caso della Spagna che in poco tempo ha passato da 4 milioni di utenti a 15 milioni.\n\nTra le novit\xe0 dell\u2019applicazione troviamo un nuovo caricamento della home, in cui sono stati aggiunti i nuovi pulsanti \u201cGiochi\u201d e \u201cScrivi al tuo amico\u201d, inoltre sono stati eliminati i due pulsanti \u201cFoto e Storie\u201d che avevano originariamente l\u2019incarico di condividere le immagini su Facebook e Instagram.\n\nIl suo funzionamento si basa su un semplice scambio di messaggi in cui l\u2019utente, dall\u2019apposita interfaccia di Facebook per Smartphone, pu\xf2 utilizzare una tastiera per comporre il testo e inviarlo ai suoi contatti. L\u2019opzione gi\xe0 si integra con il servizio di messaggistica istantanea di Facebook, WhatsApp, e con Messenger. Per effettuare una ricerca \xe8 necessario digitare il nome di un contatto nell\u2019apposita barra delle ricerche.\n\nInoltre per chi non dispone di una connessione Internet a banda larga, \xe8 stata introdotta la possibilit\xe0 di caricare un\u2019immagine senza l\u2019obbligo di avere una connessione internet. L\u2019immagine viene caratterizzata da una piccola animazione che si sviluppa gradualmente sul display del telefono, permettendo di vedere il progressivo aggiornamento dell\u2019immagine.]" time="2.202"><properties><property name="score" value="0.9998014171918234" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99980142&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99980142
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Gerry is head of commercial and operational support and has worked for the Trust since 1988.\n\nHaving been educated in Scotland, he was an apprentice engineer and a welder, specialising in the design and maintenance of special mechanical and electronic equipment used in nuclear power stations.\n\nHe worked on naval nuclear submarines at Rosyth, before working in the oil and gas industry and eventually becoming an electrician at oil rigs in the North Sea.\n\nAfter working in various positions in the engineering industry, he joined North East Ambulance Service in 1987, initially as an electrical engineer. He then became training co-ordinator, before joining Northumberland, Tyne &amp; Wear Ambulance Service in 1988 as a training officer.\n\nAs an operational support manager, Gerry now manages the administration support, communications and logistics for the emergency and non-emergency services in the area.]" time="0.470"><properties><property name="score" value="0.999751627445221" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Minister for Justice and Equality, Charlie Flanagan TD, and Minister for Social Protection, Leo Varadkar TD, today launched a joint campaign in support of the Crosscare Pathways Project, a nationwide outreach service for rough sleepers.\n\nLaunched this morning in Dublin, the \u2018Living On The Streets\u2019 campaign aims to draw attention to the fact that there are approximately 1,000 rough sleepers in Ireland. The aim of the project is to help these rough sleepers access the appropriate support services, whether it be residential or non-residential. The campaign highlights the issue of rough sleeping in Dublin city, which is rising at an alarming rate, with an increase of over 40% in the number of people living on the streets since 2013. The campaign\u2019s focus is on raising awareness of this issue.\n\nThe campaign involves advertisements on the city\u2019s LUAS and Dublin Bus, as well as featuring on O\u2019Connell Street, Portobello and Talbot Streets. It also involves the installation of posters across the city, specifically in locations where the homeless are more likely to reside.\n\nSpeaking at the campaign launch, Minister Flanagan said: \u201cRough sleeping has reached record levels and has become a visible problem in our capital city. The launch of this campaign will increase awareness of the services provided by the Crosscare Pathways Project to rough sleepers, so that they are aware of the options available to them. Rough sleepers should be referred to the services that will best suit their needs. As Minister for Justice and Equality, I am particularly concerned to ensure that rough sleepers are not penalised for begging and sleeping on the streets.\n\n\u201cThe Government is determined to prevent and reduce homelessness, to support rough sleepers to exit homelessness, and to prevent homelessness for others at risk. Over \u20ac70 million has been provided to support homeless services and prevent homelessness, which includes funding to frontline services such as the Crosscare Pathways Project, who work to prevent rough sleeping by providing outreach support to rough sleepers.\u201d\n\nMinister for Social Protection Leo Varadkar said: \u201cThe \u2018Living On The Streets\u2019 campaign brings home the reality that the street is a harsh and unforgiving place to be, particularly during the winter months. The response from the public will be crucial if we are to address this problem effectively.\n\n\u201cThere are over 1,000 rough sleepers in Dublin city and unfortunately this is a significant increase on previous years. That is why it is so important that we engage the public, using this campaign as an opportunity to highlight the services available to rough sleepers, including the Crosscare Pathways Project. It is a complex problem that requires a complex response, involving a number of different organisations, as well as a societal change in attitude.\n\n\u201cWith the help of my colleague, Minister Flanagan, we have increased funding for the homeless to \u20ac70 million this year, which will provide more than 3,000 more beds in a wide variety of services. We have also agreed to provide an additional \u20ac5 million over the next two years.\u201d\n\nCrosscare Pathways Project is a national organisation with local services based in Ballymun, Jobstown, Dublin city centre, Trim, Waterford and Wexford. It provides a range of services to homeless people and others at risk of homelessness, as well as people sleeping rough, particularly those sleeping on the streets of Dublin. The Crosscare Pathways Project also provides services for homeless people in Carlow, Cork, Donegal, Dublin North, Dublin South and Dublin West.\n\nThe project\u2019s services include outreach services, sleeping bags, socks, warm clothes, food, clothing, showers, laundry and toiletries.\n\nThe campaign is the work of the cross-agency Dublin Central Area Partnership (DCAP), which is comprised of An Garda S\xedoch\xe1na, Dublin City Council, Dublin Regional Homeless Executive (DRHE), Dublin Fire Brigade, Dublin Regional Housing Executive (DRHE), Dublin Street Pastors, Homeless Hotline, Irish Red Cross, Samaritans, Simon Community and Street Reach.\n\nThe campaign is supported by a number of major partners including Dublin City Council, Bus \xc9ireann, Dublin Bus, Dublin Airport Authority, Transdev and Luas Cross City.\n\nENDS]" time="2.154"><properties><property name="score" value="0.6971581935882568" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[During the war in Bosnia, in the capital city of Sarajevo, there was a hotel that was shelled repeatedly, and you could tell by the black marks on the facade which were caused by sniper fire and which were caused by shrapnel.\n\nHow do I know? Because my cousin owned the hotel, and I knew her back when she was a young student in Yugoslavia in the 1960s.\n\nShe was smart, but she was]" time="0.217"><properties><property name="score" value="0.81319260597229" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[See the live USD price. Control the current rate. Convert amounts to or from BABB and other currencies with this simple calculator.\n\nHow did the currency on yesterday? The average value Babble price for convert (or exchange rate) during the day was $0.009986. Min. Babble value was $0.009322. Max. BABB price was $0.011702. BABB price dropped by 7.5% between min. and max. value. Let's see what's next.\n\nOlder news about Babble\n\nThe Babble increased by 15.27% on Thursday 17th of September 2019 Let's take a look at interesting data from yesterday. BABB price increased by 15.27% between min. and max. value. Max. BABB price was $0.010422. Min. Babble value was $0.008959. The average value Babble price for convert (or exchange rate) during the day was $0.010405. Keep it up.\n\nThe Babble dropped by 17.49% on Wednesday 16th of September 2019 How did the currency on yesterday? The average value Babble price for convert (or exchange rate) during the day was $0.010164. Min. Babble value was $0.009265. Max. BABB price was $0.011671. BABB price dropped by 17.49% between min. and max. value. Let's see what's next.\n\nThe Babble dropped by 23.23% on Tuesday 15th of September 2019 How was the currency exchange rate changed on yesterday? BABB price dropped by 23.23% between min. and max. value. Max. BABB price was $0.012727. Min. Babble value was $0.009289. The average value Babble price for convert (or exchange rate) during the day was $0.012042. Watch the next day.\n\nThe Babble increased by 16.08% on Monday 14th of September 2019 And we have data for yesterday. Min. Babble value was $0.009069. Max. BABB price was $0.011378. The average value Babble price for convert (or exchange rate) during the day was $0.010781. BABB price increased by 16.08% between min. and max. value. Certainly it is good news for all.\n\nThe Babble dropped by 7.02% on Sunday 13th of September 2019 Let's see on yesterday. BABB price dropped by 7.02% between min. and max. value. Min. Babble value was $0.010600. Max. BABB price was $0.011775. The average value Babble price for convert (or exchange rate) during the day was $0.011251. Let's see what's next.\n\nThe Babble increased by 5.37% on Saturday 12th of September 2019 Let's take a look at interesting data from yesterday. Min. Babble value was $0.010724. Max. BABB price was $0.011425. The average value Babble price for convert (or exchange rate) during the day was $0.010947. BABB price increased by 5.37% between min. and max. value. Certainly it is good news for all.\n\nThe Babble increased by 7.5% on Friday 11th of September 2019 How was the currency exchange rate changed on yesterday? BABB price increased by 7.5% between min. and max. value. Min. Babble value was $0.010348. Max. BABB price was $0.011268. The average value Babble price for convert (or exchange rate) during the day was $0.010825. Keep it up.\n\nAnother conversions\n\nUsc to Babble, Uro to Babble, Uralscoin to Babble, Uscoin to Babble, USD-e to Babble, Usdq to Babble, US Dollar to Bangladeshi Taka, US Dollar to Banca, US Dollar to Bancor, US Dollar to Bankcoin, US Dollar to Bancorplus, US Dollar to Bankera, US Dollar to Banca,]" time="2.334"><properties><property name="score" value="0.01458844542503357" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01458845&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.01458845
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Mick Campbell: Some Hilarious (but Informed) Jokes About Health Reform\n\nHealth-care reform is under way in the US and health-care reform is under way in Britain. Some people may not have noticed the second fact. Over here, the NHS is now the NHS 2.0, or 3.0, or whatever they are on to. We\u2019re getting used to the idea that it will always be changing, and changing in ways we\u2019re not in total control of.\n\nWith my dry British sense of humor, I\u2019ve always enjoyed the NHS. In fact, I like it so much that I even thought of emigrating to a country with an NHS \u2013 like, say, the USA. So what a treat to discover in my in-box this morning a compilation of jokes, half of them at the expense of the USA and half at the expense of the UK, about health-care reform.\n\nFrom London, Chris Salisbury compiled them, and here are a couple of samples. The first is about healthcare reform in the USA:\n\n\u201cHave you heard about the new Congressional Health Care plan?\u201d\n\n\u201cYeah, the government\u2019s going to give you half of what it takes away.\u201d\n\nAnd the second, about the NHS:\n\nA man is at the doctor\u2019s office, and the doctor tells him, \u201cI\u2019ve got some good news and some bad news for you.\u201d\n\nThe man says, \u201cWell, I can\u2019t take the bad news right now, so give me the good news first.\u201d\n\nThe doctor says, \u201cWell, the good news is that you have an 18-inch penis.\u201d\n\nThe man looks stunned for a moment, and then asks, \u201cWhat\u2019s the bad news?\u201d\n\nThe doctor says, \u201cYour brain\u2019s in your dick.\u201d\n\nThese jokes may be inaccurate, of course. For example, the doctors\u2019 pay packets are pretty great, and some of them can do much better work in the US than in Britain. But one thing I know is for sure: it\u2019s great to see the NHS getting some serious attention in the American media. I bet it makes the doctors happy.]" time="1.185"><properties><property name="score" value="0.971766471862793" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Hello.\n\nI have been trying to download the RPM of a file that resides on a university FTP server. I can download files with wget and ftp, but not with rpm. The file is a tar file. I want to be able to download it with rpm. Is this possible?\n\nThanks!]" time="0.217"><properties><property name="score" value="0.9425084590911865" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.94250846&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.94250846
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[\n\nBuchbesprechung\n\nHeinz Wernicke: Das zerrissene Land.\n\n\n\n1\n\nEin erstaunlich gutes Buch, das sich auf das politische und wirtschaftliche Geschehen Deutschlands in der Epoche von 1870 bis 1950 bezieht. Es zeigt die Interessen, aus denen sich die in der ersten H\xe4lfte der Epoche g\xfcltigen Allianzen zwischen Reichswehr und Arbeiterschaft, den Kirchen, den Nazis und den Kapitalisten in der zweiten H\xe4lfte zu so erstaunlich verschiedenen Gebilden wandelten.\n\n\n\n2\n\nDas Buch kommt zu dem Ergebnis, dass der R\xfcstungsstaat, der heute aus der Weltkriegsgeneration stammt, sich schon in den letzten Jahren des Kaiserreichs vorbereitete und dass er nicht von der Weimarer Republik, sondern von dem inzwischen gescheiterten Gegner der Weimarer Republik konzipiert wurde.\n\n\n\n3\n\nEs schildert die Haltung des Kaisers und seines Adlatus Ebert zu der als Gro\xdfmacht \xfcberlebenswichtigen Schwerindustrie und zu dem Staat, der seit 1871 f\xfcr die industrielle Entwicklung garantiert hat.\n\n\n\n4\n\nEs stellt dar, wie und warum Hitler nicht auf einen Einmarsch in die Tschechoslowakei gefolgt ist und zeigt die Beweggr\xfcnde f\xfcr diese Entscheidung. Diese sind bis heute in der Diskussion \xfcber den Charakter des Nationalsozialismus vernachl\xe4ssigt worden.\n\n\n\n5\n\nEs berichtet \xfcber die Zusammenarbeit von milit\xe4rischer und ziviler Macht im Verwaltungsapparat der Wehrmacht w\xe4hrend des zweiten Weltkriegs, von den Vorg\xe4ngen im Inneren des Staatsapparats, der Gesellschaft und der Armee bis hin zum Prozess gegen die deutschen]" time="1.387"><properties><property name="score" value="0.9957321882247925" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99573219&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99573219
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Most of the novels I enjoy are focused on romance, in whatever form it may come. I believe that there\u2019s something for everyone to enjoy, as the author does such a great job of portraying the characters in all their strengths and flaws. Sometimes they go a little over the top with the angst, and it isn\u2019t very realistic, but I understand why. There are authors who never experience any type of conflict with their characters, and the books just feel very one dimensional. Some books can have some really crappy endings, but I have learned to stay away from books that end with the couple together and happy but someone was murdered or something along those lines. I don\u2019t care for books that have the couple separated, but they\u2019ll come back together in the end.\n\nWhat I want in a good romance book is a good, if not a great, ending that makes me feel good, no matter what happens to the characters. As I mentioned earlier, I don\u2019t like books that end in tragedy, and they usually do end in tragedy. There are some that aren\u2019t too tragic, but they just aren\u2019t the endings that I like. This is where the author\u2019s job becomes very difficult. They have to give the reader a reason to feel the way they want them to feel about the characters. If they feel bad, they\u2019ve done their job. If they feel happy, they\u2019ve done their job. As a reader, I\u2019m not looking to be outraged, nor am I looking to be happy. I\u2019m just looking for something that makes me feel.\n\nOn the flip side of that, I know a lot of people don\u2019t want to be emotionally manipulated. They just want to read a book and enjoy it. They don\u2019t want to be forced to feel one way or another. They want to feel neutral. Neutral, to me, is a huge accomplishment in a romance novel. To me, a romance novel has to be able to get a person, even if they don\u2019t think it will, to feel something for the characters. In most books, people want to know how they can make a particular character do what they want them to do. In a romance novel, that\u2019s the main character. They want to know how they can make the characters fall in love, because they see themselves in one or both of them. How would they do it? How would they fix it?\n\nThat is what a reader does when they read a romance novel. They want to know how to fix the problems that the characters have. This is why the ending is so important. It has to make the reader happy, but not to the point where they roll their eyes or feel like they were taken for a ride. I have a great example. In the movie Now and Then, a woman tells her daughter that she is happy. When her daughter tells her that she is, too, she tells her that it\u2019s because they\u2019re both lying. They don\u2019t want to make each other sad, so they\u2019re lying to each other. This is the exact thing that happens in a lot of romance novels, in my opinion. They\u2019re trying to make the readers feel good, but in the process they\u2019re lying. To me, the ending has to make the reader feel what the author was trying to make them feel, not what they want them to feel.\n\nI\u2019m not saying that the book has to end the way I want it to. What I\u2019m saying is that I want the author to be honest. I want them to give me a believable ending that isn\u2019t just thrown together so that they can end the book. I want the ending to mean something, and I think that\u2019s what most readers want. Most people like to be happy, and a good book will give them that. The ending has to be good enough that the reader doesn\u2019t feel like they wasted their time, but also good enough that they\u2019re not completely disappointed. They have to be happy with the story, even if it wasn\u2019t exactly what they wanted.\n\nWhat are some of your favorite romance book endings?]" time="2.172"><properties><property name="score" value="0.9997991025447845" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[1) Running on behalf of a person\n\nThese dates are the running dates for campaigns on behalf of clients. If you are looking to book a campaign on behalf of your brand, please email theteam@roar.co.nz\n\n2) Need to know more?\n\nIf you are looking to know more about the process of booking a campaign, you can find more information on our website. Or alternatively, get in touch with our team at hello@roar.co.nz]" time="0.276"><properties><property name="score" value="0.9998231530189514" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Ask.com gives you a list of possible answers when you type in a question. Then it shows you the number of people who searched for those questions.\n\nNot a new idea, but very handy.\n\nFiled in: Technology \u2022 Web]" time="0.173"><properties><property name="score" value="0.9977366924285889" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[There\u2019s nothing more adorable than a real life couple who gush over one another on social media and share the same love story. What\u2019s even more adorable is when you know that they found each other through social media.\n\nTake Brandon Flowers and his wife, Tana. The two met on Twitter and now, they\u2019re celebrating their five year anniversary. If you follow either of their social media accounts, you may have noticed some cute posts from them about their relationship. We want to thank them for allowing us to have a look into their love story, because they\u2019re too cute!\n\nAs fans of The Killers, we love hearing about their adventures.\n\nThey\u2019re pretty cute!\n\nThere\u2019s nothing more adorable than a real life couple who gush over one another on social media and share the same love story. What\u2019s even more adorable is when you know that they found each other through social media.\n\nTake Brandon Flowers and his wife, Tana. The two met on Twitter and now, they\u2019re celebrating their five year anniversary. If you follow either of their social media accounts, you may have noticed some cute posts from them about their relationship. We want to thank them for allowing us to have a look into their love story, because they\u2019re too cute!\n\nAs fans of The Killers, we love hearing about their adventures.\n\nThey\u2019re pretty cute!\n\nI'd just like to say that my wife is everything. \u2014 Brandon Flowers (@flowerboy) August 5, 2017\n\nWedding Day. Not on social media. So I don't know what people think. I know my wife is everything. That's all I need. \u2014 Brandon Flowers (@flowerboy) August 5, 2017\n\nOn the last night of a tour, getting some wife-y kisses. And in case anyone was wondering. This is how we roll. pic.twitter.com/zSXlfUgKXd \u2014 Brandon Flowers (@flowerboy) August 7, 2017\n\nOh yeah. Still be on tour, still be making music. Not yet a father, but that's coming. Still have some things left to do. https://t.co/CJ1vOs8lYA \u2014 Brandon Flowers (@flowerboy) August 7, 2017\n\nYes. It's true. I am getting married. And I'm marrying my best friend. And my favorite person. https://t.co/uOe3DjtMtQ \u2014 Brandon Flowers (@flowerboy) August 5, 2017\n\nThe next night, he admitted that he has not yet become a father, but he is working on it.\n\nDidn't say I was a father yet. Wasn't pregnant before. Wasn't even married. But I love my wife and my future kids. pic.twitter.com/uVEXZHScGz \u2014 Brandon Flowers (@flowerboy) August 6, 2017\n\nYep. Only day two and already love you so much more than I did yesterday. No, that's not enough. Still not enough. pic.twitter.com/cItd8b8HxY \u2014 Brandon Flowers (@flowerboy) August 8, 2017\n\nYou are a goddamn genius. No one can make me laugh like you. \u2014 Brandon Flowers (@flowerboy) August 8, 2017\n\nThe next day, Brandon shared the first picture of the newlyweds.\n\nTana and I on our wedding day. pic.twitter.com/6R5xT0P5zT \u2014 Brandon Flowers (@flowerboy) August 7, 2017\n\nI can't wait to see you two together at our wedding. \u2014 Brandon Flowers (@flowerboy) August 7, 2017\n\nI love my wife and I'm going to spend the rest of my life loving her. https://t.co/2h7LXqHxjC \u2014 Brandon Flowers (@flowerboy) August 7, 2017\n\nBrandon gave a shout out to his brother and the rest of the band for not spoiling their special day.\n\nThis is how much I love my wife. I would've wanted to come here tonight. But I didn't. https://t.co/ox7XY3sPGO \u2014 Brandon Flowers (@flowerboy) August 8, 2017\n\nJust a normal day. A normal day with my wife. She is my favorite person in the world. We have no plans. pic.twitter.com/vNNkIOCcUQ \u2014 Brandon Flowers (@flowerboy) August 8, 2017\n\nCurious to see the rest of their love story?\n\nHappy 5 years to the love of my life, my best friend and my favorite person. I still feel like I'm the luckiest]" time="2.364"><properties><property name="score" value="0.9997905790805817" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The mainline of PC\u2019s arrived on Friday in great shape. In addition to the newer No. 3 (blue) there was an earlier No. 4 (green), a No. 5, and two later No. 5\u2019s. To fill the gaps between these 5\u2019s I bought two\u2026well, three\u2026plows from another fellow who also had lots of couplers and castings. It\u2019s good to have some of each for testing fit of things, but for permanent operation I\u2019ll probably want to get a handful of No. 5\u2019s.\n\nNow it\u2019s back to finishing the minis and collecting the accessories. The insides of these last few cars could use some work, like fitting the walls for the corner curves. I\u2019ll do that at the same time as adding the doors, and a couple of weeks should have them done. Once the house is done I\u2019ll probably just keep building trains and cars for a while, rather than finishing the accessory set.\n\n-- Charlie M. &quot;Woodworking - patience = firewood&quot;]" time="0.540"><properties><property name="score" value="0.5196418762207031" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.51964188&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.51964188
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[To a certain extent, the goal of the Trump administration's trade policy, as explained in a speech given by Commerce Secretary Wilbur Ross on March 1, is the creation of a &quot;reciprocal&quot; or &quot;two-way&quot; trading system in which all nations treat each other equally. While not naming China, the intended target of that policy was clear. The Chinese government, on the other hand, was not impressed.\n\nWriting in the English-language edition of China Daily on Monday, Ma Kai, vice-premier of the State Council, or China's cabinet, and chair of the National Development and Reform Commission (NDRC), countered the Trump administration's position by arguing that China has been the most open to foreign investment in the world for the past three decades.\n\nIn Ma's view, China is playing by the rules of free trade. The problem is that the rules that America wants are not the same rules that America has been playing by. The United States wants to be able to sell its goods and services into China, but it wants to block Chinese goods and services from being sold in America.\n\nSpecifically, Ma pointed to the issue of the U.S. $347 billion trade deficit with China, which is the focus of a U.S. investigation to determine whether China's government has unfairly subsidized certain industries.\n\nBut China is not subsidizing its industries, according to Ma. The U.S. trade deficit is simply a function of America's industrial overcapacity and under-consumption, as well as Chinese market and regulatory restrictions on certain industries.\n\nIn Ma's view, there is an easy solution to the trade deficit, and that is to allow China to import more agricultural products, textiles, and auto parts from the United States.\n\nThis is not the first time Ma has been forced to explain the Chinese position in the ongoing trade war. Last week, he held an hour-long interview with Charlie Rose, the television host. In the interview, Ma also said that he would not criticize the Trump administration for the position it is taking. Instead, he is simply trying to point out to the Trump administration that it is China that has been working toward a free trade regime, and that the American government has not been as open as the Chinese government has been in allowing Chinese companies to invest in American businesses.\n\nWhat Ma is really doing is not simply trying to engage in a debate about the nature of free trade, but to make sure that Americans understand how the Chinese government is fighting to protect its own economy from the American tariffs, while at the same time, maintaining a narrative of the mutual benefits of free trade.\n\nMa's interview with Charlie Rose can be viewed here:]" time="1.349"><properties><property name="score" value="0.9997834165891012" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[WHY PEOPLE KEEP COMING BACK\n\nMany people come back again and again to our Salon. Why?\n\nThey come back because they love the salon experience. They are getting a quality cut and colour, a relaxing shampoo and spa treatment, and an enjoyable, friendly service. They come back because they know they will leave feeling good. Our customers tell us how they love our staff, our warm atmosphere, our extra touches, and our positive, welcoming approach.]" time="0.267"><properties><property name="score" value="0.9450801610946655" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Selected Test Results\n\nWhat works\n\nInstaller; not required to finish installation of wine for use in the installer. Installs just fine on linux-gnu, after adding 32-bit windows support in winecfg. For some reason the driver doesn't get installed in Windows.\n\nWhat does not\n\nInstallation of the driver in windows\n\nWorkarounds\n\nWhat was not tested\n\nFull installation of windows; only tested enough to install drivers\n\nHardware tested\n\nGraphics:\n\nGPU:\n\nDriver:\n\nAdditional Comments\n\nThis is not a great test since the windows installation is a mess, but it seems to install the driver properly on wine.]" time="0.308"><properties><property name="score" value="0.860663115978241" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.86066312&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.86066312
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[What can I say about Sweet Lou's Bakeshop and Cafe in L.A.'s Chinatown other than, well, it is a little gem. If I had to choose one thing to eat every day for the rest of my life it would be a quiche from Sweet Lou's. Seriously. When I went to the Farmers' Market on Monday I bought one. I bought a quiche at the bakery next door too but I'm not too sure how good that one is. But I am very sure that the quiche from Sweet Lou's is absolutely delicious. And it's made from scratch.\n\n\n\nLou is such a sweetie. I know he'd be a fun guy to have coffee with. He owns a gorgeous old school bakery in Chinatown and the smells of fresh baked bread and sweet treats are incredible. So if you are ever in L.A., and you happen to be in the Chinatown area, go on over and get a treat from Lou. You'll be glad you did. And he even has a menu for the cafe part of his establishment and everything on it looks pretty good too. I need to get over there soon and try it out.\n\n\n\nHis website is a little hard to find, I'm not sure if that's just me or if it's hard for everyone. But I did find it.\n\n\n\nI'm a big fan of Lou's pies. They look good enough to eat!]" time="0.748"><properties><property name="score" value="0.9239310920238495" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The popular Excel shortcut keys are Ctrl + C to copy and Ctrl + V to paste, Ctrl + B to bold, Ctrl + I to italicize and Ctrl + U to underline. Ctrl + M activates the Enter key. Pressing the F1 key opens the help file for the active program. The Windows key launches the Start menu and the arrow keys highlight and select items on the screen. For many other common key combinations, check out www.usingtech.com/excel.\n\nWhen you're using Excel's screen controls, it helps to know the shortcut key combinations. In addition to copying, pasting and formatting, the Ctrl key also serves as a modifier key that lets you move the cursor to another area of the screen or to navigate through lists.]" time="0.346"><properties><property name="score" value="0.9989129304885864" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99891293&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99891293
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The G20, the \u201cShangri-La\u201d and the Global Trade War\n\nThe G20 is currently gathering in the Chinese city of Hangzhou. China has been actively preparing for the summit and has invested in the improvement of its infrastructure and public services. However, the relationship between China and the US has deteriorated in the past few weeks. The US has asked the WTO to look into what it describes as unfair treatment of US companies in China. Furthermore, the EU, the US and Japan have called for retaliatory measures against China, after the announcement of a new Chinese law to allow China to bypass international patents. On the other hand, there are indications that the Chinese government will relax foreign ownership restrictions in its auto industry, and also encourage the outflow of capital from the country. The theme of the summit is &quot;Better Growth, Better World&quot;, and the three themes will be &quot;growth&quot;, &quot;inclusive growth&quot;, and &quot;sustainable growth&quot;.\n\nAccording to a recent report by the World Bank, China has a great economic potential, yet faces severe challenges. These challenges include maintaining the economic growth, the management of urbanisation, the fight against corruption, the increasing inequality and the aging population. On the other hand, the report also concludes that China has the opportunity to become a major contributor to global growth, thanks to the fact that China is an important producer of capital, the country is growing in the manufacturing sector, and is also important in terms of consumption, which is expected to grow by more than 6% this year. The report also highlights the fact that China has been the biggest contributor to the global GDP growth in the last 15 years, with a contribution of 30%.\n\nOn the other hand, China is also a top trading partner of the US, with trade between the two countries amounting to more than $600 billion. Therefore, the potential of a trade war between the US and China would be dramatic. As the leaders of the two countries are gathering in the same city, the G20 summit will be an opportunity to address many of the issues between the two countries. As an example, the summit will discuss the sustainability of trade, both for exports and imports.\n\nIt seems that the US and China may be heading towards a trade war. If such a trade war will occur, it will have devastating consequences on both the US and Chinese economies, as well as on the global economy. Furthermore, the WTO may face a severe problem, as it would probably have to be the court of the two parties involved in the trade war. Thus, the G20 summit will be an important meeting and will represent a key opportunity for the leaders of the two countries to reach an agreement.]" time="1.365"><properties><property name="score" value="0.9997930328051249" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Juan De La Torre, a 10-year-old artist, designed a colorful new look for the school. He won the contract over five other artists because his design most closely resembled the school's own renderings, reported Milenio.]" time="0.142"><properties><property name="score" value="0.9850605130195618" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.98506051&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.98506051
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Takes place three months after the events of Season 4, 'Brother's Keeper'.\n\nMax and Liz's lives were put on hold after George's tragic death, but now that the madness of the wedding season has passed, they're finally ready to start their life together. They've settled into their apartment, they're adjusting to working as a team, and even Chompy's finally feeling more like family than just another new roommate. They're finally back to their normal routine.\n\nExcept now, it's not.\n\nNow it's crazy.\n\nMax and Liz are supposed to be getting ready to take their first vacation in a very long time, but it turns out Liz isn't feeling up to it. She has a persistent cough, she's not sleeping well, and she's even started getting headaches. Max tries to be supportive as she talks about calling their doctor, but the more she says it, the more certain he becomes that something is seriously wrong.\n\nAnd then Liz stops talking. She won't talk to the doctor on the phone, she won't talk to Max, and she won't talk to the doctor when he comes to their apartment. She just smiles politely and lies to the doctor's face about how she's been feeling, and then makes Max drive her home as soon as the appointment is over.\n\nMax finally gets her to talk to him once they get home, but she can't tell him what's wrong. She just says that she's been feeling off, and doesn't know why. She refuses to go to the hospital, and eventually falls asleep on the couch, leaving Max alone to think about the phone call he has to make the next day.\n\nHe tells Hank, and his mind immediately goes to the worst case scenario. It doesn't help that when he went to visit Liz earlier, he saw a few of the guys from work lurking around outside their apartment. He had a feeling they weren't just paying him a social call.\n\nHe ends up calling Hank at two in the morning, rambling and stuttering and eventually saying it all in one long string of words. When he's done, Hank doesn't even need to tell him he was right.\n\nThe next day, Max wakes up to find Liz looking better than she has in weeks. There's a huge smile on her face, and she looks so happy. He's thrilled to see her finally feeling better, but he can't help but wonder if she's just hiding the fact that she's sicker than ever.\n\nHe can't bring himself to ask.\n\nShe starts working on her first assignment with a new partner a few days later. This new partner's name is Adam, and he seems like a pretty nice guy. Max watches Liz interact with him and can't help but wonder if he's doing the right thing by keeping her away from the wedding business. He knows how much she loves the job, and this is the first time he's seen her truly smile in a while.\n\nOf course, he doesn't get the chance to mention that to her. The moment he opens his mouth, she cuts him off and says she has a date with Adam.\n\nLiz goes on a date with Adam, and comes home late again. Max's heart sinks when she opens the door and he sees the smile on her face. It's not her usual smile, the one she saves for Max. It's the smile she reserves for Chompy and her art, and it's a smile he hasn't seen in a long time.\n\nMax knows what he's supposed to do, and he knows what's best for Liz. He knows that keeping her away from the wedding business is the best thing he can do for her. But he can't bring himself to do it.\n\nHe still isn't sure how she's going to take it when he tells her.\n\n&quot;Oh my god,&quot; she says, and Max's breath catches in his throat.\n\nHe doesn't get a chance to tell her that he's sorry.\n\nShe rushes off the bed and takes the two steps necessary to cross the room. She throws her arms around him, and Max's whole body tenses at the unexpected contact. He can feel her shaking in his arms, and the sound of her sobs echo through his apartment.\n\nShe tells him she loves him, and he tells her that he loves her too. She pulls away, and he can see the tears streaming down her face. She says that she has to go, and he knows he has to tell her what he did.\n\nHe knows she'll be angry with him, and he's not sure how she's going to react.\n\nShe kisses him on the cheek, and then disappears out the door.\n\nHe doesn't go after her.\n\nMax calls Liz the next day, and tells her to come over to his apartment. She's silent the entire drive, and he's not sure what she's going to do when she walks in.\n\n&quot;I'm sorry,&quot; he says, and she freezes in the middle of the living room.\n\n&quot;I have something to tell you,&quot; he says, &quot;and I'm going to need you to promise me that you'll stay calm.&quot;\n\nShe gives him a slow nod, and he knows she's not going to like what he's about to say.\n\n&quot;I called Hank, and I told him that you were sick. That's why you haven't been going to work. I told him that you needed to take some time off, because you had a really serious case of strep throat.&quot;\n\nShe opens her mouth to say something, but he interrupts her and keeps going.\n\n&quot;He didn't want to believe me at first, but I convinced him. I told him that it wasn't a big deal, but he needs to know if you start showing signs of weakness. He's going to call in a few days to check up on you. I told him you wouldn't answer your phone, and that's why he's calling here instead.&quot;\n\nHe sees the expression on her face, and he knows it's time to tell her the truth.\n\n&quot;You're really sick, aren't you?&quot; she asks.\n\nHe nods.\n\n&quot;How sick?&quot; she asks, and he knows this is the moment he's been dreading.\n\n&quot;Very sick,&quot; he says, &quot;It's Hodgkin's lymphoma. And it's probably already in your lungs.&quot;\n\n&quot;When did you find out?&quot; she asks, and Max can't stand to look at her. He turns away, and he's silent for a few moments.\n\n&quot;I knew as soon as we got home,&quot; he says, and she goes completely still.\n\n&quot;Max,&quot; she says, and he turns around.\n\n&quot;I was afraid,&quot; he says.\n\n&quot;You didn't have any right,&quot; she says, and he can see the tears in her eyes.\n\n&quot;It was a bad idea,&quot; he says, &quot;I know it was a bad idea.&quot;\n\n&quot;What about me?&quot; she asks, and he knows what she's asking.\n\n&quot;You have to promise me that you'll stay out of the wedding business for a while,&quot; he says, &quot;I can't risk you getting hurt.&quot;\n\nShe shakes her head.\n\n&quot;You can't stop me from doing this,&quot; she says.\n\n&quot;Liz,&quot; he says, and his voice cracks.\n\n&quot;I'm not giving up on my dream, Max,&quot; she says.\n\n&quot;It's not a dream,&quot; he says, &quot;It's your life. It's your job, and I can't risk you getting hurt.&quot;\n\nShe shakes her head again.\n\n&quot;You can't tell me what to do,&quot; she says.\n\n&quot;I can and I will,&quot; he says.\n\n&quot;We're not a team,&quot; she says, &quot;We're not the Millers.&quot;\n\n&quot;You're not the Miller,&quot; he says.\n\n&quot;No,&quot; she says, &quot;I'm the Keller.&quot;\n\nHe's never seen her like this before, and he realizes she's not going to let him stop her.\n\n&quot;I'm doing this,&quot; she says.\n\n&quot;No, you're not,&quot; he says, and he can see the fire in her eyes.\n\n&quot;Don't tell me no,&quot; she says, &quot;Don't tell me I can't do this.&quot;\n\n&quot;I'm not telling you that,&quot; he says, &quot;You're going to do this, no matter what I say.&quot;\n\nShe shakes her head.\n\n&quot;I can't stop you,&quot; he says, &quot;I know I can't. I'm not going to force you to do anything, but I am going to ask you to think about what you're doing.&quot;\n\nShe just stares at him.\n\n&quot;You're not going to stop, are you?&quot; he asks.\n\nShe shakes her head, and then walks out the door.\n\nLiz calls Max from work a few days later.\n\n&quot;Why did you do it?&quot; she asks.\n\n&quot;I thought I was doing the right thing,&quot; he says.\n\n&quot;I don't want to be a burden to you,&quot; she says.\n\n&quot;I thought you were sick,&quot; he says, &quot;I was just trying to help.&quot;\n\n&quot;I know,&quot; she says, &quot;And I appreciate it. It's just\u2026I'm not really sick. At least, not anymore.&quot;\n]" time="5.073"><properties><property name="score" value="0.99958436191082" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[In this section we describe and analyze the effects of different problems on the model performance. In particular we consider the following problems: the serial position effect, the order effect, the semantic (in)sensitivity, and the distance effect.\n\nFigure 6.16 shows an example of text results of three sequences of three sentences. The whole process of filling and querying a WBS model is shown in Figure 6.17.\n\nWe can observe that in all three sequences the first sentence (\u201cA citizen of A is driving a car\u201d) is more frequently retrieved than the second sentence (\u201cA citizen of B is driving a car\u201d). The number of retentions for the first sentence is also higher than the number of retrievals for the second sentence, which can be attributed to the lexical and syntactic differences between these two sentences.\n\nAnother common problem in short sentences is the inter-sentence association: if a WBS model is designed to store sentences and not words, then the grammatical order of the sentences is kept in the model. Thus, the sentences that have the same word but a different order in the two sentences will be retrieved as if they had the same word. Figure 6.18 shows an example of sentences with the same words but a different order. As expected, the same sentences will have a higher number of retentions than the same words in the same order.\n\nThe serial position effect can be observed when words that occur in a specific position in the sentences are more retrieved than the same words that occur in different positions. This effect can be attributed to the influence of the semantic context on the recognition of the words that are at the same position in different sentences, and the words that are not in the same position have a different semantic context that makes them not to be recognized as if they were in the same position. Figure 6.19 shows an example of sentences with the same words in the same order but with a different position in each sentence. In this example, the word \u201cof\u201d in the position 5 of the first sentence is more retrieved than the same word in the position 7 of the second sentence, and the word \u201cof\u201d in the position 3 of the second sentence is more retrieved than the same word in the position 6 of the third sentence. This effect is more noticeable when the word \u201cof\u201d is repeated.\n\nIn this section, we present the results of a search on the Corpus of Contemporary American English (COCA) [44] with the retrieved words and their respective sentences in a sentence. For each sentence, the average of the values of the retrieval frequencies for all the words in the sentence is computed. We show in Figure 6.20 the relative frequencies of the average values for each word-type for all the sentences of the corpus.]" time="1.377"><properties><property name="score" value="0.9997808337211609" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[In two years, after graduating with a master's in fine arts, he would do what he was paid to do: act in a studio's movies.\n\nBut then in the late 1960's, &quot;you couldn't see any kind of independent films,&quot; he said. &quot;We had two major studios, MGM and RKO. The independents didn't get a chance.&quot;\n\nSo he decided to make one of his own. After asking students at the graduate film program of the New York University Film School for help, he gathered friends and shot his first film, &quot;Mr. Mom,&quot; about a laid-off advertising executive who becomes a housewife, in 1970.\n\nThe low-budget, independent film, directed by a woman and starring a cast of talented newcomers, was ahead of its time. Many viewers laughed, he said, at the notion of a housewife as a career woman. Others did not understand the film's subtle humor.\n\nPhoto\n\n&quot;The primary thing people didn't understand was it wasn't supposed to be funny,&quot; he said. &quot;It was dramatic.&quot;\n\nThe film cost $25,000, and he sold it to Warner Brothers for $50,000. &quot;We got drunk for two weeks,&quot; he said. &quot;It was our first film. We were elated.&quot;\n\nHe went on to make six more films as a director and actor. But the money he made was eaten up by the next movie, as well as his divorce, children and a country house in New Hampshire, he said.\n\nNewsletter Sign Up Continue reading the main story Please verify you're not a robot by clicking the box. Invalid email address. Please re-enter. You must select a newsletter to subscribe to. Sign Up You will receive emails containing news content , updates and promotions from The New York Times. You may opt-out at any time. You agree to receive occasional updates and special offers for The New York Times's products and services. Thank you for subscribing. An error has occurred. Please try again later. View all New York Times newsletters.\n\nThen he met Nancy Chase, a writer and former actress who would become his second wife. She had grown up in a suburb of New York City and attended St. Agnes School in the city's West Village and a private high school.\n\nAdvertisement Continue reading the main story\n\nMs. Chase knew nothing about film, he said. But she was interested in Mr. Moore's films, especially &quot;Slackness,&quot; which came out in 1981. She asked if he would like to work on a screenplay she was writing. He agreed.\n\n&quot;I thought this was a possibility to get out of the mess I was in,&quot; he said. &quot;It was a partnership. I had a chance to work on a screenplay with her. And she had a chance to write something. And then we got married.&quot;\n\nThey met in 1983. A year later they married and went on to make several more films.\n\nMr. Moore said that he and Ms. Chase had a very &quot;civilized&quot; divorce after 23 years. &quot;We've been married long enough that we have two children who are married, who are professionals,&quot; he said. &quot;They're both people who are not into the entertainment business. So they're good children.&quot;\n\nMr. Moore also has two other children. He did not see his first wife, Kay, and their children after they divorced in 1973.\n\n&quot;I didn't find it productive to go back into my son's life,&quot; he said. &quot;I didn't want to be a hero and make up. I didn't want to see him have a second set of delusions about me. I've always believed that children should be given credit for being capable of determining who their parents are. I didn't want to be a subject of people's emotions.&quot;\n\nHe recently remarried, and now lives with his wife, Eileen Minnelli, in a three-story, six-bedroom house that sits behind a wrought-iron gate on a leafy residential street in Mill Valley, Calif.\n\nA self-described &quot;indiscriminate reader,&quot; Mr. Moore spends much of his time writing short stories and screenplays. And he has plans for at least two more films. One is about a woman who was institutionalized in the 1960's, after becoming pregnant. The other is about the son of an interracial couple who is raised by his mother's relatives and becomes a neo-Nazi.\n\n&quot;I have a lot of things to say, and I don't have a problem saying them,&quot; he said. &quot;I haven't changed my approach to making films at all.&quot;]" time="2.338"><properties><property name="score" value="0.9997521787881851" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Jets quarterback Mark Sanchez practiced for the first time this week and appears on track to start Sunday against the Tennessee Titans.\n\nSanchez's right shoulder strain limited him in practice Wednesday, but he took most of the first-team repetitions. He was 15 of 19 passing with no interceptions and no sacks against a second-team defense. He said after practice that he would play against the Titans if the game were today.\n\n&quot;It's great, it's great,&quot; Sanchez said of his arm strength. &quot;The most important thing with an injury is to get the strength back. We're seeing good results of that. ... That's the main thing.&quot;\n\nThe Jets quarterback was injured in the second quarter of Sunday's 24-17 victory over the Baltimore Ravens, and rookie Greg McElroy came off the bench and played well in relief. McElroy completed 13 of 19 passes for 153 yards and a touchdown and was sacked twice.\n\nSanchez said it's not hard to prepare to play in his first game since the injury.\n\n&quot;This is the first week I'm really getting to throw a lot,&quot; he said. &quot;I've got to get back into a rhythm. That's what it's about. ... I'm excited. I'm ready to go.&quot;\n\nThe Jets have to make a decision on their starting quarterback by Tuesday, when players return from their four-day Thanksgiving break. Coach Rex Ryan said it's not a tough call at this point.\n\n&quot;We'll see,&quot; Ryan said. &quot;We've got another day of practice. That's what we've got to look at and see how the situation unfolds.\n\n&quot;If you just look at where he is at this point, he looks good. You look at what Greg did, too. You look at what we did against a good team with a lot of different players.&quot;\n\nRyan said the decision to start Sanchez or McElroy has &quot;absolutely nothing to do with any kind of controversy or anything like that. We've got a good quarterback. We know we have two good quarterbacks. You guys want to make this some big controversy, but that's not the case.&quot;\n\nMcElroy said he is excited for Sanchez and happy for himself, as well.\n\n&quot;When you have two quarterbacks, you're going to get one of them going,&quot; he said. &quot;I'm excited to see Mark back. It's a good thing.&quot;]" time="1.255"><properties><property name="score" value="0.9524917006492615" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Sir Robert Hart, 1st Baronet\n\nSir Robert Hart, 1st Baronet (22 November 1823 \u2013 1 September 1905) was a British diplomat who worked as chief inspector of Chinese Maritime Customs.\n\nHart was born in Fermoy, County Cork, Ireland. His father was a physician, and his mother was the daughter of another doctor. He was educated at Fermoy Grammar School.\n\nHart's father died when he was twelve, leaving the family in financial difficulties. Hart was forced to earn a living and so took a job as a clerk in the Civil Service of the East India Company at a salary of \xa3100 per year. He became fluent in Urdu and Hindi. After six years, he left the service of the East India Company and moved to China.\n\nHe joined the Chinese Imperial Maritime Customs in 1851. He worked at Amoy in Xiamen and at Shanghai. He worked at Fuzhou and Hong Kong. He spent some time in Macau before moving to Canton in 1859. In 1863, he was promoted to Inspector General. Hart was promoted again in 1871 to Chief Inspector, and became Inspector General in 1882. In that position, he was responsible for administering and enforcing the Chinese Maritime Customs' rules and regulations and for conducting negotiations with other governments. He negotiated many new treaties with foreign powers. He retired from that position in 1897. He died in London in 1905.\n\nThe famous novel &quot;Lost Horizon&quot; by James Hilton has a British consul who is modeled after Hart.\n\nHart was created a Baronet, of Claremont in the County of Cork, on 10 August 1899. He received several honours and decorations from foreign governments, including the Grand Cross of the Order of the Bath from the British government, the Grand Cross of the Order of Leopold and the Legion of Honour from the Belgian government, and the First Class Order of the Double Dragon from the Chinese government.\n\nOn his death, he was succeeded as baronet by his eldest son Sir Ronald Hugh Stewart Hart, 2nd Baronet.\n\n\n]" time="1.018"><properties><property name="score" value="0.999670535326004" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Chris Van Gorder\n\nChris Van Gorder (born March 16, 1955) is the former head coach for the women's basketball team at the University of Cincinnati. Van Gorder coached at Ohio University from 1985 to 1996, where he posted a 204-163 record.\n\nVan Gorder was the head coach at the University of Akron for ten seasons. During his tenure, he compiled a record of 157-108. Van Gorder was the first head coach of the Akron Zips women's basketball team.\n\nIn 2005, he was named Mid-American Conference Coach of the Year after his Zips finished 22-9 and went to the NCAA Tournament.\n\nIn the 2006-2007 season, the Zips started the season ranked 12th nationally but dropped out of the rankings following a 65-67 home loss to Western Carolina. Despite the loss to Western Carolina, Van Gorder's Zips won the Mid-American Conference regular season title and the Mid-American Conference Tournament. Van Gorder's Zips then lost to Tennessee in the first round of the NCAA Tournament.\n\nIn the 2007-2008 season, Van Gorder's Zips were not ranked to start the season. But after two early victories against ranked opponents, Van Gorder's Zips were ranked 25th in the nation. Van Gorder's Zips remained in the top 25 for a majority of the season and won the Mid-American Conference regular season championship for a second consecutive year. After a first-round loss in the NCAA Tournament, Van Gorder was named Mid-American Conference Coach of the Year for the second time in three seasons.\n\nVan Gorder was let go at the end of the 2012 season and hired by the University of Cincinnati on April 5, 2013. He resigned on May 4, 2015.\n\nVan Gorder is married to the former Deborah Beisecker, a native of Knoxville, Tennessee. The couple has a daughter, Jennifer, and two sons, Cody and Collin.\n]" time="0.956"><properties><property name="score" value="0.9996168911457062" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Streptomyces durhamensis\n\nStreptomyces durhamensis is a bacterium species from the genus of &quot;Streptomyces&quot;. &quot;Streptomyces durhamensis&quot; produces shikimic acid, paclitaxel, antimicrobial peptides and tyrosinase.\n\n\n\n]" time="0.261"><properties><property name="score" value="0.8835775852203369" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Ammi majus\n\nAmmi majus, common names greater ammi, gousle, onion grass, greater ammi, cumin-vine, is a plant belonging to the family Apiaceae. It is endemic to Central America, South America, Mexico, West Indies, and Florida in the United States. It is cultivated as a vegetable and spice in other parts of the world.\n\n&quot;Ammi majus&quot; is an annual or biennial plant that reaches heights of and has weak stems. The leaves are thin, simple, and alternate. The flowers are white or pale yellow and borne in clusters of 2 to 4. The fruits are cylindrical capsules with five angles.\n\n&quot;Ammi majus&quot; is used as a vegetable in the Caribbean, where it is known as &quot;gousle&quot;, and in Southeast Asia. It has been found to contain goitrogenic compounds which may be problematic for people with thyroid disease. The whole plant can be consumed, but is most often eaten only the tender leaves.\n]" time="0.479"><properties><property name="score" value="0.5892606377601624" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.58926064&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.58926064
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[2010 Spanish Grand Prix\n\nThe 2010 Spanish Grand Prix (formally the XLV Gran Premio de Espa\xf1a Telef\xf3nica) was a Formula One motor race held on 28 May 2010 at the Circuit de Catalunya in Barcelona, Spain. It was the seventh round of the 2010 Formula One season. The 60-lap race was won by Red Bull driver Mark Webber, after he started from second position. Lewis Hamilton finished second for the McLaren team and Ferrari driver Fernando Alonso came in third.\n\nPirelli's prime tyre compound was the hard &quot;primes&quot;, whilst the option tyre was the soft &quot;option&quot; compound.\n\nMark Webber took pole position for Red Bull, his first since the 2009 Belgian Grand Prix, while Fernando Alonso set the fastest time in Q2 and qualified second. The two Ferraris filled the next row with Felipe Massa ahead of Kimi R\xe4ikk\xf6nen. The two McLarens of Lewis Hamilton and Jenson Button were fifth and sixth, with Nico Rosberg in seventh place, having struggled all weekend with the handling of his Mercedes. Adrian Sutil in the Force India was eighth, followed by Sebastian Vettel in the second Red Bull and Nico H\xfclkenberg in the Williams. Paul di Resta, S\xe9bastien Buemi and Michael Schumacher rounded out the top ten. At the start, Mark Webber led away from Fernando Alonso and Felipe Massa while the two McLarens of Hamilton and Jenson Button slipped backwards, with Hamilton eventually being forced to pit with an electrical problem, and Button dropping to eleventh. Mark Webber's race soon went wrong as he came into the pit lane to switch to his wet weather tyres on lap 11, but was told to remain in his pit box as he had not switched off his engine, having forgotten to do so. This meant he was stationary for 23 seconds and dropped to fourth, before recovering to third by the time he rejoined the circuit. He also had to serve a drive-through penalty for passing Alonso on the grass, although he eventually finished fourth.\n\nAlonso and Massa pitted on laps 18 and 22 respectively, rejoining in second and third places behind the McLaren of Lewis Hamilton, who had made an early pit stop for dry tyres and was running a different strategy to the two Ferraris. With the track still wet, Hamilton, Massa and Alonso all switched to the intermediate tyres on lap 25, but at this point, the gap between the top three was only 1.3 seconds. The McLarens of Hamilton and Button were then unable to pass the Ferraris on the drying track and they began to slip backwards, eventually losing their fourth and fifth places to the two Force Indias of Adrian Sutil and Paul di Resta. The latter had passed Sutil on the main straight on lap 30 after the Force India driver had made a pit stop and had a clear track in front of him, and held on to fifth place despite being passed by Sutil a few laps later. The two Toro Rossos of Jaime Alguersuari and S\xe9bastien Buemi were the first of the lead lap cars to pit on lap 28, with Alguersuari dropping behind Rosberg in the standings after pitting earlier than the Mercedes driver. The other Toro Rosso of Sebastien Buemi managed to stay out longer and maintained his eighth place, before making a pit stop at the end of lap 34.\n\nFurther back, Jenson Button was the first of the two McLarens to pit for new tyres on lap 33, with a nine-second stop due to an electronics problem with the wheel gun which meant the front left tyre was not fitted correctly. Team-mate Lewis Hamilton followed suit a lap later and rejoined the race in twelfth place. Button rejoined the race behind S\xe9bastien Buemi and passed him on lap 36, but was behind Adrian Sutil when he stopped for fresh tyres on lap 38, although Button was able to pass Sutil a few laps later, with Buemi also getting ahead of the Force India. At the end of the 43rd lap, Lewis Hamilton had managed to pass Paul di Resta and move up to eighth place, with Sutil falling to ninth, having been passed by both Buemi and Button. Hamilton's battle with Buemi came to an end on lap 44 as the two McLarens pitted together, with Hamilton getting ahead]" time="2.201"><properties><property name="score" value="0.432895764708519" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Carthage House is a historic home in Deland, Florida, United States. It is located at 709 South Catherine Street. On July 26, 1990, it was added to the U.S. National Register of Historic Places.\n\n]" time="0.187"><properties><property name="score" value="0.8760352730751038" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[S\u0103v\u0103disla\n\nS\u0103v\u0103disla () is a commune in Bihor County, northwestern Romania with a population of 2,099 people. It is composed of four villages: Aluni\u0219u (&quot;Szalonyos&quot;), Ciutura (&quot;Sztutura&quot;), S\u0103v\u0103disla and Socet (&quot;Szoket&quot;).\n]" time="0.224"><properties><property name="score" value="0.9472625851631165" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.94726259&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.94726259
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Agelasta cognata\n\nAgelasta cognata is a species of beetle in the family Cerambycidae. It was described by Stephan von Breuning in 1935. It is known from the Philippines.\n]" time="0.152"><properties><property name="score" value="0.9710140824317932" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The Don's Analyst\n\nThe Don's Analyst (, , ) is a 1968 Italian-French comedy film directed by Sergio Corbucci.\n\nDon Salvatore's sister wants to marry Don Vincenzo's son, which makes the two crime families form a union. But Don Salvatore's son is in love with the Don's brother-in-law's daughter, which makes the two families' alliance a difficult one.\n\n\n]" time="0.226"><properties><property name="score" value="0.9997363686561584" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Kurnatowice\n\nKurnatowice is a village in the administrative district of Gmina Lipsko, within Lipsko County, Masovian Voivodeship, in east-central Poland. It lies approximately south-west of Lipsko and south of Warsaw.\n\n&lt;br&gt;\n]" time="0.171"><properties><property name="score" value="0.6286781430244446" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Jane's All the World's Aircraft\n\nJane's All the World's Aircraft is an annual publication on the subject of civil and military aircraft.\n\nIt was originally produced by Charles and John Whitley in 1909. Since 1962, it has been published by Jane's Information Group. It is published in English and French editions.\n\nAt the time of writing (June 2018), the 2018 edition is in its 75th year of publication.\n\n]" time="0.232"><properties><property name="score" value="0.9996095299720764" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Eucosma plurivittella\n\nEucosma plurivittella is a species of moth of the Tortricidae family. It is found in Russia (Amur, Ussuri, Sakhalin, Primorye, Khabarovsk, the Kuriles), China (Heilongjiang, Beijing, Jilin, Liaoning, Zhejiang, Hunan, Yunnan), Korea, Japan, Mongolia and Europe, where it has been recorded from Spain, Germany, Austria, the Czech Republic, Hungary, Slovakia, Poland, Latvia, Finland, Ukraine and Romania. The habitat consists of moist meadows.\n\nThe wingspan is 12\u201317\xa0mm. Adults are on wing from May to August.\n\nThe larvae feed on &quot;Rumex japonicus&quot; and &quot;Polygonum&quot; species.\n\n]" time="0.399"><properties><property name="score" value="0.941705048084259" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.94170505&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.94170505
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Cluster switching\n\nCluster switching is a migration strategy for cloud computing, in which the users or services in a running system are transparently moved to a new cluster without disrupting their services or performance. It was first introduced in a paper presented at the 2004 IEEE International Conference on Cluster Computing (CLUSTER 2004) by the same name, and published in Cluster Computing in 2006.\n\nThe purpose of cluster switching is to maintain system services even when a node or the entire cluster fails. Cluster switching is often used in parallel or distributed systems where the nodes are not permanently connected to each other. In this case, it is possible to have one node fail without impacting the others. The challenge with cluster switching is that it requires the system to transparently move the workloads to other nodes without affecting their performance.\n\n]" time="0.397"><properties><property name="score" value="0.7680217027664185" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.7680217&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.7680217
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[2015\u201316 UEFA Women's Champions League\n\nThe 2015\u201316 UEFA Women's Champions League was the 12th edition of the European women's championship for football clubs. The final was played between Frankfurt of Germany and Lyon of France, which Frankfurt won on a penalty shoot-out following a 1\u20131 draw. The win was the first time a German team had won the competition.\n\nGerman teams won both the main round groups, and Frankfurt also won the play-off round group, while German and French teams dominated the knockout phase. Frankfurt defeated Arsenal 3\u20131 in the final, while Paris Saint-Germain defeated Barcelona 3\u20132 in the match for third place.\n\nStarting from this season, the following changes were introduced:\n\nThe allocation of teams into qualifying and qualification is as follows:\n\nUEFA has scheduled the competition as follows (all draws are held at the UEFA headquarters in Nyon, Switzerland).\n\nThe schedule of the competition is as follows (all draws are held at the UEFA headquarters in Nyon, Switzerland).\n\nThe draw for the qualifying round was held on 22 June 2015. The first legs were played on 27 and 28 August, and the second legs were played on 3 and 4 September 2015.\n\nThe first legs were played on 17 and 18 September, and the second legs were played on 24 and 25 September 2015.\n\nThe first legs were played on 1 and 2 October, and the second legs were played on 22 and 23 October 2015.\n\nThe first leg matches were played on 6 and 7 November, and the second legs were played on 20 November 2015.\n\nThe first leg matches were played on 1 and 2 December, and the second legs were played on 13 December 2015.\n\nThe first leg matches were played on 17 and 18 January 2016, and the second legs were played on 24 and 25 January 2016.\n\nThe draw for the round of 16 was held on 27 January 2016. The first legs were played on 7 and 8 March, and the second legs were played on 14 and 15 March 2016.\n\nThe first leg matches were played on 7 and 8 April, and the second legs were played on 14 and 15 April 2016.\n\nThe first leg matches were played on 12 May, and the second legs were played on 19 May 2016.\n\nThe first leg matches were played on 25 and 26 September, and the second legs were played on 2 and 3 October 2016.\n\nThe final was played on 15 May 2016 at the Stadion im Borussia-Park, M\xf6nchengladbach, Germany.\n]" time="1.398"><properties><property name="score" value="0.9989054600397745" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Kawanishi N1K\n\nThe Kawanishi N1K &quot;Ky\u014df\u016b&quot; (\u5f37\u98a8, &quot;Strong Wind&quot;) was a long-range single-seat carrier fighter aircraft, serving the Imperial Japanese Navy Air Service during World War II. It was a single-engined, single-seat monoplane fighter designed to replace the A6M Zero, which was designed for shorter ranges.\n\nThe N1K was developed to replace the Nakajima A6M Zero in the long-range fighter role, and was the first fighter in the world to use the Type 3 &quot;Hakury\u016b&quot; 18-cylinder air-cooled radial engine, which was also used in the Kawanishi N1K1-J &quot;Shiden&quot; fighter. The Japanese Navy also gave it the name &quot;Kyofu&quot; (meaning &quot;Strong Wind&quot;) as the manufacturer's symbol, the first time a name was given to a Japanese military aircraft.\n\nIn the end, only a handful of N1K1 fighters (armed with two 20\xa0mm cannon and two 12.7\xa0mm machine guns) were deployed operationally before the end of the war, and none ever engaged enemy aircraft in combat. After World War II, the Allied Occupation forces in Japan took possession of six N1K1 fighters and four N1K2 fighters. These aircraft were studied and flight tested by the Technical Air Intelligence Unit (TAIU) at Tachikawa. Test flights and subsequent analysis revealed that the N1K2 was the faster of the two types, but its lighter armament was deemed inadequate for combat, and the aircraft was regarded as lacking in manoeuvrability.\n\nFollowing the war, some of the N1K2 aircraft were provided to the USSR, and some were used by the &quot;Taiwanese Air Force&quot; (ROCAF), until all remaining aircraft were destroyed by a typhoon in 1950.\n\n\n\nThe only two remaining airworthy N1K2-J Shiden, on display in Japan, are preserved at the following museums:\n\nThe only remaining N1K1-J is on display at the Planes of Fame Air Museum in Chino, California. The airframe was recovered from the ocean in 1994 by Paul Allen and Don Arndt, restored to static display configuration and placed on display at the museum in 2003. The N1K1-J was removed from display in late 2013 for a full restoration, including flight testing, and returned to display on March 10, 2017.\n\nThe Soviet Union took possession of several of the N1K2-Ja variants and used them as the prototypes for the development of the Mikoyan-Gurevich MiG-15.\n\nTwo Shiden's survive in Taiwan. They are located in the Republic of China Air Force Museum in Taipei.\n\n\n]" time="1.434"><properties><property name="score" value="0.990880290667216" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[BLACK SABBATH star Tony Iommi says that he has &quot;no plans&quot; to reunite with Ozzy Osbourne for another album or tour following the release of the group's new album, &quot;13&quot;, in June (13).\n\nThe metal legends are about to launch their final world tour after performing in Latin America this month, but while fans can expect to see Osbourne and Geezer Butler at the shows, Iommi reveals that the band will not attempt to bring bassist Bill Ward back into the fold.\n\nHe tells Rolling Stone magazine, &quot;There's no plans for another album or tour with (Ward). And the bottom line is that I wish him well, but I just can't go back. It's been too long. I just can't do it again.&quot;\n\nHe continues, &quot;I'm not going to say that I'll never work with (Ward) again, but I'm not going to try to work with him again for the sake of the fans. It wouldn't be right. It wouldn't be fair.&quot;\n\nIommi also admits that he is baffled by the criticism Osbourne received for his behaviour during the band's Rock And Roll Hall Of Fame induction in April (13).\n\nHe says, &quot;I know he's had problems over the years with drink and that, but I don't really know. All I know is what he did at the Hall Of Fame, which was absolutely incredible, because it was one of those rare moments when you see somebody and you know they're completely sober. It was really good.&quot;\n\nHe adds, &quot;I really couldn't understand the criticism. But then again, I'm pretty thick-skinned about stuff like that. I can take it.&quot;]" time="0.891"><properties><property name="score" value="0.9975299537181854" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[&quot;When you have a lot of money and power, sometimes you think you can buy yourself out of anything.&quot; \u2015Ben Urich [src]\n\nJohn Bushmaster was a powerful crime boss and drug dealer, having inherited control of Harlem's Paradise after his brother, Quincy, was shot and killed by Luke Cage. Bushmaster was also the head of the ruthless Bushmaster drug cartel.\n\nContents show]\n\nBiography\n\nEarly Life\n\nNot much is known about John Bushmaster's early life, except that he was once involved in organized crime. He became a drug lord and took over his brother's position as head of Harlem's Paradise. He later became involved in the criminal organization known as the Hand.[1]\n\nLuke Cage's Power\n\n&quot;Tell me where they took Misty Knight.&quot;\n\n&quot;What are you going to do with me?&quot;\n\n&quot;Well, I'm not going to let you walk away.&quot;\n\n&quot;Who's going to stop me?&quot;\n\n&quot;Luke Cage.&quot; \u2015Luke Cage and Bushmaster [src]\n\nWhen Bushmaster witnessed Luke Cage throw a man out of Harlem's Paradise for harassing some young women, he asked his brother, Quincy, to help him get rid of the unruly patron. The next day, Bushmaster called the man who was threatening to sue Harlem's Paradise and threatened to kill him. When Cage returned, Bushmaster tried to persuade Cage to leave by telling him he did not belong there. Bushmaster revealed to Cage that his name was John Bushmaster and that he was the owner of Harlem's Paradise. When Cage insisted that he would be staying, Bushmaster noted that he did not know who he was messing with before Cage told him he was messing with Luke Cage.\n\nCage remained in Harlem's Paradise and caused problems for Bushmaster, much to his annoyance. Bushmaster and his men began selling the Bushmaster drug throughout Harlem. Cage overheard Bushmaster's name when it was said by his brother Quincy during a phone call. When Cage visited Bushmaster at Harlem's Paradise and accused him of being behind the Bushmaster drug, he denied it and told Cage to leave. Bushmaster's men, however, confronted Cage and attacked him. Cage easily overpowered all of them and escaped from Harlem's Paradise.[1]\n\nChasing Misty Knight\n\n&quot;I'm getting sick of your bullshit.&quot;\n\n&quot;Is that what you tell all the girls?&quot; \u2015Bushmaster and Misty Knight [src]\n\nWhen Cage attacked one of Bushmaster's drug warehouses and threatened to kill Bushmaster if he did not tell him where he was keeping Misty Knight, Bushmaster told him that he did not know what he was talking about and Cage attacked him. Bushmaster ordered his men to shoot Cage while he escaped. The bullets did nothing to Cage as they could not penetrate his skin.\n\nBushmaster visited Pop's Barber Shop and threatened to burn down Pop's Barber Shop and kill everyone inside if he did not get a haircut, including the young girl named Tilda Johnson, who could do a mean cornrows, before he left. Tilda snuck out of the shop and stole Bushmaster's car while he was inside getting a shave. Bushmaster ran outside and found Tilda attempting to get away. He then chased after Tilda in the stolen car.\n\nTilda was able to evade Bushmaster's pursuit until he was able to push her car off the road and caused her car to roll over and crash. Bushmaster dragged Tilda from her car and found out who she was. Tilda, however, was rescued by a bystander who had witnessed the entire incident and called the police. Bushmaster was arrested for kidnapping, but he was released from police custody later.[2]\n\nBusiness Proposition\n\n&quot;Now, we are making money, yes, but not as much as we should. It's a competitive market, man. You're good. And that girl, well, she's something else. I'm offering you a piece of this action. 25% for the both of you.&quot;\n\n&quot;That's not how this works.&quot;\n\n&quot;You know what? It is how this works. It's time for you to grow up.&quot; \u2015Bushmaster and Shades [src]\n\nBack at Harlem's Paradise, Bushmaster offered Shades a percentage of his drug business, but Shades refused. Bushmaster made an offer to Diamondback to sell him some of his product in exchange for a fee. Bushmaster met with Shades, who berated him for not being a better businessman and selling to Diamondback. Shades told Bushmaster that he was selling drugs on his own, but Bushmaster told him that he had a plan to sell his drugs in Bushmaster branded bullets. Bushmaster then attempted to seduce Shades' wife, Tilda Johnson, but she refused his advances.\n\nThe next day, Bushmaster gave a speech to the patrons of Harlem's Paradise as they celebrated the end of their evening. Bushmaster boasted about his power and demanded that they bow to him, but was met with little response. Shades then insulted Bushmaster, who then drew his gun on Shades. Luke Cage then attacked Bushmaster, but Bushmaster managed to get the upper hand and beat Cage into submission. As Bushmaster prepared to shoot Cage, Tilda Johnson shot Bushmaster in the chest with a concealed gun and stole his gun, with Shades shooting the other members of Bushmaster's entourage before they could react.\n\nBushmaster and his men went into hiding at a warehouse, where Bushmaster furiously scolded Shades for his betrayal. Shades explained that he had been working with Diamondback, but now wanted to make a better life for himself and his wife. Bushmaster, realizing that Shades had a point, agreed to go into hiding for the night. The next day, Bushmaster was visited by Knight, Shades, and Cage. Knight had returned to take back Harlem's Paradise. Knight gave Bushmaster a chance to surrender, but Bushmaster refused and attempted to shoot Knight, but was overpowered by Cage and Knight, forcing him to surrender.\n\nWhile Knight and Cage prepared to transport Bushmaster to Ryker's Island, Bushmaster told them that the shipment of drugs would be arriving that night. The police tried to stop the convoy, but the SWAT team was gunned down by Bushmaster's men. Knight decided to attack the warehouse to stop the shipment from getting to New York. Knight, Cage, and Bushmaster fought his men while Tilda Johnson fought Diamondback and his men, although Diamondback managed to escape from the scene.\n\nOnce Knight and Cage subdued the rest of Bushmaster's men, Knight interrogated Bushmaster about where the shipment was. Bushmaster told them that the guns had already been sold to a man named Ricky. Knight and Cage traveled to a diner to look for Ricky, where they found Ricky talking to his partner Rico. Knight threatened Ricky into giving her the location of the shipment. Once Ricky told Knight the location, she and Cage drove to the location where the guns were being shipped. Knight was able to steal the guns from the men who were shipping them, but one of Bushmaster's men threw a grenade into the truck. Cage managed to catch the grenade and toss it back at Bushmaster's men, but was then hit by the truck, knocking him unconscious. Knight ran Cage to the hospital while Bushmaster's men reclaimed their guns and returned them to Bushmaster.[3]\n\nPersonality\n\nTo be added\n\nAbilities\n\nExpert Combatant : Bushmaster was an expert hand-to-hand combatant, as he was able to easily fight off Luke Cage with little effort.\n\n: Bushmaster was an expert hand-to-hand combatant, as he was able to easily fight off Luke Cage with little effort. Skilled Leader: Bushmaster was a skilled leader, as he was able to rally his men together in order to escape Harlem's Paradise.\n\nEquipment\n\nWeapons\n\n&quot;9-milimeter. Take the cap off, and you got a little pipe bomb right here.&quot; \u2015Bushmaster to Tilda Johnson [src]\n\n9mm Pistol: Bushmaster often used this gun when threatening people. Tilda Johnson]" time="4.075"><properties><property name="score" value="0.2838244214653969" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[It's been more than a decade since the last major update to Britain's copyright law, and it's high time that the UK brought its copyright law into the 21st century.\n\nYou've probably heard that the Government wants to bring in a new Digital Economy Act, which includes a measure to require all ISP customers to opt out of a default-on filter for adult content. Here's how you can tell them that you're not happy with this proposal, using just your email address.\n\nWho?\n\nIf you're concerned about online porn, you might have already signed a petition on the government's website calling for the default-on filter to be dropped. And in the wake of the controversy around the DEA, you might also have signed a petition demanding that ISPs should automatically block sites known to be involved in the distribution of online child abuse imagery. (Both of those petitions are still open for signatures if you want to join in).\n\nBut now it's time to tell the Government about some of the other things they're doing that are not as sensible, and we'll help you do it.\n\nWhat?\n\nFor a long time now, the UK Government has wanted to bring in a new Digital Economy Act. Last summer, the new law was finally published. The Act is massive, and there's lots to argue with, but it looks like the Government may be preparing to rush it through Parliament without any further debate. That means there's a limited window of opportunity to discuss changes to the draft before it becomes an Act of Parliament, and the Government starts work on enforcing it.\n\nThe draft DEA proposes a number of controversial new laws, including the opt-out default filter that's being discussed so widely. But the Act also proposes lots of other worrying measures, and gives the Government a lot more power to interfere with our online communications. The Act's supporters will tell you that it will help to tackle online infringement, for example. But its many critics argue that it risks stifling innovation in online services, and giving the Government a free hand to censor the internet in other ways.\n\nOn Monday, we told you that you could email your MP to demand a debate on the DEA. Now we've got even more reason to tell the Government that we're not happy with the draft DEA.\n\nThis week, the Pirate Party UK submitted a response to the Department of Business Innovation and Skills (BIS) on the Digital Economy Act, which responds to the Government's consultation on the proposals. We're not the only ones with concerns about the proposals; the Open Rights Group has submitted its own response (pdf) to the Government's consultation, and Index on Censorship has produced a response (pdf).\n\nHere's why we think the Government's proposals are a Bad Thing:\n\nThe proposals for an &quot;Extended Liability for Platforms&quot; could have serious unintended consequences for small internet service providers and independent content sharing websites, making them responsible for policing their services and preventing copyright infringement. We think this is likely to stifle innovation in new services, and also has serious implications for our online privacy.\n\nThe plans for an automatic filtering system to block content that is not classified by the state could lead to a drastic over-blocking of content. It could block more than just &quot;extreme pornography&quot; - it could block access to anything the Government decides is inappropriate. There's no requirement for these filters to be accurate or transparent, so we won't even be able to tell what's been blocked.\n\nUnder the proposals, the Government would also have the power to block websites that it claims are &quot;blatantly&quot; or &quot;unlawfully&quot; infringing copyright. This could be a serious threat to online innovation, and to our online privacy. And again, there's no requirement for these blocks to be accurate or transparent.\n\nThe DEA would make it much easier for the Government to put in place more blocking orders, and extend their scope. The wording of the legislation is vague enough that it could be used to order ISPs to censor content that isn't even illegal. The Government has already said that it wants to use its new powers to stop us from accessing websites that might be used to share copyrighted material.\n\nThe DEA is the biggest threat to our online privacy since the Data Retention and Investigatory Powers Act. If the Government gets its way, then they'll have the power to block websites and track what we're doing online. And they'll be able to do it without even telling us that it's happening.\n\nYou can read the Pirate Party's full response to the DEA consultation here (pdf).\n\nNow we need your help\n\nIf you're concerned about the DEA, you can join us in telling the Government what you think about the proposed Act.\n\nThe Open Rights Group's response (pdf) to the DEA consultation is still open for signatures. Once they've got enough names on the petition, they'll send their response to the Government.\n\nAnd now we've got our own response to the Government's consultation on the Digital Economy Act. Please support our campaign to stop the Act by signing our response here (pdf).\n\nWe'll be sending our response to the Government later this week, and we hope that it will give people the opportunity to stop the Act before it's too late.\n\nThis is the one chance we have to stop the DEA. But we need your help. Please support our campaign to stop the Act.]" time="2.644"><properties><property name="score" value="0.9522371490796407" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[BevMo Beverage Superstore in California is a chain of beverage and snack retail stores. Its headquarters are located in San Carlos, California. As of May 2015, it has 41 stores in California and Nevada. All the stores offer beer, wine and liquors. They also carry various snacks, health and beauty products, household items and automotive supplies.\n\nStore Manager and Assistant Manager jobs in BevMo are offered in different parts of the state. The company also offers Benefits for its employees and Manager Trainee Program.\n\nGet more information about Manager Trainee Program at http://www.indeed.com/cmp/BevMo/jobs\n\nBevMo Beverage Superstore provides competitive pay, benefits, and other programs to help develop employees' potential and achieve their career goals.\n\nApply today!\n\nIt's fast and easy to apply online for jobs in BevMo Beverage Superstore, at http://www.indeed.com/apply?q=beverage+superstore\n\nFrequently Asked Questions\n\nCan I find the career page for BevMo Beverage Superstore on LinkedIn?\n\nNo, the company doesn't have a LinkedIn page. However, you can find more information about the company at its official website: http://www.beverage.com/About-BevMo/Careers\n\nCan I connect with BevMo on Twitter?\n\nNo, the company doesn't have a Twitter page. You can follow them on Facebook at https://www.facebook.com/Bevmo/\n\nWhat is the number for the customer service center?\n\nBevMo doesn't have a customer service center. However, you can contact the company at 1 (877) 822-6692.\n\nIf you have other questions, comments, or concerns, please leave them in the comments section below.\n\nBest Regards,\n\nIndeed.com]" time="0.904"><properties><property name="score" value="0.16468289494514465" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The Indigenization policy announced by Zimbabwe\u2019s new president, Emmerson Mnangagwa, in 2017, has raised a lot of questions about the process of nationalisation and the methods that will be used to allow for equal distribution of capital, land, and business interests. It\u2019s a policy that seems to favour the government at the expense of foreign interests and brings into question whether or not the government has a commitment to fighting corruption and seeking justice.\n\nThe new policy has been met with much scrutiny both in Zimbabwe and internationally. Many have expressed concern that the implementation of the policy will be flawed and that it will ultimately lead to a worsening of the economic situation in Zimbabwe. The Zimbabwean opposition party, Movement for Democratic Change (MDC), has expressed its disapproval of the policy, stating that it will further marginalise Zimbabweans who aren\u2019t involved in the government or who support the opposition.\n\nOn the international level, the Indigenisation policy is one of the latest changes to the Zimbabwean government that have raised concern. Since he took office in November 2017, President Mnangagwa has brought the central bank and the national oil company under government control. This has led to further speculation about the political and economic direction that the government intends to take.\n\nIs the policy legal?\n\nThe Indigenisation policy was initially announced in 2010 and later replaced by another policy called the Zimbabwe Agenda for Sustainable Socio-Economic Transformation (Zim-Asset). The Zim-Asset policy, which was issued in 2013, introduced a set of guidelines for the process of indigenisation. It requires that 51% of all Zimbabwean businesses be transferred to the country\u2019s black majority. However, the policy has never been enforced.\n\nThe decision to revive the policy in 2017 has been widely criticised. Human Rights Watch (HRW) issued a statement urging Mnangagwa to repeal the policy. HRW believes that the policy will only exacerbate the poor economic situation in Zimbabwe. The statement also notes that the government has never properly implemented the policy. It states that \u201cthe fact that no businesses have been indigenized in the past six years suggests that it is unworkable, and that it will not contribute to economic development or job creation.\u201d\n\nAs a member of the African Union, Zimbabwe is bound to uphold the African Charter on Human and People\u2019s Rights, which contains specific provisions regarding the protection of private property. These provisions state that the Charter guarantees the rights to private property ownership and the right to participate in the management of natural resources]" time="1.320"><properties><property name="score" value="0.9998134970664978" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[We believe that these arrangements will provide the school with sufficient financial and resource security to deliver a quality and innovative curriculum for our children and young people.\n\nThis has been a difficult decision to make. The Governing Body has been unanimous in its support for the proposal.]" time="0.161"><properties><property name="score" value="0.9895671010017395" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[W naszym roczniku najwa\u017cniejsze s\u0105 sprawy wzgl\u0119dne, dlatego zdarza si\u0119, \u017ce nie doceniamy swoich przyjaci\xf3\u0142 lub ich nie szanujemy. Je\u015bli jednak czujesz si\u0119 s\u0142abo, koniecznie powiedz, czy nie chcesz, by kto\u015b Ci\u0119 pociesza\u0142, czy te\u017c jak najbardziej. Czy\u017cby to by\u0142a mo\u017cliwa pr\xf3ba?\n\nGdy jest si\u0119 zmartwionym, trzeba wyci\u0105ga\u0107 wnioski. Zdecydowanie lepiej jest przyj\u0105\u0107 kogo\u015b do serca, ni\u017c zwalnia\u0107. Mo\u017cesz te\u017c zaufa\u0107 osobie, kt\xf3ra nigdy nie wyrzuca\u0142a z siebie nic niechcianego. Sp\u0119d\u017acie wi\u0119c czas wolny wraz z rodzin\u0105, przyjaci\xf3\u0142mi lub po prostu samemu. To na pewno b\u0119dzie dobry pomys\u0142.]" time="0.608"><properties><property name="score" value="0.9691646099090576" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.96916461&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.96916461
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Why do I love Saav? It's not because they sent me these for review. These are on the top of my head that I will put on at the first sign of sun when I go out this summer. Yes, they look this good on.\n\nMade from 100% organic cotton, the Organic Star Trunk Brief is the perfect summertime style for your perfect beach body. These will feel like you are not wearing anything. The material is super soft and has a lot of stretch. They will last you a long time and are very well constructed. They come in either blue or black, and are available in Small, Medium and Large. They come in a gift box, so they make a great gift too. You can purchase them from www.saavdesigns.com for $40 a pair.]" time="0.406"><properties><property name="score" value="0.9997830986976624" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[CALCUTTA, March 9\u2014 The daughter of an Indian freedom fighter who died recently in a Calcutta hospital may never find out the cause of her father's death because the hospital has no autopsy facility and the family has no money to send his body to another hospital.\n\nA.B. Ray, a professor of economics at Calcutta University, died last Thursday in Calcutta Medical College Hospital after a long battle with pneumonia.\n\n''I am running from pillar to post to try and get a post mortem,'' said Mrs. Ray, whose husband is also a professor of economics at Calcutta University. ''My father was very much against autopsy and it was his express wish not to have it performed.'' Ray A.B. Ray, 82, was born on the campus of Presidency College, one of the oldest seats of higher education in India. He came from a wealthy and cultured family and had gone abroad to study economics in London.\n\nOn his return he became a prominent figure in Calcutta's intellectual and literary circles, editing newspapers and translating the ''Futurology'' series of lectures by the great French writer Andre Malraux.\n\nHe was an adviser to Indira Gandhi in the early days of her political career. He served in Parliament as a member of the Congress Party until his retirement from active politics in 1976.\n\nHe was admitted to Calcutta Medical College Hospital on Feb. 4, and received treatment in the hospital's intensive care unit. His condition gradually deteriorated and he died last Thursday.\n\nHis daughter, Dr. Suchitra Ray, said she was not at all satisfied with the treatment her father received in the hospital and thought that medical staff had not done their job properly.\n\nShe said that her father had been suffering from pneumonia for three years and that he had been given an injection a few days before he was admitted to the hospital.\n\n''My father was extremely allergic to penicillin,'' she said. ''Injecting him with penicillin at a time when he was in a very bad condition was in my opinion a highly irresponsible act.''\n\nThe doctor in charge of the intensive care unit, Dr. Abhijit Gupta, said that he could not discuss individual cases but that he had had no complaints about the treatment of Mr. Ray.\n\n''We did all we could,'' he said. ''It was very sad. His condition was very grave. It was a heart attack that killed him.'']" time="1.243"><properties><property name="score" value="0.32926323016484577" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[RIGS Creator Platform Playstation 4\n\nPC Genre Shooter, Sports Release Date 29th October 2016 Price $59.99 (PSVR Only) $19.99\n\nRIGS is an action sports game exclusively for PSVR. It was developed by Guerrilla Cambridge (and published by Sony) and released on October 29th, 2016. It takes place in a futuristic sci-fi setting, where you play as one of a number of different characters. Each character uses a different kind of battle suit, allowing them to take part in 3 different sports. Those sports are essentially the 3 main games. They are called RIGS, Megaton, and Gravity Wars. These games play as though they\u2019re real sports. They play more or less how you\u2019d expect them to. You score points by hitting goals, and your objective is to take the ball, and score it in the goal.\n\nThe first game is RIGS, which is essentially a shooting game. There are multiple game modes. Each mode is the same, except you\u2019re playing different characters, and there\u2019s a different goal for you to hit. RIGS is the most fun game in the package, simply because it\u2019s a FPS, which is something you don\u2019t get a lot of in VR games.\n\nMegaton is a sort of bumper-car football game. It\u2019s the second most fun game in the package, just because it\u2019s an interesting spin on something we all know. It is also a lot easier to master than RIGS.\n\nGravity Wars is a very interesting concept, and very hard to master. The game has no physics, and you have to control your direction by using the stick on the right. It\u2019s very challenging, and once you get the hang of it, it can be fun.\n\nRIGS: Mechanized Combat League is a game that is good fun for those who have VR. There\u2019s nothing else like it. It has an interesting concept, and it makes for an interesting single player experience. The multiplayer is there if you want it, but if you don\u2019t, you can use bots, and they will be enough to let you master the game. It does have a lot of potential, and if Guerrilla Cambridge do a sequel, it will most likely be worth it.\n\nPros:\n\nGreat for VR\n\nVariety\n\nWell made\n\nGood concept\n\nCons:]" time="1.259"><properties><property name="score" value="0.9997790455818176" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Phantom ~ Opus 4 ~ Raw\n\n~ Chapter One ~\n\nTitle: Phantom ~ Opus 4 ~ Raw\n\nAuthor: Starfyre\n\nBeta: freddie-m\n\nPairing: Eventual MeixLu\n\nSummary: After 2 years of therapy, Liu is released back into the world and he returns to the only place he calls home. Only a few things have changed. He finds himself adjusting to the changes with more difficulty than he ever expected.\n\nGenre: Angst, Romance, Tragedy, Drama, Supernatural\n\nWarnings: none\n\nWord Count: 15,611\n\nAuthor Notes:\n\nA while ago I read a story called &quot;Bloodlust&quot; by Inu'sLurker. It's a beautifully written story and I recommend you all check it out. I loved it so much I was inspired to write my own story of 2 boys on a trip to Shanghai. Since I am in no way the author that Inu'sLurker is, I apologize for my lack of originality. But I thought that perhaps if I wrote about something that happened to me I would find it easier to write a story like this.\n\nAnd what happened to me was that one year, the day before my birthday, my mother and I went to China. Now that I look back on it, it was a very exciting trip, though at the time I was rather sad because I was turning twelve. I had been dreading the day, and it was nice to be with my mom because she always made me feel better.\n\nIt was also an interesting trip because we were staying with my Grandfather for a few weeks. He lived in Shanghai, and he was the one that took us to China in the first place. So it was an adventure for me to be in a place that he lived. I only saw my Grandfather on holidays, and it was a great experience.\n\nI hope you all enjoy this story as much as I enjoyed writing it.\n\n- - - - -\n\nIt was just before eleven in the morning when we arrived at my grandfather's house. We had been traveling for quite some time and I was very excited to see him. He came out of the house to greet us]" time="1.127"><properties><property name="score" value="0.9998137950897217" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Rideshare companies like Uber and Lyft are under increased scrutiny after a self-driving Uber killed a pedestrian last week.\n\nThey\u2019re called autonomous, but are they ready? It turns out not even close.\n\nYou might be surprised to learn that self-driving technology is more or less as old as the automobile itself. In fact, there are a number of cars on the road today with self-driving features, but they are only meant to assist the driver and not take over completely.\n\nOne such system, called Autopilot, was first launched by Tesla in October 2015. The main feature allows the car to drive itself on a highway, but it\u2019s still supposed to be monitored by the driver.\n\nUnfortunately, Tesla and Elon Musk have become a lightning rod for scrutiny after an accident last week that killed a Model X driver using Autopilot. A series of emails have revealed that Tesla has attempted to make it clear that Autopilot is not a \u201cself-driving\u201d feature.\n\nIn a recent letter to shareholders, Musk wrote, \u201cWhen used correctly, it is already significantly safer than a person driving by themselves and it would therefore be morally reprehensible to delay release simply for fear of bad press or some mercantile calculation of legal liability.\u201d\n\nWhat exactly does this mean? For one, this was an admission that the technology was not yet ready for public consumption.\n\nTesla is also being called to answer to whether or not the driver should have known that he was using the Autopilot feature. The company says the driver would have known since he was sitting in the driver\u2019s seat and the car\u2019s steering wheel was not in a locked position.\n\nA photo taken by a witness has been released, showing the driver looking down at something with his hands not on the wheel.\n\nThe automaker has since issued a recall to fix the issue and said it will be updating the Autopilot system to ensure that drivers stay engaged with the vehicle\u2019s operation.\n\nUber said it will be disabling the self-driving feature in its fleet of Volvo XC90s after one of its vehicles ran a red light in San Francisco last week, killing the woman. The car was in self-driving mode and the Uber employee behind the wheel wasn\u2019t able to take control of the vehicle in time.\n\nTesla and Uber are not alone in pursuing self-driving cars. Most automakers have some form of autonomous vehicle in the works, as well as companies like Google and Apple. Tesla is still the only automaker with a Level 5 vehicle that has no steering wheel or pedals, however.\n\nAccording to the National Highway Traffic Safety Administration, Google\u2019s self-driving cars have only been involved in 11 accidents over the course of 1.5 million miles of driving, but most were caused by human error.\n\nGoogle has said that the biggest hurdle for self-driving cars will be the creation of a trust between car owners and their vehicle. To that end, the company has been testing a number of prototypes over the last several years, which they claim have driven millions of miles without causing any accidents.\n\nVolvo has been developing its own version of an autonomous vehicle for some time now, but the company is not going at it alone. They\u2019ve partnered with Uber, which bought some XC90 vehicles from Volvo, to run a fleet of self-driving Volvos for ride-hailing in Pittsburgh, where Uber\u2019s research center is located.\n\nWith the recent incidents involving Tesla and Uber, the questions of what\u2019s right and what\u2019s safe have only been heightened. Tesla is going to fix the problem with its Autopilot system and Uber is pausing its own self-driving testing, but it\u2019s clear that these vehicles are a long way from being ready for mass adoption.\n\nWill the driverless cars revolution ever happen? The answer seems to be a resounding yes, but it won\u2019t be anytime soon.]" time="2.012"><properties><property name="score" value="0.576872706413269" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Truly great parties start with quality lighting. Party city and Little Tikes, combine that with other decorations and you will be sure to make your little one's birthday the best party ever. Party Supplies are not just for kids. We have a great selection of adult and kids party supplies.]" time="0.217"><properties><property name="score" value="0.8917667269706726" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[July 13, 2015\n\nNot so great...\n\nI have this straightener and while it is a nice colour, and heats up fairly quickly, it is not really worth the price. It straightens okay, but not much better than cheaper ones, and it is quite heavy so it feels like it is burning your hair when you use it, which makes it really unpleasant to use. Also, I found it did not stay very straight the day after straightening, so I ended up having to straighten my hair every day instead of twice a week. Overall, I would not recommend this straightener.]" time="0.304"><properties><property name="score" value="0.6038280725479126" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.60382807&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.60382807
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The actual intent of the changes in the UK\u2019s regulation and taxation of ISAs and pensions has become muddled in translation.\n\nThe UK\u2019s changes in pension taxation may well reduce your contributions.\n\nThe investment management industry\u2019s drive to promote personal pension products as an alternative to ISAs may have got a little over excited.\n\nFirstly, the changes that the UK government has made to the regulation of pensions and ISAs has not in fact reduced the level of tax relief on ISAs. It has however changed the way in which tax relief on contributions to personal pensions is calculated.\n\nThe current situation, whereby the actual value of the tax relief received on the sum of a pension contribution, as opposed to the rate of relief, is included in the calculation of income tax, has led to some confusion.\n\nThere is a belief that the relief that is received on the sum of a pension contribution is reduced to 10 per cent and therefore is not as good as the relief received on ISAs, where the actual value of the relief is not included in the calculation of income tax.\n\nThis misunderstanding may well have led to some people contributing to ISAs rather than pensions in order to maximise tax relief.\n\nIn addition, if the actual value of the tax relief received on the sum of a pension contribution is used as a deduction from the gross amount of the pension contribution then the level of relief has not changed at all. The actual value of the relief received on the sum of a pension contribution in the income tax calculation has increased in line with the amount of gross income that is subject to tax relief.\n\nFor example, a gross contribution of \xa310,000 would result in an actual tax relief of \xa37,500 under the current arrangements (the gross contribution is grossed up by 50 per cent to give the net pension contribution of \xa37,500).\n\nUnder the new arrangements, the gross contribution would remain at \xa310,000, but the actual tax relief received on the sum of a pension contribution would be \xa37,000. The relief on the actual value of the pension contribution is not reduced to 10 per cent.\n\nThe amount of gross income that is subject to tax relief is now \xa340,000. If you earned this amount in income before your pension contributions, then the actual tax relief on the sum of the pension contribution would be \xa330,000, rather than \xa37,500.\n\nWhat has changed is the way in which the actual tax relief received on the sum of a pension contribution is calculated. The net pension contribution (the amount actually deducted from the gross income) will no longer be grossed up to produce the actual tax relief received on the sum of the pension contribution.\n\nTherefore, the actual tax relief received on the sum of a pension contribution will always be lower than the actual tax relief received on the actual value of a pension contribution.\n\nIt is important to note that it is not the rate of relief that is affected. It is the actual value of the relief that is reduced. The rate of relief is not affected.\n\nPension contributions are still exempt from tax, although the amount of income that is excluded from tax is less than it would have been if the relief was given on the actual value of the pension contribution.\n\nSecondly, the changes to the regulation of pensions and ISAs have not increased the level of tax relief on pensions. The actual level of tax relief on pensions has not changed. The maximum level of tax relief is still 40 per cent.\n\nThe regulation of pensions has changed. Under the new rules, the maximum amount that can be paid into a pension pot, subject to tax relief, is \xa31.8 million. The maximum amount that can be paid into an ISA, subject to tax relief, is \xa37,000.\n\nBoth the maximum contribution level of pensions and ISAs has increased, but the maximum tax relief that can be obtained on both types of products is the same.\n\nIf you have already opened a pension plan, then you can only increase the contribution amount by \xa32,880 per year. Any extra contribution beyond the annual limit will not be eligible for tax relief. The full value of the contribution, regardless of how much it is, will have to be paid from your gross income.\n\nThe actual level of tax relief on ISAs has not changed. The maximum amount of ISA contributions that can be paid into an ISA in any tax year is \xa37,000. This has not changed. Any amount that is paid into an ISA in excess of this amount will not be eligible for tax relief. The full value of the contribution, regardless of how much it is, will have to be paid from your gross income.\n\nSo in summary, the actual level of tax relief on ISAs has not changed. The maximum amount that can be paid into an ISA, subject to tax relief, is \xa37,000.\n\nThe actual]" time="2.361"><properties><property name="score" value="0.9998119235038757" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[How has music from Spain been perceived in its country of origin? After World War II, a new generation of composers, nurtured by the European avant-garde, began to write in a new, progressive, experimental vein. The growth of technology, which had already made its mark on the theatre, the visual arts, architecture and the cinema, had also started to make itself felt in the music world.\n\nSpanish composers, like their European colleagues, were attracted to the possibilities offered by the new electronic and experimental instruments, and soon took their place among the most important of the international avant-garde. This programme presents a series of fascinating compositions from this period of dynamic change, providing a fascinating insight into how this new generation of Spanish composers reacted to the environment in which they lived.]" time="0.394"><properties><property name="score" value="0.9998140931129456" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[During the first year of life, children change more than they ever have before. At this stage, parents play an important role in influencing their child\u2019s development.\n\nPregnant mothers are able to greatly influence the infant\u2019s life in many ways. During pregnancy, a woman is able to optimize her diet and avoid environmental pollutants. Even while pregnant, it is important for her to stay healthy and happy, as this will in turn make the baby happier. All of this is important in a child\u2019s development.\n\nBelow are five milestones in a baby\u2019s development.\n\nMovement\n\nYour child has been feeling his own kicks and punches for a while now. He may have even kicked his mom and caused her some pain. The child will also be moving his arms and legs in a jerky fashion. As the baby grows, his movements will become more fluid.\n\nChild\u2019s Hand and Finger Development\n\nBabies are able to grab things because their hands are strong. They are also able to curl their fingers into a fist. During this time, the baby may be grasping onto a lot of things that are near.\n\nThe baby may also have strong movements, such as pinching your fingers. Babies begin to explore their world through their hands. They are also able to control their hands and fingers with ease.\n\nSitting\n\nA baby can begin to sit up on his own. This is because he has grown to a point where he is able to do this. The baby may also be able to hold his head up by himself.\n\nBabbling\n\nThe baby may also begin to make sounds that can be referred to as babble. Babies are able to make a lot of sounds and combine them together. This is a part of speech development.\n\nFeeding\n\nYour child may be starting to eat solid food. This may include mashed vegetables, cereal, and baby food. If your baby is too young to eat this type of food, it may not be a good idea to force it. This can cause choking or other complications.\n\nSource: BabyCenter]" time="1.006"><properties><property name="score" value="0.9995244741439819" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[This lecture introduces the practice of mining associations, relations, and dependencies from massive data sets.\n\nRelated Information\n\nAdditional Resources\n\nThe exercises in the course support the textbook Data Mining and Predictive Analytics in the Biomedical Sciences, by Nicolas Le Nov\xe8re, Yan Liu, Jeffrey Leek, and Balaraman Ravindran.\n\nFor background, also check out Jeffrey Leek's introductory lecture to this course.\n\nThis course is based on the book Data Mining and Predictive Analytics in the Biomedical Sciences, by Nicolas Le Nov\xe8re, Yan Liu, Jeffrey Leek, and Balaraman Ravindran.]" time="0.354"><properties><property name="score" value="0.9750154614448547" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.97501546&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.97501546
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[New !\n\nTo access the complete course listing, please log in as a member of the Medical Imaging &amp; Radiation Oncology Learning Community.\n\nMembership is free.\n\nCME Credit: 0.5 CME Credit\n\nNumber of credits allocated to this activity. The sponsoring society for this activity,\n\n, has certified that the learner may receive up to\n\n0.5 Continuing Medical Education credit(s)\n\nfor completing this activity.\n\nAccreditation Statement:\n\nSteris Corporation is accredited by the Accreditation Council for Continuing Medical Education to provide continuing medical education for physicians.\n\n\n\n\n\nThe Medical Imaging &amp; Radiation Oncology Learning Community is accredited by the Accreditation Council for Continuing Medical Education to provide continuing medical education for physicians.]" time="0.336"><properties><property name="score" value="0.6787474155426025" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[&quot; Very nice high quality item, exactly what I expected, and quick delivery. &quot; Very nice high quality item, exactly what I expected, and quick delivery.\n\n~Brian S., Australia\n\n&quot; It arrived. It fits great and looks great. Thanks! &quot; It arrived. It fits great and looks great. Thanks!\n\n~Paul L., San Luis Obispo, CA\n\n&quot; The case arrived today as promised; it fits great! I\u2019m very impressed with the quality of materials and workmanship. I plan on ordering more from you in the future, as I have one other custom amp, and will probably get more vintage ones as time goes by. Thanks again for your assistance and awesome product! &quot; The case arrived today as promised; it fits great! I\u2019m very impressed with the quality of materials and workmanship. I plan on ordering more from you in the future, as I have one other custom amp, and will probably get more vintage ones as time goes by. Thanks again for your assistance and awesome product!\n\n~Mark S., Harford, WI\n\n&quot; I received the covers today....very nice job, they look great! I haven't tried them on the amp yet but I will be for a gig Sat. Thanks for everything! PS, I bought a Pma Board a few months back and ordered one of your covers with it.....nice fit! &quot; I received the covers today....very nice job, they look great! I haven't tried them on the amp yet but I will be for a gig Sat. Thanks for everything! PS, I bought a Pma Board a few months back and ordered one of your covers with it.....nice fit!\n\n~Joe K., Williamson NY\n\n&quot; The slip arrived today and fits perfectly, you hit the nail right on the head. Thank you for your professionalism and look forward to buying one for my 4x12 flextone cab!! &quot; The slip arrived today and fits perfectly, you hit the nail right on the head. Thank you for your professionalism and look forward to buying one for my 4x12 flextone cab!!\n\n~Derek P\n\n&quot; I got the new cases for my pedal steel and they are awesome, they fit perfectly. &quot; I got the new cases for my pedal steel and they are awesome, they fit perfectly.\n\n~Earl F., Houston, TX\n\n&quot; My padded cover arrived today for my Morgan PR12. The quality is superb, and the fit is perfect. It is refreshing to get a product delivered that exceeded my expectations. &quot; My padded cover arrived today for my Morgan PR12. The quality is superb, and the fit is perfect. It is refreshing to get a product delivered that exceeded my expectations.\n\n~Larry G., Saskatoon, Canada\n\n&quot; Hey just wanted to say i got the covers last week and they are awesome!!! i will definitely purchase more in the future.my amp and cabinets are bandmate protected &quot; Hey just wanted to say i got the covers last week and they are awesome!!! i will definitely purchase more in the future.my amp and cabinets are bandmate protected\n\n~Pete G., Nova Scotia CANADA\n\n&quot; The covers arrived yesterday and they are wonderful. The Two Rock &amp; the Princeton covers fit like a glove. Thanks for your patience in getting the measurements right and the quick turnaround on my order! Appreciate it very much! &quot; The covers arrived yesterday and they are wonderful. The Two Rock &amp; the Princeton covers fit like a glove. Thanks for your patience in getting the measurements right and the quick turnaround on my order! Appreciate it very much!\n\n~Mark W\n\n&quot; I just wanted to say a great big THANK YOU for my recent order. I thoroughly am impressed with my new studio slips. &quot; I just wanted to say a great big THANK YOU for my recent order. I thoroughly am impressed with my new studio slips.\n\n~Tim S., Nashville, TN\n\n&quot; The cover fits perfectly and looks great too! &quot; The cover fits perfectly and looks great too!\n\n~Keith R., San Diego CA\n\n&quot; I received the amp cover, you do fabulous work, thank you! &quot; I received the amp cover, you do fabulous work, thank you!\n\n~Michael S., West Chester PA\n\n&quot; I received the amp cover today. Perfect fit and fantastic materials and quality work as always! Thank you for your work! &quot; I received the amp cover today. Perfect fit and fantastic materials and quality work as always! Thank you for your work!\n\n~Mel B., Taunton, MA\n\n&quot; I received my order in perfect condition, the cover fits and looks great ! It was a pleasure doing business with your company. &quot; I received my order in perfect condition, the cover fits and looks great ! It was a pleasure doing business with your company.\n\n~Jerry W., New Berlin WI\n\n&quot; Received the studio slips cover today, perfect fit! Thanks Again for a great product! Great doing business with you again..... &quot; Received the studio slips cover today, perfect fit! Thanks Again for a great product! Great doing business with you again.....\n\n~Joe V., Lindenhurst, NY\n\n&quot; I just wanted to let you know I am thrilled with the new cover. It arrived today and fits perfect. Good job and well done!! &quot; I just wanted to let you know I am thrilled with the new cover. It arrived today and fits perfect. Good job and well done!!\n\n~Kenny M.,]" time="2.800"><properties><property name="score" value="0.18003066778182983" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[For many years, and at many different homes, the grand game of monopoly was a nightly staple of our game playing regimen. Most of the time the game was played with one of my kids as the banker. The game required a bit of accounting to keep track of all the money. More often than not the game was a heated one and everyone, particularly the banker, enjoyed counting out money while money was being piled up all around the board.\n\nMonopoly is an interesting game for a number of reasons.\n\nFirst, it requires the basic accounting function of making change. The cashier has to be able to count out the correct change for all the players. This seems like a very simple, but important function. If a cashier can\u2019t count change, you would have to return the money, and you might not get it back.\n\nAlso, with any money that comes into the game there are some decisions that have to be made. For example, with a loan, how much should be charged to each player. More importantly, how much can the players afford to pay back?\n\nIn addition, how much does the bank loan to the players and what are the terms? What does the bank require to be repaid? At what time? What happens if there is not enough money to pay back the bank?\n\nThis seems like an easy concept, but a lot of my kids had a difficult time trying to figure out what they had to do to get out of debt.\n\nFinally, there are times when players simply get too deep into debt. They are forced to sell off a building, or even the entire game.\n\nThis seems like a pretty simple concept, but I found it interesting because it taught some of my kids a very valuable lesson about life. They realized how important it is to have a budget. More importantly, they realized how much money they had to spend. They had to spend wisely because they couldn\u2019t go into debt.\n\nI think it\u2019s easy to teach kids to balance their checkbook and keep track of their spending, but it\u2019s much more difficult to teach them to spend wisely and stay out of debt.\n\nThe same goes for a family\u2019s bank account. Families, just like the player\u2019s in monopoly, need to keep track of the money coming into the bank account and what is going out. They need to make sure there is enough money to cover the things they need to pay for and that there is enough money to cover the things they want to spend money on. They need to be responsible with their money and realize they can\u2019t go into debt.\n\nThe game of Monopoly teaches a lot of these principles. It is not a game of instant gratification. There are no easy money schemes. It\u2019s a game that will test the mind and the ability to be responsible with money.\n\nIf you have kids, Monopoly is a great way to learn about personal finance. You don\u2019t need to be an expert, but you do need to be willing to show them how to count and account for money.]" time="1.527"><properties><property name="score" value="0.9998129407564799" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The Bengals fell just short of a win in Kansas City Sunday, falling to the Chiefs 31-13. The loss drops the Bengals to 7-9, a number that still includes a loss to the Jacksonville Jaguars in London.\n\nOne of the biggest stories to come out of this loss is the fact that star running back Jeremy Hill was benched in favor of Rex Burkhead. Hill didn't have a good game, and the Chiefs were able to put him in his place. Burkhead was also pretty good, and scored the only Bengals touchdown on the day.\n\nBoth running backs ran for over 100 yards, but the key difference was Hill fumbled twice, including once inside the Chiefs' 20. Hill had carried the ball 39 times before the benching, while Burkhead had just nine.\n\nIt was announced Monday that Hill is still the Bengals' starting running back going forward.\n\n#Bengals Jeremy Hill named starting running back in 2018, Mixon to see more action. https://t.co/jZ9624gBwO pic.twitter.com/jOKGewN4zq \u2014 Kevin Clark (@bykevinclark) January 1, 2018\n\nThis is an interesting decision, considering the way the Bengals used the two running backs in Kansas City. Hill has been a solid running back throughout his career, but has shown lapses in concentration, as well as fumbling issues. He's not been consistent, and that has made the Bengals hesitant to hand him the starting job.\n\nThe Bengals will likely be looking to get more out of their running backs next season. We saw how talented Joe Mixon is this year, and he's the one who should have the starting job next season.\n\nHill will still be involved in the offense, but this change is a sign that he won't be a starter for the Bengals next season.]" time="0.849"><properties><property name="score" value="0.9848792552947998" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[This is the brand new video for \u201cSave Our Souls\u201d from the German melodic rockers CASANOVA. This is the third video release of the band, and you can watch it below.\n\nCASANOVA\u2019s new album, \u201cMonte Carlo\u201d will be out on January 22, 2013 via Escape Music.\n\nCASANOVA\n\nFacebook: http://www.facebook.com/casanovaofficial\n\nWebsite: http://www.casanova.de/\n\nEscape Music: http://www.escape-music.com\n\nDirk]" time="0.335"><properties><property name="score" value="0.9994240999221802" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[To: SeekAndFind\n\nThose were the good old days. Now we have NDAA, signed by \u201cwho\u2019s in charge now\u201d, saying that if you are a terrorist, you can be shot on sight, anywhere in the world, even on American soil. All you have to do is become a terrorist, and that takes less than one day in this country. It\u2019s like signing your own death warrant. The silver lining is that you might be one of the ones killed before you can become a terrorist.]" time="0.249"><properties><property name="score" value="0.6352401375770569" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.63524014&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.63524014
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Dr. Mackle and her staff are the best. I am an ex military retiree and have been getting crowns for over 10 years and this was my best experience yet. The entire staff were all friendly and professional. And they were all on time. I felt like I was part of their family. And I really liked their prices. I will use Dr. Mackle in the future. You are all awesome.\n\n\u2013 B.S. / Google]" time="0.205"><properties><property name="score" value="0.7283971905708313" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Transcript\n\n1 ADNOTAMENTO A FOGLIO NUMERO INDAGINI C.P.C.O. PROFESSIONE DISPONIBILE ORDINE CLINICO ED ORTOPEDICO 1. MEDICO ORTOPEDISTA 1 MEDICO ORTOPEDISTA ORTOTTISTA 1 MEDICO ORTOPEDISTA ORTOPEDISTA 3 MEDICO ORTOPEDISTA PSICOLOGO 1 MEDICO ORTOPEDISTA RIABILITATORE VISO 1 MEDICO ORTOPEDISTA ODONTOIATRA 1 MEDICO ORTOPEDISTA NEUROPSICOLOGO 1 MEDICO ORTOPEDISTA ASSISTENTE SOCIALE 1 MEDICO ORTOPEDISTA PSICHIATRA 1 MEDICO ORTOPEDISTA 1.1 FISIOTERAPISTA 1.1 FISIOTERAPISTA INFERMIERE 1.1 FISIOTERAPISTA OPERATORE SANITARIO 1.1 FISIOTERAPISTA DIETISTA 1.1 FISIOTERAPISTA ASSISTENTE SOCIALE 1.1 FISIOTERAPISTA COACH 1.1 FISIOTERAPISTA ELETTROMIOGRAFO 1.1 FISIOTERAPISTA DENTISTA ORTOPEDICO 1.1 FISIOTERAPISTA INFERMIERE CLINICO 1.1 FISIOTERAPISTA OSS 1.1 FISIOTERAPISTA TERAPISTA DELL AUTOMOBILE 1.1 FISIOTERAPISTA TECNOLOGO DEI MATERIALI 1.1 FISIOTERAPISTA 1.2 ORTOTTISTA 1.2 ORTOTTISTA ASSISTENTE SOCIALE 1.2 ORTOTTISTA ODONTOIATRA 1.2 ORTOTTISTA 1.3 OSTETRICA E GINECOLOGO 1.3 OSTETRICA E GINECOLOGO ASSISTENTE SOCIALE 1.3 OSTETRICA E GINECOLOGO ODONTOIATRA 1.3 OSTETRICA E GINECOLOGO 1.4 PROSTETICO IGIENE ORALE 1.4 PROSTETICO IGIENE ORALE ODONTOIATRA 1.4 PROSTETICO IGIENE ORALE 1.5 RADIOLOGO 1.5 RADIOLOGO ODONTOIATRA 1.5 RADIOLOGO 1.6 REUMATOLOGO 1.6 REUMATOLOGO MEDICO DELLA PREVENZIONE 1.6 REUMATOLOGO 1.7 OSTEOPATA 1.7 OSTEOPATA ASSISTENTE SOCIALE 1.7 OSTEOPATA ODONTOIATRA 1.7 OSTEOPATA\n\n2 1.8 PSICOLOGO 1.8 PSICOLOGO MEDICO DELLA PREVENZIONE 1.8 PSIC]" time="1.741"><properties><property name="score" value="0.19956344366073608" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.19956344&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.19956344
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Will the traditional establishment defend democracy?\n\nBy Alexandra Morton\n\nUBC-O Sea Around Us Project\n\nI am scared to my bones. For the first time in my life I am terrified that what I learned as a child is wrong. There is no more mother or father to turn to. There is no more food to count on. This is my last chance to sound the alarm that we are being systematically robbed and cheated and lied to and I don\u2019t know what we are going to do about it.\n\nA few years ago I could not have imagined that I would find myself writing this. I was born and raised in the village of Old Massett on Haida Gwaii, also known as the Queen Charlotte Islands, in Canada. We fish and hunt and gather food and make our own shelter and build our own community. This is how it has been for thousands of years. In modern language this is called subsistence living. We were the last people in Canada to sign treaties with our colonial oppressors in the 1800s. These documents were the foundation of our way of life.\n\nThen it all changed in the early 2000s when the first commercial aquaculture license was issued on our coasts. At first, like most people on Haida Gwaii, I thought that farming fish would provide income for our community. Little did I know that the promise of wealth would come with a heavy price.\n\nThe first evidence I saw that something was wrong was in 2006. The sea lice were everywhere and a fish farm had started using a pesticide called Slice. I saw the sea lice on the fish and sea lions. I saw sea lions coughing up blood, the bones on their backs visible through their skin.\n\nIn 2007, the open net pen at salmon farm salmon feedlots one kilometre off the coast of my village collapsed. An estimated two to three million Atlantic salmon, with a value of $6.5 million to $8 million, swam into the ocean. There was no monitoring, no analysis of the risk to wild salmon, not a word to local communities.\n\nAnd then the toxins appeared. In 2008 and 2009 people in my village became sick and died. In the village of Klemtu next door, children were born with flipper-like hands and feet. In Norway, which had a higher concentration of fish farms, brain tumours and other cancers were rising and women were giving birth to deformed babies.\n\nI looked at these stories and thought, \u201cThis can\u2019t happen to us.\u201d\n\nThis was followed by the closure of the Haida Nation\u2019s only grocery store. People in my village and community started suffering from serious health problems, especially neurological ones. We were told it was due to influenza or SARS. We had no idea that a toxin, which our bodies have not evolved to handle, was being released into the air and our water. The Sea Around Us reported that fish farms were creating the most toxic area in the world, right next to the First Nations in Haida Gwaii. This happened, and no one told us.\n\nI realized I had to do something. For the first time in my life I realized that what my ancestors had always taught me was wrong. It was not nature that took care of us. It was we who were responsible for taking care of nature.\n\nI wrote my first letter to the local newspaper. It was my story, and I was afraid to publish it.\n\n\u201cHow many people have to get sick and die before they take this seriously?\u201d I asked.\n\nWe waited and waited, and nothing happened. I wrote my second letter. Nothing.\n\nWe waited and waited. Still nothing.\n\nPeople in my village got sick and died, and the Norwegian scientists kept warning that their food supply was being contaminated. We were told that this was the price of progress, that it was our job to clean up the mess. And we had to do it on our own.\n\nAnd then the baby seals came. There are no young sea lions, so the seals started eating the herring. At first we thought it was cute. Then we realized they were killing the herring. There were too many seals.\n\nI wrote my third letter. I looked at the numbers and saw that Norway was killing 80,000 seals a year and there were only 80,000 seals left.\n\nI started to organize rallies and a campaign called Save the Herring, which caught the attention of major news outlets across the globe. By this time I had a voice. I was a professor of biology at UBC and one of the world\u2019s experts on fish farming, open net pen fish farms, and marine pollution. I knew the science.\n\n\u201cHow many baby seals have to get cancer before they take this seriously?\u201d I asked.\n\nIn the midst of this, Canada approved a massive new open net pen farm. This was the largest salmon farm in the world. Our local paper warned us that \u201cthe massive farm will increase marine pollution in the area.\u201d This farm started dumping blood and feces into our waters. The evidence was mounting that we were at risk of a catastrophic collapse in our food web.\n\nBut no one cared. The aquaculture industry is worth $400 million a year. We can\u2019t compete with that.\n\nA new form of ocean zoning called an impact assessment was supposed to keep salmon farms out of sensitive areas, but it was all a lie. The people who wrote it are in the industry. They gave themselves exemptions.\n\nThen I realized that I could make a living as a scientist. So I created a lab at the University of Victoria to study the toxins from fish farms. We started to find the pollutants in the mussels and the herring. I published more and more scientific papers.\n\nWe held more rallies, and I wrote more and more letters to the local paper. We were ignored and dismissed.\n\nSo we went to the Supreme Court. Our case was thrown out, but the judge was extremely critical of the Canadian government and gave them 18 months to fix their laws. But the government is now granting the aquaculture industry a blanket exemption from environmental laws and protections. The judge told them that they could not do this. But they did.\n\nI have studied toxicology for 30 years and I have never seen this. We know that these toxins are not getting into our environment accidentally. The only reason they are there is because someone put them there. It is like they are trying to poison us.\n\nAnd then we started to see a new toxin. This time it was a by-product of the feed used in the salmon pens.\n\nA Norwegian scientist tested a by-product of salmon feed. The test showed that it was toxic. The Norwegian government stopped using the feed. In 2015, that same scientist tested all the salmon feed used in the British Columbia industry. The results showed that there was another new toxin, this time a carcinogen. The Norwegian scientist said, \u201cIt is extremely unlikely that it could happen at such high levels by chance.\u201d\n\n\u201cHow many toxins have to get into our food before we take this seriously?\u201d I asked.\n\nThis time I knew that there was something wrong with my brain. We found out that Norway and Scotland were using the same feed. We had a toxic lab here in British Columbia, so what about Norway and Scotland? And the feed is approved by the Canadian Food Inspection Agency and used in these salmon farms.\n\nIt\u2019s hard to see who we are going to turn to.\n\nThen a woman in Ketch Harbour started asking questions. She wrote to the local paper and was ridiculed and insulted. She wrote to the Minister of Fisheries and Oceans. He never replied.\n\nShe was not from my village. She was a white woman from Newfoundland. Her grandfather was a fisherman.\n\nI wrote to her, and then I realized I could write to everyone. And I did. I started to get hundreds of people to write letters, and we wrote to the Minister of Fisheries and Oceans and to our Premier, the head of our government. We wrote to every MLA, the Members of the Legislative Assembly, and we were ignored and dismissed.\n\nSo we went to court again. This time we were supported by 30 affidavits from fishermen and local citizens.\n\nI went to the hearings. The minister did not even show up. His lawyer said that we should stop questioning him and just accept that he is doing the right thing. He had no explanation as to why the 30 affidavits were not enough evidence to put a stop to this industrial salmon farming.\n\n\u201cIf I had a million dollars, I could save Haida Gwaii,\u201d I said.\n\nBut I have never been able to raise that money. The man who tried to stop me is now running for Premier, and he is the frontrunner.\n\nSo I wrote my fifth letter to the editor. I warned that if we continued to use this feed, we would all get sick and die.\n\nAnd then I realized that I could put this on Facebook.\n\nA few months ago, the BC government ordered the feed company to remove this toxin from their products. But they still sell it in other provinces. There is still no ban on the feed, even though Norway has banned it, the FDA has banned it, and Health Canada has banned it.\n\nAnd then I realized I could just get on a bus and go to Toronto and speak to people directly.\n\nSo that is what I am doing. I am going to make this my life\u2019s work. I am going to stand in front of people, in universities, in public hearings, and tell them my story. I am going to explain that it is not]" time="4.813"><properties><property name="score" value="0.5924304922421774" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[I am a very fair skinned male and this tanning lotion has worked wonders on my pale body. It was quick and easy to apply. It also was very easy to clean off my hands after I applied it, which was a big plus.]" time="0.194"><properties><property name="score" value="0.977544903755188" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Veggie burgers are great! The more veg I add the better I feel about my nutrition. Plus I am not that great at taking vitamin supplements. I do eat all the veggies I need though.\n\nMy first veggie burger was just lentils, eggs, and bread crumbs. No extra seasonings or flavours. They were ok but really needed salt. This recipe is so much better. I used this recipe as a basis but made a few modifications. I used 3 cups cooked quinoa, 2 cups kale, 1 cup spinach, 4 large carrots, 4 celery stalks, 1 onion, and 1 cup corn. I added a tsp. of thyme and 1 tsp. of cayenne pepper. I had a tiny bit of ground pork that I needed to use up so I also added that. I used black beans instead of pinto beans because I prefer them. I added extra bread crumbs and an egg to bind it. I think I put in 3 Tbs. of olive oil as well.\n\nI am happy with the result. The texture is really nice and the flavour is delicious! I served it with some sriracha mayo that I made. YUM!\n\nAdvertisements]" time="0.546"><properties><property name="score" value="0.9996895790100098" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[JUMBLE CONFIDENCE\n\nDefinitions\n\nConfidence: A term describing the likelihood that a decision-making system will reach a good decision. It refers to the quality of a decision, not to the quality of the process that leads to that decision.\n\nJumble: A pattern of naturalistic decision making. It does not include extraneous information or create false information.\n\nExample\n\nA fighter pilot is confident that she will destroy the enemy aircraft even though she has only her own memory of previous aerial encounters to rely on.]" time="0.293"><properties><property name="score" value="0.9753872156143188" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Anselm and his ontological proof, is one of the most important Christian philosophers of the Middle Ages. He is also well known for his attempts to fuse the sciences with theology, something of a controversial topic for medieval theologians. The ontological argument itself is based upon the idea that God\u2019s nature is that of existence. Therefore, to say that God exists is to say that God has the greatest nature, which in turn leads to the conclusion that God must exist. Some scholars argue that this is the proof\u2019s most fundamental flaw, namely that God\u2019s nature is unprovable and thus, cannot be equated to the definition of existence. Other critics point out that the greatest nature is really a metaphysic and therefore cannot be understood by people. The latter claim has been supported by Immanuel Kant and many other philosophers since. Kant\u2019s work was not known to Anselm\u2019s, but it is clear that the German philosopher would not have found much use in Anselm\u2019s philosophy.\n\nThe foundation of Anselm\u2019s ontological argument is a branch of philosophy known as natural theology. Natural theology is the search for a knowledge of God through the use of reason. The argument\u2019s goal is to prove that God exists in some way, and as such is a concept that is widely accepted in the world of Christian philosophy. In many ways, Anselm\u2019s argument for God\u2019s existence shares many similarities with the arguments of many of his contemporaries. Anselm claims that God exists and that this conclusion can be proven using the method of defining terms. The proof proceeds from the claim that God is the greatest thing conceivable to the claim that God must also exist. Anselm\u2019s argument does not satisfy many people because it does not provide any evidence to support the claim that God is in fact the greatest conceivable thing. This is one of the major weaknesses of Anselm\u2019s argument, but it is also one of the weaknesses of most arguments for the existence of God. The ontological argument can be seen as an attempt to prove God\u2019s existence without supporting any specific philosophical view of God\u2019s nature.\n\nMuch of Anselm\u2019s work focuses on the attempts to explain and prove God\u2019s existence. He was one of the earliest of many philosophers to attempt to explain and prove God\u2019s existence. The debate over the existence of God has continued over the years, and the arguments have been supported and refuted on many different occasions. Anselm\u2019s work is important because he was one of the first philosophers to put forth the idea that existence itself was proof of God. He was also one of the first Christian philosophers to try and explain the idea of God in a non-theological way. Many of Anselm\u2019s contemporaries had rejected his view on God\u2019s nature, and they also criticized his work on natural theology.\n\nOne of the major debates over Anselm\u2019s work was that of the ontological argument. This is a popular concept that states that God\u2019s existence is a self-evident truth. The name of the argument comes from the phrase \u201csomething that is greater than which cannot be conceived\u201d. This concept became widely accepted in the world of Christian philosophy, and it is still being debated today. The concept of God\u2019s existence as a self-evident truth has been supported by many Christian philosophers, including Thomas Aquinas. Aquinas\u2019 view of God was one of the most influential views of God\u2019s nature in the Middle Ages. The ideas presented by Anselm and Aquinas on God\u2019s nature were among the first attempts to view God in a logical way.\n\nMost of Anselm\u2019s work focused on the idea of God and the way in which God was perceived. He focused on the idea that God was perceived through faith and that this idea had a real and substantial foundation. Many critics viewed Anselm as a philosophical follower of Augustine, and he was certainly influenced by many of Augustine\u2019s works. The Confessions of Augustine were written in the fifth century, and they discussed the views of the time on God and the manner in which people should understand God. Anselm\u2019s view of God is therefore quite different from many of his contemporaries, and he was heavily criticized for his work on natural theology.\n\nThe ideas presented by Anselm were extremely influential to his contemporaries. Although his ideas were at times criticized, many of his contemporaries found them to be the most logical views of God. His argument for God\u2019s existence was extremely popular among his peers, but it is one of his most controversial works.\n\nAnselm of Canterbury Biography\n\nAnselm of Canterbury was born in 1033 in Aosta. He was the second of four sons and his parents were wealthy. Anselm\u2019s father was an Italian who owned a noble estate. His mother was the niece of the archbishop of Rheims. Anselm was well educated, and he studied both rhetoric and literature as a young man. He studied law in 1057 at the cathedral school in Laon. During this time, he became close friends with Roscelin of Compi\xe8gne. Roscelin was an expert on logic, and he would be one of Anselm\u2019s best friends and mentors for the rest of his life. He was also close to Lanfranc and eventually became his student.\n\nAnselm studied theology in the early 1060s at the cathedral school of Chartres. He also spent a great deal of time studying theology and grammar at the school of Benevento. Anselm spent two years at the school and was then called to the cathedral school of Canterbury in 1062. Anselm was called to the cathedral school to replace the archbishop. The archbishop of Canterbury was the highest ranking prelate in England and he was charged with the administration of the Church in England. Anselm quickly gained respect in his new position, and he wrote a new rule for the cathedral school. The rule was based on the Rule of St. Benedict, which was the most important rule for monasteries at the time. He was also responsible for the cathedral school\u2019s many other duties and its administration. He established the system of electing canons and was also in charge of its finances. Anselm had a great deal of power in his new position, but he also was a reformer. He placed emphasis on the importance of education and scholarship, and he ensured that the students received high quality instruction. The rule at the cathedral school was seen as being high quality and the students at the school were trained in Greek, Latin, rhetoric, dialectic, and arithmetic.\n\nAnselm also established a library at the school. The library contained many ancient works, which were frequently read by students. He became an expert on the writings of Augustine, and he began his own writing career in the early 1060s. His writings were influential in the development of later philosophy, and his ideas on natural theology were both influential and controversial. In 1063, Anselm was appointed archbishop of Canterbury. This was an extremely important position, and Anselm would play a major role in the Christianization of England. He was a major contributor to the revival of learning in England, and he attempted to bring education to all of his subjects.\n\nHe was a skilled theologian, and he produced several important theological works. These included the Monologion and the Proslogion. The Monologion addressed the questions of the nature of God. The Proslogion addressed the questions of the nature of God\u2019s knowledge and the proof for God\u2019s existence. Anselm was influenced by Augustine in his writing of the Monologion]" time="3.986"><properties><property name="score" value="0.9955440387129784" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Advantage Resourcing Ltd - Northampton\n\nWe are currently recruiting for a Production Operative on behalf of our client based in Northampton. This role is starting immediately on a temporary to permanent basis. If you have experience in Manufacturing then please get in touch. You will be required to carry out your duties in a high quality...]" time="0.181"><properties><property name="score" value="0.9997707009315491" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.9997707&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.9997707
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The Blastlite Lite BTM (Blastlite Thermal Monitor) is a small, single-output controller that supports a wide range of time-programmable and event-triggered outputs. The time programmable outputs are available as dry contacts (NO/NC) or as solid-state relays.\n\nThermal sensors are available for a wide variety of applications and locations. Select the correct sensor for the task at hand using the Sensor Selector Guide.\n\nThis thermostat was designed to easily mount to the side of any standard 1 1/2 inch (38mm) junction box with the included mount bracket.\n\nBlastlite Lite BTM Features\n\nLED status indicators\n\nEasily programmable for time-triggered and event-triggered outputs\n\nThermal sensors for flame detection and appliance temperature monitoring\n\nBlastlite Lite BTM Specifications]" time="0.430"><properties><property name="score" value="0.9326208829879761" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[\u201cDark\u201d will be the third in a series of single-pitch sport routes at Boulder Zephyr in Georgetown, Ontario. The sport crag is located a few kilometres west of Blue Mountain, Ontario. Here are the proposed grades and estimated lengths of the routes, which are currently being developed:\n\nDark (5.12a) \u2014 35m\n\nMorning Light (5.11c) \u2014 35m\n\nDawn Patrol (5.11b) \u2014 35m\n\nLunch Break (5.10b) \u2014 35m\n\nVibes (5.10a) \u2014 35m\n\nThough none of these routes are bolted yet, and they are still only in the development stage, Boulder Zephyr is already a great place to go bouldering!\n\nSource: Peter Croft (Bouldering Ontario)]" time="0.371"><properties><property name="score" value="0.9951702952384949" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.9951703&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.9951703
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[In California, a convicted pedophile is on the loose. The worst part? He's a convicted pedophile.\n\nFrancisco Felix-Jimenez was released from prison last week. He was convicted of molesting an 8-year-old boy in 2006.\n\nWell, it turns out that the Los Angeles County Sheriff's Department forgot to register Felix-Jimenez as a sex offender.\n\nYou may be asking yourself, &quot;How does that happen?&quot;\n\nThey may have simply forgot. It happens all the time, although it's usually because a sheriff's department employee who should have registered the convicted sex offender doesn't think the guy will do it again.\n\nSometimes a new chief or sheriff takes over the department, doesn't like the policy, and cancels the sex offender registry.\n\nHere's another scenario: When the convicted sex offender is released from prison, he is given a deadline by the parole board to register.\n\nIf the parole board gives the guy a year, the sheriff's department is probably going to give him a year.\n\nHowever, if the parole board gives the guy a year and the county sheriff wants him off the street, he will register the convicted sex offender that day.\n\nIn the case of Felix-Jimenez, the sex offender registry is mandatory and the deadline wasn't up yet, so how he is able to be on the loose is still unclear.\n\nThe good news is that Felix-Jimenez hasn't gone into hiding.\n\nIn fact, the most recent report shows that he is staying in a homeless shelter in Van Nuys.\n\nCalifornia doesn't want him there.]" time="0.830"><properties><property name="score" value="0.8694588840007782" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Granted, my current policy does not meet what I want to do, so I have done some thinking on what changes would be nice. There are two major changes I think should be made.\n\nThe first would be to include the deletion policy as the first part of the \u201cFAQ\u201d section, as opposed to the second, where it is now. The issue here is that it is currently hidden away, and unless you know to go there, you might not even know it exists. This is why I believe the policy should be in the FAQ, in addition to the other policy changes.\n\nThe second major change would be to change the time periods for deleting and undeletion requests. Currently, these time periods are a month for both and can be extended for a total of 2 months. These time periods need to be shortened. If a user creates an account, and decides it is not for them within a week, they should be able to delete it without the issue being made public, since no other users have commented on it. If a user creates an account and decides that it is not for them within the first month, they should also be able to delete it without it being made public. Also, it is not fair to users who are newer, and decide that they do not wish to use the site for the first time period to only have it deleted. These users should be able to delete them after this time period if they choose. I believe these time periods should be a month for deletion requests, and 2 weeks for undeletion requests.]" time="0.766"><properties><property name="score" value="0.9946683347225189" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Perhaps best known for their exploits during the First World War and the Gallipoli campaign in particular, the ANZACs, or Anzacs, are the legendary members of the Australian and New Zealand Army Corps.\n\nTheir contribution to the military campaign in the Middle East, including the battle at Gallipoli in 1915, will be recognised this year, with the 100th anniversary of the Gallipoli campaign being commemorated by the nations involved.\n\nThe events of the Gallipoli campaign resulted in heavy casualties, with a staggering 258,000 deaths and an additional 156,000 injuries for the Allied forces, including 44,000 British deaths and 29,000 French deaths.\n\nThe staggering casualty figures have subsequently been seized upon by those who question the strategic worth of the campaign.\n\nOne figure, British historian Niall Ferguson, wrote that \u201cthe Dardanelles fiasco was one of the greatest military blunders of all time, a story of incompetence, vanity and wishful thinking out of which could have been constructed a textbook case of how not to run a war\u201d.\n\nSo why were the Gallipoli campaign and the actions of the ANZACs so significant for Australia and New Zealand, and what did the battle mean for the two countries?\n\nArrival of the ANZACs\n\nFollowing the outbreak of war in 1914, the Australian and New Zealand forces began arriving in Egypt. On February 25 1915 the first ANZAC troops arrived at Gallipoli, a total of 11,000. The rest of the 20,000 strong force arrived over the following weeks.\n\nMany of the ANZAC troops had volunteered at the outbreak of war, lured by the romantic and patriotic view of warfare and by a yearning for adventure.\n\nJust as the troops were largely made up of volunteers, the casualties at Gallipoli were also predominantly made up of volunteers.\n\nIt has been estimated that 10 percent of all the Australian and New Zealand troops at Gallipoli were killed during the campaign.\n\nAustralian Prime Minister Sir Robert Menzies commented that \u201cthe experience of war for Australians would always be Gallipoli\u201d.\n\nPrior to the arrival of the ANZAC troops, the British had launched an attack on the Gallipoli peninsula on the 25th of April 1915.\n\nThe Allied attack consisted of five divisions, with most of the troops coming from the British Army. It was hoped that the landings would help the British Navy to get through the Dardanelles, a narrow strip of water separating the Aegean Sea and the Sea of Marmara, which leads to the Bosphorus and the Black Sea.\n\nHowever, the initial landings met with great opposition from the Turkish defenders, and the Allied forces soon became pinned down on the beaches and unable to advance.\n\nThe ANZAC troops arrived at Gallipoli, the capital of the Ottoman Empire\u2019s European territories, just over a month later on the 19th of May 1915.\n\nThe ANZACs had initially been part of the British and French force in Egypt. They were moved to a separate command under General Ian Hamilton, with Australian General William Birdwood as his deputy.\n\nA General Order issued on the 3rd of May stated that the Australian and New Zealand troops were to be given \u201cseparate and distinct\u201d duties from those of the British.\n\nThis was not intended to be interpreted as any form of segregation. Rather, it was due to the fact that the Australians and New Zealanders, despite being part of the British Empire, were soldiers who had volunteered from their own country, and who had formed a unique sense of national identity and pride.\n\nIn order to maintain a clear distinction between the Allied forces and the enemy, the ANZACs also adopted their own distinctive insignia, such as a yellow or red poppy on their badges and hats.\n\nHowever, despite the greater efforts to set the ANZACs apart from their Allied allies, the treatment of the soldiers at Gallipoli was the same as that received by the British.\n\nThe ANZACs were given responsibility for the defence of the Anzac Cove, at the northernmost tip of the Gallipoli Peninsula.\n\nThe cove was strategically important to the Turks because of its proximity to the Narrows, the strip of water which separated the peninsula from the Dardanelles.\n\nThe Anzacs were also expected to hold on to the various hills and ridges along the peninsula, despite the fact that the Allied positions were in constant danger of being shelled]" time="2.352"><properties><property name="score" value="0.9980734705924987" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[V6 engines had too much power to safely make use of the front wheels alone. V8 engines didn\u2019t fit between the front and rear axles, making them virtually useless for most people.\n\nHowever, a minivan with four-wheel drive would have been a dream come true.\n\nSource]" time="0.203"><properties><property name="score" value="0.971993625164032" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[I love hand-cranked coffee grinders, but a fully automated device that grinds directly into the portafilter has always been one of the holy grails of coffee brewing. The Ekobrew takes a lot of the guesswork out of the whole process. Fill the hopper with whole coffee beans, twist the top to open, drop the portafilter in, turn the handle a few times and you\u2019re ready to make your espresso.\n\nThe container is made of a non-porous material, so you can put ground coffee in it. It also has a couple of holes on the side, so that you can clean it and store it without having to remove the portafilter. The concept is pretty ingenious, but I had some difficulty in actually getting it to work properly.\n\nIn theory, the machine takes a lot of the stress out of making espresso. When it works, the espresso is good. It has a big enough capacity that I could probably make it twice before having to refill it. It would be nice if there was a better way to tell how full it was.\n\nUnfortunately, it\u2019s a fairly unreliable method of extracting espresso. You can only have it set to brew for a specific amount of time. You can\u2019t use any sort of fancy steaming wand to make sure the pressure is exactly right, or adjust the grind to get a specific flavour. I think this device is designed for making a quick shot in the morning and then tossing the puck after the shot is made.\n\nThis is the kind of device that I can imagine being useful to take on a trip, or maybe for an RV or camper, but not in a regular espresso bar or cafe. At the price of $50, I can\u2019t really recommend it, though I do think that it has potential. If it were a bit more inexpensive, I think it would be worth experimenting with.]" time="0.869"><properties><property name="score" value="0.9998034834861755" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[I like it!\n\nI like it!\n\n$750 DSQUARED2 Short dress is on sale! $375\n\nI like it!\n\nI like it!\n\n$170 CRISTINAEFFE Short dress is on sale! $72\n\nI like it!\n\n$290 CHRISTIES \xc0 PORTER Short dress is on sale! $206\n\nI like it!\n\nI like it!\n\nI like it!\n\nI like it!\n\nI like it!\n\n$340 ERMANNO DI ERMANNO SCERVINO Short dress is on sale! $254\n\nI like it!\n\n$440 Jil Sander Short dress is on sale! $323\n\nI like it!\n\n$138 TROU AUX BICHES Short dress is on sale! $85\n\nI like it!\n\nI like it!\n\nI like it!\n\n$460 Alberta Ferretti Short dress is on sale! $322\n\nI like it!\n\nI like it!\n\nI like it!\n\n$219 CRISTINAEFFE Short dress is on sale! $117\n\nI like it!\n\nI like it!\n\nI like it!\n\n$104 JUST FOR YOU Short dress is on sale! $70\n\nI like it!]" time="0.660"><properties><property name="score" value="0.782539427280426" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.78253943&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.78253943
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Horses, rabbits and fish swimming around in glass globes are an enticing attraction at Hanging Lake Lodge in Fruita. But, for more than a century, the people have had a passion for Hanging Lake.\n\nChildren, horses, deer and pet dogs are all welcome to stay at Hanging Lake.\n\nDogs are welcome at Hanging Lake Lodge, as long as they don't swim.\n\n&quot;There were (people) that were always trying to get to the lake,&quot; said Dale Hanks, whose great grandfather built the original Hanging Lake Hotel.\n\nHanks has been running Hanging Lake Lodge since 1972, and says his customers are almost always from out of town.\n\n&quot;They've been told about it by their neighbor, they've been told about it by their son who visited last year, they've been told about it by friends in Colorado, and when they're here they just have to come,&quot; Hanks said.\n\nMany visitors are drawn to Hanging Lake by the nearby Red Canyon, as well as the Colorado National Monument. The lake is located in a canyon along the Colorado River, nestled at the base of a rock formation called the Palisades. The National Monument, also known as The Dots, is a rock formation of mostly red and orange that spans for about a mile along the Colorado River.\n\nThe Dots.\n\nHanging Lake is a water body that was formed by lava over millions of years, according to the National Park Service. The water body is fed by a small underground river and water seeps out through the canyon walls to form the lake.\n\nHanging Lake is known as one of the most picturesque lakes in the country, and often the photos seen online or on Instagram have a filter on them. The color of the water is bright blue, the rocks are a fiery red and orange, and the waterfalls are turquoise blue.\n\nA tourist sits by the Hanging Lake waterfall.\n\n&quot;It's an excellent opportunity to do some waterfall photography,&quot; said Ben Lin, a Denver resident. &quot;That is what drew me here, the waterfall, and I don't know why, but it is very hard to take a bad picture here.&quot;\n\nLin has been to Hanging Lake several times, the first time in 2015. He came with a small group of about 10 people, and the people in his group were in awe at the sight.\n\nThe fall\n\n&quot;This is the best,&quot; Lin remembers one person saying. &quot;This is the best picture, we are so lucky.&quot;\n\nThe water temperature in the summer is about 45 degrees, and not many people are willing to dive into the water, which Lin says is a shame.\n\n&quot;I feel like you should be able to swim here, it's a beautiful place to swim,&quot; Lin said. &quot;The problem is the lake is only as deep as your chin. When you're done swimming there is all that debris and gunk on the lake, you can't see it from up above, but as you're swimming you feel the gunk and rocks in the bottom of the lake.&quot;\n\nAlthough swimming is prohibited, people do still swim in the lake, which causes the algae to increase, according to the National Park Service.\n\nHanging Lake is one of the first stops for many travelers, but after leaving the picturesque sight, many people don't realize there is so much more to Fruita than Hanging Lake.\n\nNext time you are passing through the small town of Fruita, consider stopping at Hanging Lake Lodge for lunch, a short hike, or an afternoon in a hammock. The history behind the town and the visitors will not disappoint.\n\nEditor's note: This article has been updated to correct information about Hanging Lake. The photos seen on Instagram are not enhanced in any way.]" time="1.997"><properties><property name="score" value="0.9945235997438431" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Fantasy baseball: Six more players who may not be healthy\n\nSportsPulse: USA TODAY Sports' fantasy baseball experts tell you which players you should buy low and sell high as the season gets into full swing.\n\nJoin the conversation or respond by signing in\n\nFantasy Home\n\nPopular Tags\n\nPlease enable JavaScript to view the comments powered by Disqus.]" time="0.217"><properties><property name="score" value="0.8311710357666016" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[You won't see me taking the same chances\n\nYou can't expect me to start taking your drugs again\n\nAnd now I'm out of the darkness I'll never be lost again\n\nI'll never be lost again\n\n\n\nYou never knew my secret name\n\nYou never heard of my all seeing eyes\n\nYou never knew my secret name\n\nYou never heard of my all seeing eyes\n\n\n\nYou know you're walking on the ceiling and now you're kicking my ass\n\nYou're not the only one who has had the best of me\n\nNow that I'm out of the darkness I'll never be lost again\n\nI'll never be lost again]" time="0.310"><properties><property name="score" value="0.9997739195823669" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Corporate Counsel is partnering with several organizations to host an event on October 20, 2017, on the future of diversity in corporate law, featuring Keisha Lance Bottoms, Mayor of Atlanta; Deidra Lee, Atlanta District Attorney; Aisha Dew, Founder &amp; CEO of We The Change; and Connie Maysonave, CEO of Atlantans for a Fair Future.\n\nThe event will feature a moderated conversation among Mayor Bottoms, Ms. Dew, Ms. Lee, and Ms. Maysonave, followed by a networking session and a reception. The event will take place at the Loews Atlanta Hotel from 5:30 to 7:30 p.m.\n\nTo register for the event, click here.]" time="0.342"><properties><property name="score" value="0.9025541543960571" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[He has won numerous awards and has been honored by both the people and the government of Jamaica. These accolades are well-deserved and confirm the love and respect the people of Jamaica have for this music genius. In December of 2005, Williams won the prestigious Golden Arrow award for his music and contribution to the music industry in Jamaica. In February of 2006, he was awarded the honorary Order of Jamaica by Prime Minister Bruce Golding at a gala event held at Jamaica House.\n\nWalter &quot;Sweetie&quot; Williams has been a major force in the growth and popularity of Jamaican music, in particular]" time="0.280"><properties><property name="score" value="0.8236011266708374" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.82360113&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.82360113
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[During my first month of graduate school, I received a call from my aunt on a weekend when I was in the middle of working on my thesis. She needed help moving out of a rental property she owned. I went over that weekend and helped her clear out the house.\n\nAfter it was all done, she said to me, \u201cYou know, a lot of people would be really upset at me for making them do this.\u201d\n\nI smiled.\n\n\u201cThat\u2019s why I am here.\u201d\n\nIt\u2019s funny how a simple answer can change the dynamics of a relationship.\n\nI continued to help my aunt after that weekend. She would ask for my help, and I would drop what I was doing. She asked, I gave. I took it for granted that it was my responsibility to help her out, to be available. I also took it for granted that she was comfortable asking me for help.\n\nA few years ago, I went to see my aunt in the hospital. She was having a difficult time recovering from her medical procedure, and I was her first call. I didn\u2019t think twice about dropping everything and going over to her place to take care of her and her dog, especially since she had just moved to an apartment that didn\u2019t allow dogs.\n\nBut I got to the hospital and found out that my uncle was also at the hospital. I didn\u2019t realize that he was in the hospital, and, by the time I found out, he was asleep in a nearby room. I talked to my aunt and we came to the realization that it was important for her to have my uncle there. She needed him in a way that she didn\u2019t need me. I left.\n\nI don\u2019t think my aunt and uncle even realized how much I had changed. I was not as available to help as I used to be, and I knew my time was better spent on other things. I was a graduate student, after all, and I had a whole other life to live.\n\nAs a result of all of this, my aunt and uncle started to depend on others to help out more. They hired a house cleaner and a housekeeper. They began to call people in order to receive the help they needed. I had thought that they would not want to go through the hassle of finding someone to help, but they did.\n\nIt\u2019s as if they needed to be reminded that it was okay to ask for help. They needed to realize that asking for help wasn\u2019t a sign of weakness, but a sign of strength. That asking for help meant you were confident in who you were and what you had to offer, and that you weren\u2019t afraid to ask for help if you needed it.\n\nIn doing so, my aunt and uncle showed me what it meant to ask for help, and I realized that asking for help could be a sign of strength, not weakness.\n\nAdvertisements]" time="1.562"><properties><property name="score" value="0.9992287158966064" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Une analyse de la CNIL\n\nBac d\u2019information, traitement automatis\xe9, robots, IA\u2026 nous avons tendance \xe0 penser que les syst\xe8mes automatis\xe9s ne sont pas faits pour r\xe9pondre \xe0 nos besoins de compr\xe9hension, de sensibilit\xe9 et d\u2019empathie. Or, \xe0 l\u2019occasion de l\u2019assembl\xe9e g\xe9n\xe9rale de l\u2019institut Data &amp; Society qui a eu lieu la semaine derni\xe8re, nous avons eu l\u2019opportunit\xe9 d\u2019entendre les explications d\u2019Omer Tene, chercheur en sciences cognitives, et professeur associ\xe9 en intelligence artificielle. Il nous a d\xe9montr\xe9 que les machines n\u2019\xe9taient pas forc\xe9ment con\xe7ues pour remplacer l\u2019homme. Bien au contraire. L\u2019apprentissage automatique est aujourd\u2019hui une v\xe9ritable valeur ajout\xe9e et permet de d\xe9passer les limites des technologies humaines.]" time="0.594"><properties><property name="score" value="0.9997802376747131" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99978024&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99978024
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[C\xe9sar D\xedaz \u2022 29/10/2018 - 10:08\n\nSebasti\xe1n, yo te recomiendo que te hagas el seguimiento de los indicadores y que no hagas mucho caso de los medios tradicionales. El \xfaltimo sondeo se\xf1ala que las elecciones se definir\xe1n por un muy muy muy estrecho margen de diferencia. La probabilidad de que el Frente Amplio gane el 23 es muy alta y cuanto m\xe1s alta es, mayor es la probabilidad de que gane. En estos momentos hay dos variables que van a determinar que el FA gane:\n\n- Cambios en la opini\xf3n p\xfablica (es decir, la propaganda, la capacidad de atraer m\xe1s votos, etc).\n\n- Efecto S\xe1nchez (no s\xe9 qu\xe9 les pasa, se equivocaron en las \xfaltimas elecciones de Espa\xf1a y ahora si no es el chileno el que entra a saco contra Vidal, el brasile\xf1o tambi\xe9n).]" time="0.731"><properties><property name="score" value="0.9650061130523682" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.96500611&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.96500611
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Why do we need 2 statements? Isn't the first one enough to start a transaction?\n\nThe first statement is an Open. It does not actually do anything yet. It sets up the transaction, but does not actually start it. In order to do something, you must start the transaction.\n\nIt can be done, and in fact, I think that is how it is done in your example code. However, I prefer to use a begin statement for the transaction and then perform the Open within the transaction.\n\nI do this for 2 reasons:]" time="0.307"><properties><property name="score" value="0.8617724776268005" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Overview\n\nGreta Esta\xf1ol's Carmina \xc1urea are sacred poems meant to inspire devotion in the faithful. They are simple in form, yet profound in content. The author has combined medieval traditions with a contemporary perspective on the faith, creating a simple, sincere, and authentic rendition of the Catholic religion. It was in this vein that she wrote these lyrical and enlightening poems. The collection consists of approximately two hundred poems, many of which have been translated from Esta\xf1ol's Spanish original by Michael Coffey. Carmina \xc1urea are an introduction to Christian spirituality. This collection contains the following poem: To Learn the Way\n\n\n\nThe Value of Purity\n\nLove Conquers the World\n\nHoly Bible\n\nThe Sacraments\n\nLiturgical Calender\n\nMy Jesus\n\nA Prayer for Peace\n\nGod Loves Me\n\nThe Feast of the Immaculate Conception\n\nThe Sacrament of Baptism\n\nFaith and Confession\n\nThe Sacrament of Holy Communion\n\nOur Father, Who Art in Heaven\n\nThe Story of Our Lord's Birth\n\nThe Easter Vigil\n\nThe Holy Mass\n\nThe Sacrament of Marriage\n\nThe Sacrifice of the Mass\n\nThe Sacrament of Confirmation\n\nHoly Mary, Our Mother\n\nSacred Heart of Jesus\n\nOur Lady of Sorrows\n\nThe Our Father and Hail Mary\n\nThe Great Love\n\nBlessed Virgin, My Mother\n\nThe Holy Rosary\n\nThe Sacrament of Penance\n\nThe Gifts of the Holy Spirit\n\nHoly Spirit, Breathe on Me\n\nThe Day of Judgment\n\nThe Second Coming of Our Lord\n\nThe Epiphany\n\nOur Lady of Guadalupe\n\nThe Holy Spirit\n\nHoly Trinity, Trinity\n\nO Lord, I Will Not Keep Silent\n\nThe Most Holy Eucharist\n\nThe Trinity\n\nO Most Holy Trinity\n\nJesus, Mary, Joseph\n\nAn Angel of God\n\nOur Father, Who Art in Heaven\n\nOur Lady of Guadalupe\n\nHoly Father, Have Mercy on Me\n\nThe Sacrament of Extreme Unction\n\nThe Incarnation of Our Lord\n\nIn the Beginning\n\nAmen\n\nThe Power of Prayer\n\nTo a Loved One\n\nThe Love of God\n\nThe Holy Spirit\n\nWho Is the Holy Spirit\n\nHoly Spirit, the Breath of Life\n\nGlory Be to the Father\n\nThe Holy Family\n\nOur Lady of Lourdes\n\nBlessed Mother, Behold Your Son\n\nHail, Holy Queen\n\nTo Mary, My Mother\n\nPraise the Lord\n\nHail, Holy Queen\n\nTo the Crucifix\n\nThe Angelic Salutation\n\nA Prayer for Parents\n\nTo the Sacred Heart of Jesus\n\nA Prayer of Thanksgiving\n\nCome, Holy Spirit\n\nJesus Christ, My Love\n\nThe Holy Spirit\n\nHoly Spirit, Dwell in Me\n\nO Holy Spirit, Thou Who Art Thrice Holy\n\nO Holy Spirit, Unction of God\n\nO Spirit of Love, Uniting Soul and Body\n\nCome, Holy Spirit, Come\n\nHoly Spirit, Light of Love\n\nHoly Spirit, Send Your Spirit of Love\n\nHoly Spirit, The Spirit of Love\n\nThe Holy Spirit\n\nThe Trinity\n\nAll Creation Loves You\n\nThe Baptism of Jesus\n\nAn Angel of God\n\nHe Is a Man Who Sees\n\nBehold the Word Made Flesh\n\nSing a Song of Incarnation\n\nBehold the Door of the Kingdom of Heaven\n\nMay the Glorious Mysteries of Jesus\n\nCome to Us, O Jesus\n]" time="1.900"><properties><property name="score" value="0.9954071839650472" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The BB&amp;T Building (also known as the 100 North Tryon Building) is a 17-story skyscraper in Charlotte, North Carolina, United States. It is the tallest building in Charlotte outside of Uptown. It was originally known as the Bank of North Carolina Building, when completed in 1921. In 1973 it became known as the NationsBank Building. In 1998 it became the Wachovia Building and in 2004 it became the BB&amp;T Building.\n\nArchitecture\n\nThe BB&amp;T Building is designed in the Art Deco style, which became popular during the 1920s and 1930s. The building features limestone and terra cotta facing, and metal and glass curtain walls. The top of the building features two pyramid-shaped ornamental caps.\n\nTenants\n\nThe building was built in 1921 for the Bank of North Carolina. The bank grew to be the largest bank in the South, before merging with First Union in 1998, becoming Wachovia. Wachovia's Charlotte headquarters occupied three floors in the building until the company merged with Wells Fargo in 2009.]" time="0.557"><properties><property name="score" value="0.993791401386261" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.9937914&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.9937914
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Investment fund transfers are back to normal now that all new subscriptions have been processed for the month of July.\n\nOnce you've logged into your account, you'll notice that your portfolio has been adjusted to reflect the purchases and redemptions of the month.\n\nWe hope that you found this update helpful. If you have any further questions, please don't hesitate to reach out to our support team.]" time="0.211"><properties><property name="score" value="0.9994052648544312" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The Canonical Path for Classes\n\nLast Updated on Wed, 04 Jan 2016 | Excel 2003 VBA\n\nWhen you create a class module, you must specify the following items in the declaration:\n\n\u25a0 The class name\n\n\u25a0 The objects that are instances of the class\n\n\u25a0 The properties that the class will support\n\n\u25a0 The methods that the class will support\n\n\u25a0 The events that the class will support\n\nAll classes in Excel derive from a generic base class called Object, which is an instance of the class of the same name. The Object class is an abstract class and doesn't have a constructor. If you don't have any idea what I'm talking about here, don't worry about it for now. I'll explain it in detail in Chapter 5. You can't instantiate an object from the Object class, so the only real purpose of the Object class is to serve as a base class for other classes. You can think of the Object class as being like a template that all other classes derive from. If you want to learn more about how classes work and how you can customize Excel's classes to create your own customized classes, check out my Microsoft Press book Microsoft Excel VBA and Macros (ISBN 0-7356-1845-0).\n\nWhen you create a class module in your Excel application, you must specify the class name as a string literal. For example, if you create a class module called Class1, Excel uses the following syntax in the declarations area of the class module to specify the name of the class:\n\nPublic Class Class1 End Class\n\nAs with any other module in Excel, a class module doesn't do anything until you create an object from the class module. Creating an object is a two-step process: You first instantiate the object, and then you call the methods and properties of the object. For example, to create an object from the class named Class1, you use the following code:\n\nDim obj1 As New Class1\n\nOnce you create an instance of the object, you can call the methods and properties of the object by preceding the method or property name with the name of the object. For example, if you created an instance of the class named Class1, you would use the following code to call the method named Display:\n\nobj1.Display\n\nIf the object contains an object variable that represents another object of the same class, you can also call the methods of that object. For example, if the object named obj1 created in the preceding code contained an object variable named obj2 that represented another object of the same class, you would use the following code to call the Display method of the object named obj2:\n\nobj1.obj2.Display\n\nThe property and method names of a class are preceded by the name of the class, as shown in the preceding examples.\n\nIf you create a class module that contains a custom collection object, such as a collection of cells in a worksheet, you must use a new operator to create the collection object. For example, the following statement creates a collection object named myCol1 that is a collection of worksheet cells:\n\nDim myCol1 As New Collection\n\nWhen you create a collection object, you specify the type of object that the collection can contain when you define the type of the collection.\n\nFor example, the following statement creates a collection object that can contain only the cells in a worksheet:\n\nDim myCol1 As New Collection(Worksheet.Range)\n\nWhen you create a collection object, you can add items to the collection object by using the Add method of the collection object.\n\nFor example, the following statement adds a range named &quot;Region&quot; to the collection object named myCol1:\n\nmyCol1.Add mySheet.Range(&quot;Region&quot;)\n\nThe Add method takes two arguments:\n\n\u25a0 The object that you want to add to the collection object\n\n\u25a0 The collection that you want to add the object to\n\nWhen you create a collection object, you can add an object to a collection object that is an instance of a different class, but the type of the objects in the collection must be the same.\n\nThe following code example adds the range named &quot;Region&quot; to the collection object named myCol1. This code example assumes that you've created the objects in the declaration section of the class module.\n\nSub AddObjectToCollection() Dim myCol1 As New Collection(Worksheet.Range) Dim myRow As Long, myColumn As Long Dim strName As String ' Obtain a reference to the collection ' that contains the first worksheet in ' the active workbook. Set myCol1 = Sheets(1).Range(1, 1). _ SpecialCells(xlCellTypeVisible) ' Initialize the name of the cell ' to the value of the Name property. strName = myCol1.Cell(1, 1).Value ' Determine the address of the cell ' and store it in the myRow and myColumn ' variables. myRow = myCol1.Cell(1, 1).Address myColumn = myCol1.Cell(1, 2).Address ' Create a new Range object. Set myRange = Worksheets(&quot;Sheet1&quot;). _ Range(&quot;Region&quot;) ' Add the new range to the collection. myCol1.Add myRange End Sub\n\nThe following code example calls the AddObjectToCollection method to add the range named &quot;Region&quot; to the collection object named myCol1. This code example assumes that you've created the objects in the declaration section of the class module.\n\nSub AddObjectToCollectionWithMethodCall() Dim myCol1 As New Collection(Worksheet.Range) Dim myRow As Long, myColumn As Long Dim strName As String ' Obtain a reference to the collection ' that contains the first worksheet in ' the active workbook. Set myCol1 = Sheets(1).Range(1, 1). _ SpecialCells(xlCellTypeVisible) ' Initialize the name of the cell ' to the value of the Name property. strName = myCol1.Cell(1, 1).Value ' Determine the address of the cell ' and store it in the myRow and myColumn ' variables. myRow = myCol1.Cell(1, 1).Address myColumn = myCol1.Cell(1, 2).Address ' Create a new Range object. Set myRange = Worksheets(&quot;Sheet1&quot;). _ Range(&quot;Region&quot;) ' Add the new range to the collection. myCol1.AddObject myRange ' Add another range to the collection ' and display the number of elements in the collection. myCol1.Add myRange MsgBox myCol1.Count End Sub\n\nThe final statement in the AddObjectToCollectionWithMethodCall subroutine is a good example of how you can use a method that belongs to a class to call the methods of the collection object.\n\n0 0]" time="3.545"><properties><property name="score" value="0.7163082261880239" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Title\n\nInsurance Leads\n\nDescription\n\nWe are a leading Insurance Agency in Central Ohio offering all types of insurance. Our insurance agency offers insurance services for all types of customers from those who have been with the same insurance agency for many years to new customers that have not been with an insurance agency before. Our goal is to provide outstanding customer service and the best insurance services to our customers.\n\n\n\nPlease visit our website at http://www.affordableinsuranceservices.net/\n\nKeywords\n\nBusiness Name: Affordable Insurance Services\n\nSearch Phrase: Affordable Insurance Services\n\nLocation\n\n1 Water Street, Delaware, OH 43015, USA\n\nCategory\n\nBusiness Name\n\nAddress\n\n1 Water Street, Delaware, OH 43015, USA\n\nContact Information\n\nPhone:\n\nUrl:]" time="0.418"><properties><property name="score" value="0.9998266100883484" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[India has seen a significant spike in energy demand in recent years, driven by economic growth, rapid urbanisation and the increasing electrification of the country. Energy production has increased over the same period, but the country\u2019s domestic energy supply has not kept pace. The growth of energy demand outpacing domestic production resulted in a significant reduction in oil and natural gas stocks. The country became increasingly dependent on imported energy.\n\n\n\nDomestic energy production is also likely to experience a plateau]" time="0.251"><properties><property name="score" value="0.508491039276123" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.50849104&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.50849104
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Many people who oppose the minimum wage assume that raising the minimum wage is bad for the economy and will cause a lot of unemployment. This is mostly false. A new study has revealed that raising the minimum wage does not kill jobs. This is not just good news for the people who will now get a raise, but also for the economy as a whole.\n\nOpponents of the minimum wage often say that it will create a lot of unemployment and that this is proof that the minimum wage should be abolished. They argue that it creates a job loss on an economy-wide scale. However, this new study has proved that the minimum wage will not cause massive job loss. This is not just great news for the people who get a raise, but it also means that the economy will benefit.\n\nThe study comes from the National Employment Law Project (NELP) and reveals that raising the minimum wage does not cause massive job loss. In fact, only two percent of the workforce is impacted by a minimum wage increase. This means that the economy will benefit, but not that much. However, it is still a positive impact.\n\nHowever, the study also shows that the country would have to raise the minimum wage by $2.13 to achieve the same number of jobs as a similar increase in the 1990s. This is not really a surprise, as the economy was growing more rapidly back then.\n\nRaising the minimum wage was also seen to boost the economy. The study found that a $1 increase in the minimum wage would increase the GDP by $2.62. This is a significant amount and would be great for the economy as a whole. In fact, it could even boost the economy so much that it could lift more people out of poverty.\n\nWhile a raise in the minimum wage does not have a massive impact on the economy as a whole, it still has a positive impact. This means that the people who will now get a raise will enjoy it. This will also help the economy, and make the country a better place to live in.\n\nThose who oppose the minimum wage should know that it does not create massive unemployment. This is also great news for the economy, and it means that raising the minimum wage is great for everyone.]" time="1.063"><properties><property name="score" value="0.9100094040234884" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[\u201c\n\nI have no hesitation in recommending you to the body of workers in Australia, either in Railway or Country Tramways, or Electrical construction. Your outfit has been constantly at work for nearly three years and has been found fully to justify my own and your own high recommendation of it, both for its efficiency and economy of working. The Cableways I have had in operation for two years are found to work in the most satisfactory manner, both in operation and maintenance, in fact I do not know of any system that I would prefer for tramways, and I am glad that I have them.\n\nJ.R. Lucas, Chief Engineer,\n\nQueensland Railways\n\nNashua, New Hampshire, August 20, 1898]" time="0.429"><properties><property name="score" value="0.9902728199958801" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[READ MORE Wall Street: Online poker rooms are fading, but who is 6/10/2018 \xb7 As online poker room software and banking options face an uncertain future in the U.S., many of the sites that are still online are trying to\n\nREAD MORE Exclusive: Where do U.S. poker players play online poker 7/10/2018 \xb7 With the April 15th deadline for Nevada-based online poker rooms to stop offering real-money games, where will the displaced players go next?\n\nREAD MORE Online poker room ratings, reviews, and deals - Poker In the wake of Black Friday, you can\u2019t walk three feet in a Las Vegas casino without bumping into someone who\u2019s looking for a blackjack\n\nREAD MORE Nevada Online Poker Rooms On Verge Of Exiting Market Online poker rooms in the United States are in a precarious position right now. They still have a month left to get their affairs in order,\n\nREAD MORE What Happens To U.S. Poker Players Now? - Card Player 4/3/2019 \xb7 For a site to be eligible for inclusion on the Upcoming US-facing Online Poker Sites list, it must be confirmed to launch in the United States. This means\n\nREAD MORE Who are the best US poker sites in 2019? | Best US poker 3/11/2019 \xb7 The only active and legal online poker room in the U.S. is currently the 888 Poker. The American online poker has not gone anywhere since Black Friday,\n\nREAD MORE Gambling in the United States - Wikipedia Online Poker Reviews - Top Rated Online Poker Sites for 2018. Online poker rooms in the United States are not as popular as in other countries due to\n\nREAD MORE Online Poker Sites in the United States, All Active and Are Online Poker Rooms Still Profitable? Find out more about the best online poker rooms in the United States today.\n\nREAD MORE Poker News &amp; Articles: Read Online &amp; Watch the Live Videos A ranking of the best online poker rooms in the United States, including casino sites for U.S. players and trusted online poker rooms with\n\nREAD MORE Casino Game Reviews - Top Rated Online Poker Sites for 5/22/2016 \xb7 Though not as big as the gambling industry in other parts of the world, the United States does offer a number of legal gambling options to its\n\nREAD MORE Online Poker News - Latest News from Online Poker Rooms 2018 Online Poker Sites 2018 is a collection of the best online poker rooms available today for US players. You can find all of the important info\n\nREAD MORE Where are the best US online poker rooms right now? 7/17/2017 \xb7 The legal US online poker landscape could be getting a little bit clearer, as Nevada has once again made a bill that would create an online poker licensing\n\nREAD MORE Online Poker News &amp; Articles - Poker News | Upswing Poker Blog 4/29/2015 \xb7 Online poker in the U.S. has long been an industry in limbo, with the legality of the business depending on how you define it. Now the U.S.\n\nREAD MORE Las Vegas Poker Rooms, Poker Tournaments and Game The list of best US poker sites in 2019 includes the latest online poker rooms that cater to United States-based poker players. Read about the top U.S. poker\n\nREAD MORE Online Poker in the U.S. \u2013 The Ultimate Guide | Upswing Poker Online poker room reviews featuring top sites for US players. Exclusive poker bonuses, promotions and expert reviews.\n\nREAD MORE Online Poker Guide for U.S. Poker Players - US Poker Guide 4/29/2015 \xb7 Online poker is now more popular in the U.S. than in Europe. In the U.S., state laws and gambling regulations differ by state, and some states have\n\nREAD MORE Vegas Palms Online Casino, Online Poker &amp; Online Slots Join PokerStars and enjoy top quality online poker. New players get a 100% match bonus up to $600. Play on Windows PC and Mac with our instant download client.\n\nREAD MORE Gambling in the United States - Wikipedia The U.S. online gambling market is the largest in the world. By some estimates, the U.S. accounts for one-third of all online gaming revenue.\n\nREAD MORE Online Poker in the U.S. \u2013 The Ultimate Guide | Upswing Poker Online Poker Guide for U.S. Poker Players. If you are a poker player looking to play online poker in the United States, here you will find the latest\n\nREAD MORE Online Poker and Gambling Laws for 2018 - Casinolistings.com 4/15/2017 \xb7 In the wake of Black Friday, the number of online poker rooms in Nevada dwindled from six to just one, and the number of players dropped\n\nREAD MORE Poker Players on a Dime: What US Online Poker Rooms Are 2014 Online Poker Sites - US Online Poker Guide. List of Top USA Poker Sites accepting U.S. Players including all the latest bonuses and promotions.\n\nREAD MORE Online Poker USA \u2013 Online Poker Sites for US Players A brief introduction to online poker and online gambling in the United States.\n\nREAD MORE USA Online Poker Guide - UsaPokerOnline.com 5/12/2019 \xb7 &quot;We are going to be online, we are going to be regulated,&quot; said Robert Capecchi, director of federal legislation at the Marijuana Policy Project, who led a\n\nREAD MORE Online Poker and Gambling Laws for 2018 | GamingLawOnline.com Gambling in the United States has been legal in each of the 50 U.S. states since 1911, except for a brief prohibition after World War I. At the\n\nREAD MORE 5 Must-Visit Online Poker Rooms for U.S. Players 4/15/2017 \xb7 Exclusive: Where do U.S. poker players play online poker now? We asked, and you answered. We asked, and you answered.\n\nREAD MORE Online Poker in the US in 2017 - Online Poker Guides, Tips U.S. online poker has a dark history, littered with fraudulent players and hacked accounts. This guide will help you find a safe place to play.\n\nREAD MORE Online Poker Rooms - ThePokerBank.com Nevada-based poker rooms have until April 15th to exit the US market or risk being shut down. Players are looking for a new online poker home.\n\nREAD MORE Online Poker Room Rankings - ThePokerBank.com Learn more about top online poker sites and poker rooms in the United States.\n\nREAD MORE USA Online Poker Guide | Play Poker Online in the US U.S.A. Online Poker Sites. A list of all the best USA online poker sites, casino sites, and sportsbook sites with information on bonuses, games, and more.\n\nREAD MORE USA Online Poker Guide: Reviews, News, &amp; Tips - Worldwide 7/16/2015 \xb7 The status of online poker in the United States is likely to remain a confusing mess for a]" time="3.678"><properties><property name="score" value="0.703465461730957" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Borodino (village), Moscow Oblast\n\nBorodino () is a village (&quot;selo&quot;) in Odintsovsky District of Moscow Oblast, Russia, located on the Tvertsa River.\n\nThe village of Borodino was first mentioned in 1611 as part of the Tsaritsyn &quot;oprichnina&quot;.\n\nIn 1812, the Battle of Borodino, a major battle of the French invasion of Russia took place near Borodino. The village is now a site of the museum and the memorial. The French lost the battle, but went on to win the war in 1814.\n\nIn 1898, Borodino was the place where Nikolai Basargin and Nikolai Morozov organized a commune and founded a school of commune life that lasted from 1898 to 1902.\n\nIn the middle of the 20th century Borodino was also the place where Andrey Voznesensky had a summer dacha.\n\nThere is a railway station in Borodino.\n\n\n]" time="0.544"><properties><property name="score" value="0.9998043179512024" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Michal Ha\u0161ek\n\nMichal Ha\u0161ek (born April 8, 1980) is a Czech professional ice hockey goaltender who is currently playing with HC Pardubice of the Czech Extraliga. He previously played in the National Hockey League for the Detroit Red Wings, Chicago Blackhawks, Colorado Avalanche, Ottawa Senators, and Florida Panthers.\n\nHa\u0161ek spent the 1996\u201397 season with HC Pardubice's junior team, and then joined the Czech Extraliga club HC Vset\xedn, appearing in 29 games. Ha\u0161ek was a member of the Czech Republic national under-18 team at the 1997 European Junior Championships, playing in two games and posting a 1.51 GAA.\n\nThe next season, Ha\u0161ek joined HC Pardubice, which was then playing in the Czech 1. Liga. Ha\u0161ek played well for the team, and was chosen by the Colorado Avalanche in the sixth round, 162nd overall, in the 1998 NHL Entry Draft. He remained with HC Pardubice, and in his final season with the club, he posted a 2.30 GAA in 37 games. Ha\u0161ek was signed by the Avalanche, and on March 9, 2000, he made his NHL debut, against the Detroit Red Wings. He played two more games for the Avalanche before returning to the Czech Republic to join HC Pardubice for the 2000\u201301 season.\n\nHa\u0161ek played in 11 games for the Avalanche the next season, but spent the majority of the 2001\u201302 season with the Hershey Bears of the American Hockey League, where he posted a 2.20 GAA in 44 games. He remained in North America the next season, playing five games for Colorado and 59 for Hershey, and was then traded to the Chicago Blackhawks for a sixth-round pick in the 2004 NHL Entry Draft on March 10, 2003. Ha\u0161ek played well for the Blackhawks in 2003\u201304, with a 2.11 GAA in 37 games.\n\nThe next season, Ha\u0161ek played well for the Blackhawks, with a 2.22 GAA in 36 games. However, he was unable to maintain his form, and was replaced as Chicago's starting goaltender by Nikolai Khabibulin, who had been acquired by the Blackhawks in a trade with the Tampa Bay Lightning. Ha\u0161ek was then traded to the Ottawa Senators on March 9, 2005, for a second-round draft pick in the 2005 NHL Entry Draft.\n\nIn his first season with the Senators, Ha\u0161ek played in 40 games, and posted a 2.26 GAA. He then missed the majority of the 2005\u201306 season after suffering a shoulder injury. He appeared in just five games for the Senators, and was traded to the Colorado Avalanche for Patrick Eaves, a second-round draft pick in the 2006 NHL Entry Draft, and a third-round pick in the 2007 NHL Entry Draft on September 25, 2005.\n\nHa\u0161ek quickly returned to form with the Avalanche, playing in 54 games and posting a 2.22 GAA. In 2006\u201307, he played in 57 games, and posted a 2.30 GAA. Ha\u0161ek was an important factor in the Avalanche's playoff run, as he posted a 1.65 GAA in eight games.\n\nOn June 19, 2007, Ha\u0161ek signed a two-year contract extension with the Avalanche. The following season, he played in just 35 games due to a groin injury, and his GAA rose to 2.88. He was then sidelined for the rest of the season after having shoulder surgery.\n\nOn June 30, 2009, Ha\u0161ek signed a two-year contract with the Detroit Red Wings. He recorded the first win of his Red Wings career on January 2, 2010, in a 3\u20131 victory over the Chicago Blackhawks.\n\nOn January 1, 2011, Ha\u0161ek was named the First Star of the Week for the week ending December 26. Ha\u0161ek compiled a 3\u20130\u20130 record, a 0.67 goals-against average and a .984 save percentage. He led the Red Wings to their fifth four-game winning streak of the season. During the streak, Ha\u0161ek stopped 135 of 138 shots and yielded only one goal in a span of 140 minutes and four seconds.\n\nOn January 20, 2011, Ha\u0161ek was selected as the NHL's Third Star of the Week. He went 2\u20130\u20131 in three appearances, with both victories coming against Central Division rivals (January 14, 3\u20130 win at Chicago and January 16, 4\u20131 win at St. Louis), and recorded a 1.00 goals-against average and a .962 save percentage.\n\nHa\u0161ek finished the 2010\u201311 season with a career-high 41 victories, just one win behind league leader Martin Brodeur.\n\nOn March 29, 2011, Ha\u0161ek scored a goal against the Los Angeles Kings. The goal, a result of a breakaway due to an errant pass by Kings defenseman Slava Voynov, was the second of Ha\u0161ek's career. His first goal was scored against the Tampa Bay Lightning in 2008. The goal also made him the oldest goaltender in NHL history to score a goal in a game. The record was previously held by Tommy Salo, who scored a goal in a game at the age of 35.\n\nOn May 31, 2011, Ha\u0161ek underwent surgery to repair a torn labrum in his right shoulder.\n\nHa\u0161ek was injured during the 2011\u201312 preseason, and on October 2, 2011, was placed on long-term injured reserve by the Red Wings. He returned from his injury to back up Ty Conklin for a game against the Phoenix Coyotes on November 29, 2011.\n\nOn July 5, 2012, Ha\u0161ek signed a one-year, $2.05 million contract with the Florida Panthers. He was named the starting goaltender in the Panthers' inaugural game in the 2012\u201313 season against the New Jersey Devils. He was subsequently pulled in the first period after allowing 4 goals on 15 shots. He was replaced by Scott Clemmensen, and on January 24, 2013, Clemmensen was named as the Panthers' starting goaltender for the remainder of the season.\n\nHa\u0161ek was placed on waivers by the Panthers on June 13, 2013. Ha\u0161ek was then bought out from the final year of his contract on June 15, 2013. He was then signed by the Detroit Red Wings for the 2013\u201314 season to backup starter Jimmy Howard.\n\nHa\u0161ek became an unrestricted free agent in the off-season, and on September 11, 2014, signed his first contract abroad, agreeing to a one-year deal with HC Pardubice of the Czech Extraliga. After one season in Pardubice, Ha\u0161ek opted to continue in the Extraliga, signing a two-year deal with HC V\xedtkovice. In his second season with V\xedtkovice, Ha\u0161ek helped his club claim its first league title since 1999, appearing in 15 playoff games. He left as a free agent to sign a one-year contract with HC Pardubice on June 19, 2017.\n\nHa\u0161ek played for the Czech Republic at the 2002, 2003, 2004, and 2005 World Championships, winning bronze in 2004 and 2005. Ha\u0161ek was the starting goaltender for the Czech Republic team that won the gold medal at the 2004 World Cup of Hockey.\n\nIn 2003, Ha\u0161ek became a Canadian citizen, but had already represented the Czech Republic in international competition, and he was therefore eligible to play for either national team. On November 7, 2005, he announced that he would continue to represent the Czech Republic, even though he had been approached by the Canadian team and was eligible to play for Canada. Ha\u0161ek's younger brother Ond\u0159ej also plays professional hockey.\n]" time="4.019"><properties><property name="score" value="0.9473894323621478" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Helium leak\n\nA helium leak is a leak of helium from its container.\n\nHelium is a diatomic gas and therefore more difficult to confine than many other gases. The inert gas may diffuse through even very small leaks into other parts of a vacuum system, or it may leak into the atmosphere. Diffusion is facilitated by higher pressure on one side of a barrier. One way of countering this effect is to use helium's very low solubility in common materials. Helium also may leak into vacuum systems from poorly-made, improperly-applied, or damaged seals. \n\nHelium leak detectors use a small electric current through a coil inside the container to detect leaks. They are often found in high-vacuum and ultra-high-vacuum systems.\n\nAt room temperature, helium leaks through most materials at a rate of about one molecule every 10 seconds. If a]" time="0.424"><properties><property name="score" value="0.9993553757667542" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Cons\n\nCompany uses a very low standard of ethics in most cases. They do not care about employees, or helping them improve their lives. They have some of the poorest HR departments in the industry. They have a very slow turnover, and the company is very willing to allow people to be mistreated for long periods of time before taking action.]" time="0.201"><properties><property name="score" value="0.9972398281097412" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[$46,924\n\n$23/hr\n\n$2,533\n\nThe average publicity announcer salary in Alabama, United States is $46,924 or an hourly rate of $23. This is 11% lower (-$5,552) than the average publicity announcer salary in the United States. In addition, they earn an average bonus of $2,533.]" time="0.201"><properties><property name="score" value="0.9994416832923889" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99944168&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99944168
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Bitcoins are widely considered a great way to make purchases, but it is much harder to pay someone with them. That is exactly the case with Andreas Antonopoulos. He has reportedly been bombarded with requests from individuals who want to make donations to his projects. But he has no means of receiving them.\n\nIs Andreas Antonopoulos Being Misled?\n\nMost people know Andreas Antonopoulos by now. He is one of the most popular and prolific figures in the Bitcoin world. However, the widespread adoption of Bitcoin has led to a lot of new users, many of whom do not have a great understanding of the technology. That is not necessarily a bad thing, though.\n\nIt now appears as if a lot of people think Andreas Antonopoulos accepts donations, whereas he doesn\u2019t. As such, people keep bombarding the Bitcoin community icon with messages asking how they can make a donation. This leads to quite a few jokes being made about it on social media. However, Andreas Antonopoulos is well aware of the situation, as he keeps getting the same questions.\n\nHowever, the most intriguing part of this situation is how Andreas Antonopoulos makes it clear he doesn\u2019t accept donations. For some reason, some people continue to pester him about it. Some individuals even claim they are only able to donate a small amount of money, which would make no sense whatsoever.\n\nThe Andreas Antonopoulos Donation Account\n\nInterestingly enough, Andreas Antonopoulos is quite open about Bitcoin donations, but only to the extent of signing them over to projects. That is not what most people would expect, but it makes a lot of sense to give this money to a good cause. For example, Antonopoulos has donated over US$5,000 worth of Bitcoin to help rebuild the war-torn Marawi in the Philippines. It is evident this money came from donations he received, which is a very interesting development.\n\nThat being said, donations can still be made to Andreas Antonopoulos, even though it is done in a different way. Instead of sending money to his personal wallet, a public Bitcoin wallet has been created for donations to be sent to. That wallet has been active for some time now, and users can send money to a wide range of Bitcoin and Bitcoin Cash wallets.\n\nDo you think donations should be sent to Andreas Antonopoulos? Let us know in the comments below.\n\nImages courtesy of Shutterstock\n\nread original article here]" time="1.158"><properties><property name="score" value="0.8956253925959269" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Rottnest Island\n\nThe second island in the chain, after its neighbours Fremantle and Rottnest, is Barrack Street. In 1831, a convict ship called the \u201cDuke of Wellington\u201d (there are other tales about the origins of the name) ran aground here and was eventually abandoned. The women were moved to Fremantle Prison. The wreck is still there, but it is being eaten by the sea. You can walk to it on a path from Waterhouse Beach. The best way to see the island is by bike, though, because it is so small and flat. The bikes are provided by a company called Island Experience Rottnest. This is a great way to see the island. It costs around $25, which gets you a bike and the unlimited use of it for three hours. They give you a map and instructions, but you can see the whole island by walking in about an hour.\n\nThere are very few buildings here, but what there are have a very interesting history. For example, there is the Weeroona Hotel, the oldest operating hotel in the state. It is run by a company that is owned by the island\u2019s inhabitants. The oldest building is called the Round House. It was built in 1833 and has been restored to the 1830s period. It used to be the men\u2019s quarters. There is also a small cathedral that was built in 1901, and is called St Mary\u2019s Cathedral. There is also a cricket ground here that was originally established in the 1840s. This island was once the penal colony. The ruins of two of the prisons, the Governor\u2019s residence and the convicts\u2019 huts can be seen. They are located at a place called Settlement.\n\nCamel Rock\n\nThis is the most important landmark of the island. It is visible for many miles, and is sometimes mistaken for a camel. The sandstone is interesting, as it was formed from quartzite, which is a very hard stone. Some of the grains have been turned to iron oxide, which is a reddish brown. This was done by chemical reactions that were the result of oxygen and iron.\n\nSome rock formations can also be seen at the Quobba station, which is a restored 19th century railway station. There are a few lookouts here, but the best is at Arch Rock, which is a short distance from the main road. There are about 200 species of birds here, and there is a sanctuary that is managed by a conservation group called the Rottnest Island Conservation Society. This is home to the Tawny Frogmouth, the Black-breasted Button-quail, the Malleefowl, and the Rottnest Island White-tailed Black-Cockatoo.\n\nThere is also an Interpretive Centre that explains the history of the island and its different birds. Another reason to visit this island is the fact that the Aboriginal people, the Wadjuk, were here first. There is an Aboriginal reserve that is in the northern area of the island, and there is also a sacred site that can be seen there. The site is called Yanjul, which means \u201cman with spear\u201d. The local Aboriginals still refer to the island as Wadjemup, which means \u201cIsland of Peace\u201d. The First Fleet stopped at this island, and as such it is the only island in the world where two different nations fought for control.\n\nPlaces to Stay\n\nThere are a few different places to stay on this island. Most of the accommodations are self catering. There are also a few bed and breakfasts, and there is even a four star hotel called the Sheraton Rottnest Island Resort. There are over 250 different cottages to stay at, some of which have been converted from old army barracks. The Great Western Hotel is the only hotel here that has electricity and telephone connections, so you may want to stay there if you want to be near the center of things. There are many hotels here, but most of them are more expensive than the average.]" time="2.134"><properties><property name="score" value="0.9997844696044922" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[This great vinyl adhesive cloth tape is waterproof, flexible and conformable to curves. It features easy-to-remove release liner and 1,000 per roll. Use this tape to adhere vinyl cloth to flat or curved surfaces.]" time="0.170"><properties><property name="score" value="0.9489731788635254" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[I can't believe it. Yesterday, I got a letter from Robert D'Orta, the writer who's suing Paramount for copyright infringement over the Star Trek reboot. He wants to settle. And I've got to say, my heart just jumped. A settlement means that Paramount's going to have to pay some kind of damages, and I get to share in it. Maybe I get a check, maybe I get some percentage of whatever they owe to Robert. I can't believe it. That one line. &quot;Paramount's counsel has been in contact with me about settlement.&quot;It just keeps me going. It's the first time since this all started that I can say for sure that I'm not just some crazy person running after a scam. Robert D'Orta and I were going to be in the history books.I can't believe it. I'm really going to have to look at a settlement now. Who do I want? What's it going to take? What are the chances of me getting something from this?What's next?]" time="0.494"><properties><property name="score" value="0.7534622550010681" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[So, I never thought this day would come, but after several years of wrestling with this series, I have finally been forced to give up on it.\n\nI read The Fellowship of the Ring many years ago, and loved it. It was one of the books that inspired me to get into fantasy fiction. The author, J.R.R. Tolkien, was a legend in the world of fantasy fiction, and it was considered one of his greatest works.\n\nAt the time, I hadn't heard of this series. If I had, I probably would have thought &quot;Ew, it's another fantasy book.&quot; It would take me several years to learn that it's not another fantasy book. This is the one that started it all.\n\nThe story was great, and I really liked how Tolkien put all his imagination into it, and wrote a story that was unique and innovative. The story itself was fascinating. I loved the idea of Hobbits, which are little people that lived in an idyllic area, with their own language and customs. This book had a dark side, and even the humor in it was dark. There was a lot of death, evil, and destruction in the book. That's something that stuck with me, even though I really didn't appreciate it at the time. I was in my tween years, and just didn't understand it.]" time="0.654"><properties><property name="score" value="0.9906856715679169" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Reviews (15) Write a Review\n\nCool costume Perfect fit and great quality. Looks just like the picture. Happy with purchase\n\nDoesn\u2019t fit like the other girls ones! The body piece is way too long for a normal sized girl, there\u2019s so much extra material, you can barely make it look like the girl in the picture. It doesn\u2019t look as cute as the other girls ones, and the extra length isn\u2019t easy to shorten, because of the netting on the bottom.\n\nFun! Love this costume! My daughter is small and the medium was perfect! Can\u2019t wait to see her in it at school on Halloween!\n\nBeautiful! This costume is beautiful! It fits true to size and looks great!\n\nGorgeous costume I love this costume and it fits true to size. I have a slender 7 year old, ordered the size 6-8, and it fits well with some room to grow into. The only drawback I see is the flippers are sewn into the gloves so she won\u2019t be able to wear them separately, but that\u2019s a small sacrifice for the beauty of this costume!\n\nCutest! Love it! My 5 yr old loves her new swimmer costume. It's made well, super cute and just the right amount of &quot;bling&quot;!\n\nSO STINKING CUTE! This costume is exactly what I was looking for! It\u2019s super cute and fits well. My daughter is a 4t and I got her a 5-6 and it fits great! It\u2019s also not super thin. It\u2019s very well made!\n\nWonderful costume! Beautiful! Super cute and made well.\n\nExcellent quality! I\u2019m always apprehensive when ordering on-line, but was very pleased with the quality of this costume!\n\nBeautiful This costume is so beautiful! The colors are vibrant and fun, the quality is fantastic and the fit is perfect!\n\nSo pretty I bought this for my 4 year old daughter and she loves it! She is tiny so I ordered her a 4-6 which fit perfectly. It's well made and super pretty. Can't wait to see her wear it for Halloween\n\nadorable It is very pretty. The top is pretty sturdy. The bottom is light. It fits my 4 year old perfectly. My daughter is little and will not fit for very long. I love it!\n\nsuper cute This costume is so cute! I ordered the size 4-6 for my daughter who wears a size 4 in girl clothes and it fit perfectly. The material on the body piece is thick and the arms and feet are easy to get on and off, without being too loose.\n\nVery pretty I am very pleased with this costume. The colors are so vibrant and the quality is good. The size fits perfect.]" time="1.571"><properties><property name="score" value="0.979120671749115" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[2013 Crime Rate Indexes North Little Rock, AR 72223 Arkansas United States Total Crime Risk 170.0 91.0 100 Murder Risk 165.0 94.0 100 Rape Risk 164.0 90.0 100 Robbery Risk 157.0 93.0 100 Assault Risk 162.0 92.0 100 Burglary Risk 155.0 89.0 100 Larceny Risk 105.0 92.0 100 Motor Vehicle Theft Risk 132.0 89.0 100\n\nThe data for North Little Rock, AR 72223 may also contain data for the following areas: North Little Rock\n\n\n\nCrime Risk Index (100 = National Average): Index score for an area is compared to the national average of 100. A score of 200 indicates twice the national average total crime risk, while 50 indicates half the national risk. We encourage you to consult with a knowledgeable local real estate agent or contact the local police department for any additional information.\n\n\n\nCrime Indexes are based on numerous current and historical datasets as well as proprietary modeling algorithms which estimate values at more granular geographic levels when specific data is either unavailable or impractical to aggregate. While every effort is made to ensure accuracy, these are estimates and should only be used as a guide. For detailed information regarding crime and safety in a community, please contact local law enforcement agencies.]" time="0.716"><properties><property name="score" value="0.9133971929550171" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.91339719&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.91339719
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Jonathan Daniel/Getty Images\n\nAccording to a source, the Miami Heat has signed Mike Miller to a two-year, $5.5 million contract, and may be on the verge of completing a deal with Rashard Lewis for the veteran minimum.\n\nThe Miller signing comes as the Heat are preparing to play the Chicago Bulls in Game 4 of the Eastern Conference quarterfinals.\n\nMiller averaged a career-low 4.8 points and 1.8 rebounds in just 22 games during the regular season.\n\nHe did not play in Miami\u2019s first-round series against the Milwaukee Bucks, after undergoing surgery to repair a torn tendon in his left foot.\n\nHe played in just two games during the second round series against the Indiana Pacers.\n\nThe Heat's move is somewhat surprising because they're already stacked at the wing position, with LeBron James, Dwyane Wade, James Jones, Mike Miller, and James Jones.\n\nMiller will be limited to roughly 25 minutes a game, according to a source.\n\nMiller has won two NBA titles with the Heat.\n\nHe was the MVP of the NBA Finals in 2006, and has averaged 10.7 points and 4.5 rebounds for his career.\n\nThe Heat are in need of frontcourt depth, because they only have three true big men on the roster, and could use more depth, with Greg Oden having his contract purchased by the team, but not likely to play during the playoffs.]" time="0.642"><properties><property name="score" value="0.9996192455291748" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Riot seems to be slowing down its champion creation process. Since the release of the current champion rotation, only five champions have been released. Each of these champions (except for Xerath) was tested in the PBE at some point, so we knew what to expect when the champion was released.\n\nXerath was the next to last champion to be released. He came out in the middle of the season and, because of his kit, has been popular for quite a while. He didn't really have that many people against him, and it didn't take him long to climb into the current meta. He was picked a total of 11 times in the North American Championship Series and showed a 65 percent win rate.\n\nZed, like Xerath, was released at the beginning of the season, but there are still people who have not really mastered him yet. His popularity in the NA LCS was fairly low and his win rate was only 58 percent. Zed's popularity has since risen to the top and he is now played regularly in the NA LCS.\n\nBrand is another interesting case. His play rate has risen from 0 percent in the NA LCS to 3 percent. His win rate is not that great (44 percent), but because he is a burst caster, he fits in with the current meta. As long as you have the ability to lock down one person for a period of time, Brand is pretty strong.\n\nFinally, Vi was the last champion to be released, but her popularity is still a bit low. She has been played nine times in the NA LCS, with a 43 percent win rate.\n\nWith those three champions, Riot has shown that it will not just dump out new champions on the game, and that champions will still have to be tested on the PBE before being released.]" time="0.969"><properties><property name="score" value="0.9995488226413727" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Photo by Gracie Malley for Cannabis Now\n\nAs California rolls out its new cannabis rules \u2014 the most relaxed marijuana laws in the country \u2014 state and local law enforcement officials are attempting to implement the new system with an unprecedented blitz of raids on illegal pot businesses.\n\nWith the stroke of Gov. Jerry Brown\u2019s pen, California became the largest state to legalize marijuana for recreational use and the sixth in the nation to make cannabis available to the general public. Yet, the history of cannabis in California has been a long one \u2014 and not without its rocky moments.\n\nCalifornia\u2019s medical marijuana laws were first approved by voters in 1996 and ushered in a sea change in cannabis culture and use. Yet even in the heyday of Proposition 215, there were raids on medical dispensaries, law enforcement shutting down shops for violation of local zoning codes and municipal zoning laws, local municipalities levying taxes on dispensaries and even offering up community space in which to open shops.\n\nThe result of these early legislative setbacks? The displacement of a long-established network of dispensaries and cultivators. Those who could no longer get by in the state moved to Oregon, Washington, Colorado and other legal states \u2014 essentially exporting the cannabis culture they had created. Now, a new generation of cultivators and distributors are starting businesses in California, with new methods and processes that are a result of the new cannabis industry.\n\nEven as recreational marijuana use became legal in California in January 2018, there has been an uptick in raids of these newly legal businesses, which remains illegal under federal law. In the first week of 2018, there were a reported 26 raids of legal dispensaries in the Bay Area, with 25 arrests and the seizure of $16,000 in cash and seven pounds of cannabis.\n\n\u201cLaw enforcement is going to go after illegal operations in any case,\u201d said Sean Donahoe, executive director of the California Cannabis Industry Association. \u201cThese folks who are doing illegal activities, that\u2019s something they have been doing all along.\u201d\n\nIn California, the rules on commercial cannabis businesses are numerous and complex, he said. The fact that a business may be operating legally, that doesn\u2019t mean it isn\u2019t being targeted for a raid, according to Donahoe. In order for businesses to get permits to operate legally, they have to first get permits to be established.\n\n\u201cSome of the communities, their local government entities have decided]" time="1.164"><properties><property name="score" value="0.9976101915041605" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Hillary Clinton was a no-show at the opening of a new museum dedicated to women's history and equality in New York City on Tuesday, a day after she cancelled campaign events because of a stomach virus that left her dehydrated and dizzy.\n\nClinton's Republican opponent Donald Trump and his wife Melania did not attend the high-profile opening either.\n\nClinton's doctor said in a statement Monday that the Democratic presidential candidate had been &quot;diagnosed with pneumonia&quot; on Friday.\n\nThe museum is dedicated to women's history and feminist causes. Clinton's ties with the National Museum of Women in the Arts go back many years.\n\nShe presented the museum's founding patron, Mercedes T. Bass, with a leadership award in 2009. The museum's chair, Susan Henshaw Jones, has been a top Clinton donor.\n\nA spokesman for Clinton said on Monday that the former secretary of state &quot;continues to feel better,&quot; but he didn't indicate when she would resume campaigning.\n\nCampaign spokesman Nick Merrill said the Democratic presidential nominee has had an allergy-related cough, and that &quot;several hours of sleep&quot; and some antihistamines helped.\n\nThe illness struck just after Clinton flew back to New York City from a weekend trip to California.\n\nOn Monday, her doctor Lisa R. Bardack said in a statement that Clinton had had an examination of her pneumonia in a chest CT scan.\n\nBardack said the Democratic nominee &quot;is recovering nicely&quot; and has been &quot;advised to rest and modify her schedule&quot; as she continues to recuperate.\n\nClinton abruptly left a 9/11 anniversary ceremony Sunday in New York City after feeling overheated, according to her campaign. She retreated to her daughter's nearby apartment, but later emerged and told reporters: &quot;I'm feeling great. It's a beautiful day in New York.&quot;\n\nShe left Sunday's event at Ground Zero shortly after she arrived. Her campaign said she felt &quot;overheated&quot; and departed for her daughter's apartment in New York's Flatiron neighborhood.\n\nVideo captured by a bystander showed Clinton apparently struggle to get into a vehicle as staff members held her up by the arms.\n\nClinton, 68, did not immediately disclose the pneumonia diagnosis. Her aides initially said she had overheated.\n\nBardack said Clinton is &quot;re-hydrated and recovering nicely&quot; after attending a 9/11 anniversary event Sunday.\n\nThe health issue was not mentioned at a fundraising event Monday in Manhattan for Democratic Senate candidate Ted Strickland of Ohio, according to participants.\n\nAsked if the candidate's health was a campaign issue, spokesman Jesse Ferguson said, &quot;No.&quot;\n\n___\n\nAssociated Press writers Jonathan Lemire in New York and Ken Thomas in Washington contributed to this report.]" time="1.292"><properties><property name="score" value="0.3460589249928792" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.34605892&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.34605892
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The former CEO of St. Louis-based Boeing said he will plead guilty to charges that he approved the improper use of millions of dollars in company funds on home furnishings, a casino and a separate investor-relations matter.\n\nThomas W. Jones, 67, was charged with approving improper expenditures at the request of his predecessor as CEO, Philip M. Condit, 63.\n\nBoeing Co. fired both men after the company's internal investigation uncovered what prosecutors called &quot;improper and illegal conduct.&quot;\n\nProsecutors said that, in addition to Condit, four other Boeing executives and a Chicago interior designer would plead guilty to various charges related to the kickback scheme.\n\nJones and Condit each face up to 10 years in prison and fines of $250,000 on the conspiracy charge. The other defendants face up to five years in prison and $250,000 fines.\n\n'Systemic fraud'\n\nU.S. Attorney James A. Graves Jr. described the alleged conspiracy as &quot;systemic fraud.&quot;\n\n&quot;This was a company that had an ethical culture, and we find it deeply disturbing that a company like that could have employees commit such fraudulent acts,&quot; he said.\n\nCondit, Boeing's former chief operating officer, said in a statement that he agreed to plead guilty &quot;to the company's disappointment, but to help resolve the situation and put this issue behind us.&quot;\n\nA number of the alleged kickbacks involved funds that were earmarked for aircraft engine development, according to the criminal information filed in the case.\n\nBoeing is based in Chicago, but the scheme occurred in St. Louis, which was home to its defense and space unit until 2003, when the company moved it to Washington state.\n\nBoeing spokeswoman Bonnie Rodney said Friday that the company has cooperated with authorities and overhauled its expense approval procedures. She said the company is not aware of any illegal expenses related to the company's government contracts.\n\n'Widespread misconduct'\n\nIn a filing Friday in U.S. District Court in St. Louis, prosecutors said that &quot;certain Boeing employees, including several executives, had conspired to engage in the improper and illegal approval of corporate expenditures.&quot;\n\nThe filings said that Boeing officials believed there was &quot;widespread misconduct&quot; that included theft of company funds and illegal personal use of company funds.\n\nIn a &quot;factual resume&quot; filed in court, the government alleged that the executives conspired to authorize company expenses for a variety of items, including theater and concert tickets and interior furnishings for Condit's home, without any business justification.\n\nThe money was funneled through a design company in St. Louis. Prosecutors alleged that about $2.5 million was spent on the house, including $130,000 for a floor-to-ceiling stone fireplace, and $18,000 for a home theater with stadium seating.]" time="1.386"><properties><property name="score" value="0.9889350136121114" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[\n\n\n\n\n\n\n\n\n\n\n\n\n\n\n\nThis is a group for anyone interested in all things data science.\n\n\n\nWe host a variety of talks and events, including:\n\n\n\n\u2022Talks: Short tech talks, panels and panel debates, and longer presentations on a wide range of topics\n\n\n\n\u2022Workshops: Data science and coding workshops on Python, R, data exploration and analysis, visualization, and other topics.\n\n\n\n\u2022Meetups: Events around the world with data science-focused talks, panels, and panel debates\n\n\n\n\u2022Hackathons: Hackathons where attendees work on data science projects using Python or R.\n\n\n\n\u2022Data science team building: Escape the room and board games to improve teamwork, communication, and data science skills\n\n\n\nThis group is a project of the Data Science DC meetup (http://www.meetup.com/Data-Science-DC/).]" time="0.507"><properties><property name="score" value="0.809055507183075" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The mooring cable holding the stricken Rena, which has been stranded on the Astrolabe Reef off the north east coast of New Zealand's North Island since October 5th 2011, is fraying at an &quot;unacceptably rapid rate&quot; and the ship could break up in as little as a week.\n\n\n\nIt could even go as early as Tuesday night, according to New Zealand authorities.\n\n\n\nThey had been expecting the ship to sink, as currents beneath the sea would pull the vessel into the reef, and it would begin to break apart as the tides took it apart. However the &quot;churning&quot; around the ship has actually been slowing the process of the ship disintegrating.\n\n\n\nMeanwhile the crews on board are clearing the decks of the ship, as the removal of the fuel is to begin today, a job which should take a week or two. The removal of the fuel would then free up the mooring cables and allow the ship to break apart and eventually sink.\n\n\n\nHowever if the cable did break, the pieces of the ship could drift away, or worse, float across the Hauraki Gulf and into the Bay of Plenty.\n\n\n\nThe most recent attempt to free the ship from the reef failed, with the bow slipping from the rocky coral. No one has been aboard the ship since a New Zealand Navy ship fired shots at it in January, when the Rena's crew threw their passports over the side.\n\n\n\nThe Chinese-flagged vessel broke free from its tugs in the Bay of Plenty and hit the Astrolabe Reef, spilling 1,700 tons of fuel into the sea. The damage to the reef has been assessed as being very severe.]" time="0.945"><properties><property name="score" value="0.9245613813400269" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[To help raise awareness of depression and suicide, a group of engineering students from the University of Waterloo have designed a robot that can create artwork with the help of artificial intelligence (AI).\n\nMental illness is a huge problem that often gets ignored, or underplayed. Although the majority of the time you can tell if a person is suffering, there are some cases where it can be harder to identify, especially if you're just talking to them on the phone.\n\nThe team of eight, led by Chris Cui, spent four months designing their robot, which is called DepressionBot.\n\nDepressionBot is now available to help raise awareness of depression and suicide 1:03\n\n&quot;The robot is meant to help raise awareness for depression and suicide,&quot; said Cui, a third-year industrial design student, in an interview with CBC's The Morning Edition host Craig Norris.\n\nIt's no secret that mental health issues affect a lot of people, but it's not often that we see people, especially students, take time out of their day to help those affected.\n\n&quot;It's a really personal subject to a lot of people, but there's a lot of stigma around it, so a lot of people don't want to talk about it,&quot; said Cui.\n\nSketchbook artwork from DepressionBot. (Submitted by Chris Cui)\n\nGetting the right artwork\n\nIn order to create the artwork, the team taught the robot to look for certain patterns in its black-and-white drawings, then draw similar patterns in its next sketch.\n\n&quot;We had a bunch of example pictures, some]" time="0.956"><properties><property name="score" value="0.9883027374744415" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[To be blunt, the video you linked to does not really prove anything one way or the other. That video is trying to make it seem like some statements made by Dan Abnett about the use of the term &quot;boy&quot; or &quot;son&quot; when applied to the Raven Guard during the Horus Heresy was a form of apologism for using &quot;boyz&quot; when referring to Orks. This is not the case. While it is true that Abnett's statements were inappropriate, the content of that video fails to really show it. For example, where the video claims that Abnett said &quot;Raven Guard don't ever call themselves boys&quot;, Abnett was in fact saying that Raven Guard did not ever call themselves &quot;boyz&quot; or any other similar slang. He specifically mentioned &quot;boyz&quot; in the context of Orks as an example, and it is the video that is now misrepresenting his words. In other words, Abnett's statements on the subject of Orks are accurate, and the video's attempt to show him in a negative light is based on a misrepresentation.\n\nAdditionally, your claim that the Imperium is &quot;far from perfect&quot; is not true, as the Imperium's culture is based around the worship of the Emperor, who has deified himself and is often referred to as a god by Imperial citizens. This means that the Imperium has absolutely no separation between religion and government, which means that your claim that the Imperium is &quot;far from perfect&quot; is also a form of apologism, as it is ignoring a key element of the Imperium that should be readily apparent to any Imperial citizen. The use of &quot;Apologetic&quot; is a misuse of the word and also misrepresents what the author is trying to say, as you have misrepresented the claim he is making.\n\nThe author's second &quot;claim&quot; is incorrect on many levels, and is also based on a misrepresentation of facts. Firstly, let's take a look at the quote he links to:\n\n&quot;If we wanted to do a straight-up Space Marines game, we would have done Space Marines, right? They\u2019re the coolest guys.&quot;\n\nThis quote is taken out of context. The quote actually goes on to explain why they chose to do the Space Marines as they did. In fact, the quote was in response to a question about why they decided to make Space Marines so &quot;stereotypically heroic&quot; in terms of a tone. The statement that follows in the quote states that it is because Space Marines are the &quot;coolest guys&quot; in the galaxy. What the author is doing here is taking a single quote and presenting it as if it is the entirety of the explanation of why Space Marines are the way they are in this game, when that is not the case.\n\nThe author's next claim is that he does not agree with the statement about the inclusion of certain races, but instead of explaining why he disagrees, he simply says that it is &quot;wrong&quot;. This does not actually prove that the statement is incorrect, and is a form of dog-whistle politics, as it is deliberately vague and meant to be interpreted as &quot;political correctness&quot; or &quot;SJW&quot; or &quot;Cultural Marxism&quot; and similar things, all of which are highly charged words. He could very easily have explained why he disagrees with the statement, but he did not. Instead, he chose to attempt to shut down any further discussion of the topic with a rather obscure quote, while also using a misrepresentation of the quote to try and do so.\n\nThe author's final claim is based on another misrepresentation of the source material. In his article, the author linked to an article that states that the developers actually took fan-feedback into account when designing the Space Marine Chapters. In fact, the developers mentioned that they were &quot;watching&quot; the fans' reactions to the game, and were taking note of how the game was received and what fans' concerns were, which is in line with what the developers have said about their reasons for making the Space Marine Chapters they did.\n\nOnce again, the author is trying to use his article as a form of dog-whistle politics, as he is taking these claims and trying to make it seem as if the developers were lying. However, even if the developers were lying about this, which they are not, that does not invalidate the author's argument. In other words, the author's claims that the developers are lying about this are irrelevant to the argument the author is making, and should not be considered when assessing the validity of his arguments.\n\nIn the author's first sentence, he claims that it is &quot;common knowledge&quot; that the Imperium of Man is based on Nazi Germany. However, I do not believe that this is actually true, as I have never heard this before, and the idea seems rather outlandish. In fact, the idea that the Imperium of Man is based on Nazi Germany is so outlandish that it makes me question the author's sanity. Furthermore, his claim is an appeal to emotion, as the author does not really explain why he thinks this, and instead simply uses an emotionally charged accusation.\n\nIn the author's second sentence, he claims that the Space Marine Chapters, as well as the Emperor, are based on &quot;fascist leadership&quot;. I have already addressed the author's claim about the Space Marine Chapters, and I will now address his claim about the Emperor. While I cannot speak for the Emperor's actual policies, he is still worshiped as a god by most of the citizens of the Imperium, which does not actually fit with the definition of &quot;fascist leadership&quot;. Furthermore, the author does not really explain how he came to the conclusion that the Emperor is based on fascist leadership, and instead simply asserts it without any real explanation.\n\nThe author's third claim is based on a misrepresentation of facts, and is yet another form of dog-whistle politics. His claim that a representation of the logo of the White Legion is on the flag of the Imperium is a misrepresentation of the source material. In the source material, it is explicitly stated that the White Legion and the Imperial flag are two separate things, with the author's claim making it seem as if the White Legion is the official flag of the Imperium of Man. This is, of course, untrue. In fact, the White Legion is not even in the same star system as the Imperium, so the White Legion's flag is actually irrelevant to the argument that the author is making.\n\nThe author's fourth claim is also based on a misrepresentation of facts, as well as yet another form of dog-whistle politics. The author claims that the developers stated that the Imperium is &quot;an alliance between Humans and aliens&quot;. While I cannot speak for what the developers actually stated, I can point out that they did not state that the Imperium was an alliance between Humans and aliens. The author's statement that this claim is &quot;common knowledge&quot; is false. In fact, if this was common knowledge, then the author would not have needed to claim that this was what the developers had said. Instead, he would have just cited the developers as having said it. In other words, his claim that this is &quot;common knowledge&quot; is also a form of dog-whistle politics, as it is a claim that is designed to be interpreted as &quot;political correctness&quot; or &quot;SJW&quot; or &quot;Cultural Marxism&quot; and similar things, all of which are highly charged words.\n\nIn the author's final claim, he says that the Orks are portrayed as a generic enemy of the Imperium, when in fact they are a threat to the Imperium. The author claims that &quot;Boyz&quot; is used as an insult by the Imperium, when in fact it is used to refer to any Ork. In other words, his claim that the Imperium is &quot;far from perfect&quot; is incorrect, and in fact is a form of dog-whistle politics.]" time="4.138"><properties><property name="score" value="0.999787375330925" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The general consensus in psychology is that intelligence is a singular entity that can be measured and is unchangeable, and that any issues with intelligence will remain constant throughout a person\u2019s life.\n\n\n\nIn general, the accepted view of intelligence is based on a series of experiments carried out in the 1940s and 50s, commonly referred to as the IQ tests. These intelligence tests measured intelligence by identifying patterns and memory, and tested mental agility. But the modern view of intelligence is much broader, and focuses on a person\u2019s ability to learn and their capacity to understand their environment.\n\n\n\nResearch carried out by a group of psychologists in New York have challenged the accepted view, and shown that people\u2019s IQ levels do not necessarily remain the same throughout their lives.\n\n\n\nThey found that if a person has an early traumatic experience, then the level of intelligence is likely to reduce significantly.\n\nThe study\n\nThe study used a database of IQ scores for each person, which was then divided into three groups: people who were not exposed to early trauma, people who had experienced trauma before they were 18 and people who had experienced trauma before they were 16.\n\n\n\nThe study then looked at how the group with early trauma had scored after 16 years. What they found was that people who experienced trauma before the age of 16 scored lower on intelligence tests at the age of 30. Those who experienced trauma before they were 18 scored even lower than those who were traumatized before they were 16.\n\n\n\nThey also found that the difference between the group who experienced trauma and those who did not was three to five points, which is a significant amount.\n\n\n\nThese findings highlight the importance of understanding trauma and how it can affect people throughout their lives. It also shows that traumatic events can influence how a person learns and that this will have a lasting impact on their lives.\n\n\n\nThe other important thing to note from this study is that the effect of trauma was found to last for 20 years. Even though trauma had occurred so early in a person\u2019s]" time="1.011"><properties><property name="score" value="0.4799564878145854" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Unfortunately, in both cases the occupants were intoxicated. The first house we went into was this guy's.\n\n\n\n\n\n\n\n\n\nHere's a look at the bar. The beer fridge was in the shed in the back.\n\n\n\n\n\nWhat's a bachelor pad without a recliner?\n\n\n\n\n\nThe back yard had a trampoline. That's all the beer this guy was going to need for a while.\n\n\n\n\n\nNow we move to the house that had the party. There were lots of people at this house, and not a lot of time to take pictures. However, I did get a few that I think will give you an idea of the festivities.\n\n\n\n\n\nThe Docks!\n\nGood spot for the porta potty.\n\n\n\n\n\nWho's ready to rumble?!\n\n\n\n\n\nThe front lawn!\n\nGolf tournament!\n\n\n\n\n\nWhen all was said and done, we had collected 8 trucks, a garbage truck, and a car from both houses.\n\n\n\n\n\nOh, and no injuries.\n\nThe general area of the party was right next to a creek. There were numerous signs saying how you would be arrested if you went in the creek. Well, as it turns out, there was a reason for those signs.Last year, the creek flooded, and one of the houses nearby was torn from its foundation and carried downstream. This was the home's fate.]" time="0.676"><properties><property name="score" value="0.12367662787437439" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Giant scale model\n\n\n\nSide note\n\nBought in New York. My first gigapan, and I don't know if it worked. I used an auto timer on the camera, and moved the camera to a tripod as well, but I think I moved too quickly. I did a slight &quot;edit&quot; on it by changing the colors a bit and used a slightly different size to get a more accurate view of the whole object. I think it worked well. It has lots of things I don't recognize. Probably because I'm from a small town in southern Indiana, USA. I'm actually kind of suprised that it even fits in the frame.\n\nFull Resolution, if you want to look at a close up of the item\n\ngigapan.org/gigapans/72917/\n\nDone]" time="0.459"><properties><property name="score" value="0.9998251795768738" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[A human being is a part of the whole, called by us &quot;Universe,&quot; a part limited in time and space. He experiences himself, his thoughts and feelings as something separated from the rest - a kind of optical delusion of his consciousness. The striving to free oneself from this delusion is the one issue of true religion. Not to nourish the delusion but to try to overcome it is the way to reach the attainable measure of peace of mind.\n\nAlbert Einstein (1938)\n\nWho would have thought that the the famous scientist, whose ideas took a leap of imagination that seemed to defy common sense, had so much wisdom to share.\n\nEinstein's wisdom, delivered in a simple, direct manner, is of a philosophical nature. And yet, as we have seen, his wisdom is based on a scientific perspective, not a philosophical one. It is a perspective that, as we shall see, stands in contrast to the beliefs of the average person.\n\nThe ultimate source of Einstein's wisdom is his sense of awe. He speaks of being in awe of what he observed in nature, as if these observations of nature had moved him to the core of his being. But this awe was not the end of the matter for Einstein. He does not speak of being in awe and then simply stop and say: &quot;What a wondrous world it is.&quot; Rather, he looks to this sense of awe as a source of inspiration to do more.\n\nEinstein's wisdom is the result of a personal approach to understanding and the sense of awe that this approach has produced. It is, therefore, both a personal and a shared source of wisdom. The person sharing this wisdom is not speaking on behalf of some outside authority, like a philosopher or a scientist. His wisdom is simply the result of a personal quest, not a more generalized, impersonal, public investigation.\n\nEinstein did not discover his wisdom by himself. He spoke about his awe at the &quot;dignity&quot; of nature, by which he meant that the world of nature is majestic, not cheap. Einstein was aware of the sense of awe that his discoveries had evoked in him and he wished to share this feeling with others. In this way, Einstein's wisdom was, at its roots, a shared resource, because the more people who can feel and express awe at the majesty of nature, the better it is for all of us.\n\nIt is as though Einstein had an innate wisdom that he wished to share with others, and as though he felt compelled to do so. Einstein's quest for wisdom was not an abstract exercise in which he wished to understand nature or the meaning of life. Rather, Einstein was driven by a need to share his insights. As the physicist Stephen Hawking has pointed out, the very nature of the scientific enterprise is a social activity, and this is as true of Einstein as it is of Hawking himself.\n\nLike Hawking, Einstein had the good fortune of being born into a social context that could support his quest for wisdom. He lived in an environment in which he was exposed to an intellectual culture that had been nourished by many philosophers who were also his contemporaries. He could, therefore, both draw from the wisdom of his predecessors and challenge them by articulating the insights that he had achieved.\n\nIn addition, the new physics that he helped to develop has spawned many technologies that have radically transformed our understanding of the world and has introduced new scientific disciplines, such as quantum physics, that have enriched our understanding of nature. Our scientific knowledge has become a shared resource that can be used as a basis for developing a new kind of wisdom, one that does not exist in the history of human thought before the twentieth century.\n\nAs we can see, the search for wisdom and the sharing of wisdom are based on a wide array of factors. To create a context that will allow for the growth of a new kind of wisdom, one that can benefit the common good of humanity, it is necessary to invest in science and in a new philosophical perspective.\n\nIn other words, a new kind of wisdom can only arise as the result of a new synthesis, one that links scientific knowledge with a philosophical perspective. It is this new synthesis that will be explored in the final chapter of this book.]" time="2.277"><properties><property name="score" value="0.9362875580787658" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Gone are the days of fighting with finicky and slow 3D printers; with the new Prusa i3 MK3S, you get the benefits of simplicity, accessibility, and a whole lot more. This third generation printer is the successor to the very popular i3 MK2. This time around, Prusa has made some notable improvements.\n\n\n\nThe Prusa i3 MK3S prints in high quality with a dual extruder. It\u2019s also compatible with a wide range of filaments including PLA, ABS, PETG, Nylon, and more. This model is an upgraded version of the original, and it has all the perks that will get you printing with no hassle.\n\n\n\nThe printer has a noise-reducing design that won\u2019t bother anyone, and it features a beautiful enclosure that makes it a joy to look at. This sturdy machine is compatible with both Linux and Windows.\n\n\n\nIf you\u2019re looking for a durable, efficient, and reliable 3D printer, this model is just what you need.]" time="0.564"><properties><property name="score" value="0.999776303768158" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Onze melkveehouderij\n\nIn 1990, wij waren de oprichters van Lijst Harkema en verhuisden in 2004 naar Aagtekerke, waar de melkveehouderij in het verleden ook zat. Met de komst van de nieuwe melkveehouderij kwamen de melkveehouders van het dorp ook in de gemeenteraad. In 2011 ging het dorp verder met \xe9\xe9n kandidaat, die meteen werd gekozen.\n\nDe voorbije vier jaar, sinds het vertrek van Gert Mulder, heeft Lijst Harkema vier gemeenteraadsleden gehad: Eenmaal Dorine Schrier, daarna Ad van Haren, Peter Nieuwenhuis en uiteindelijk verkozen wij twee van de twee\xebnhalve tonne melkveehouders: Gerrit van der Poel en Marco Beutels. In onze partij Lijst Harkema mogen alle beslissingen eerst goed worden doorgerekend en beoordeeld door deskundigen voordat er ook maar \xe9\xe9n stap wordt gezet. Bij onze melkveehouderij in Aagtekerke is altijd aandacht voor natuurbeheer en het landelijk gebied.\n\nWe werken vanaf een bepaald aantal dieren ook samen met FrieslandCampina. Dit omdat we geen einde hebben aan het gebied waar we melkvee houden, ook al zijn wij de grootste houderij van heel West-Friesland. Het houden van meer dan 600 melkkoeien vraagt bijna dagelijks toezicht. Onze stalen zijn zorgvuldig afgestemd op de huidige voer- en veevo]" time="1.185"><properties><property name="score" value="0.9358958303928375" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.93589583&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.93589583
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Jonathan Chait at New York Magazine says we should start ignoring Donald Trump.\n\nTrump gets endless media coverage, which he then uses to insist he should be treated fairly by the media.\n\nThe fact that Trump is now president makes his demand for fair coverage a valid concern. So Chait\u2019s right that ignoring Trump isn\u2019t the solution.\n\nBut Chait is also wrong that Trump doesn\u2019t deserve any media coverage. He\u2019s just like any other president.\n\nJust not as good at it.\n\nChait notes that Trump uses the press to get out his message:\n\nThe president of the United States is an extremely powerful figure. Even if the entire country agrees that the president of the United States is, in fact, a buffoon who speaks only in unintelligible gibberish, the president of the United States is still very powerful.\n\nChait is correct. Trump may have been a complete loser in the private sector. But in politics, Trump\u2019s clearly been a winner.\n\nThere are many reasons Trump has been a success in politics. Some of these reasons have to do with his political positions. But most of the reasons have to do with his personality.\n\nTrump is the ultimate salesman. He is a chameleon, able to change his tone and style to appeal to his audience. He\u2019s also the consummate bullshit artist, able to make false promises sound plausible.\n\nAnd Trump\u2019s made one of his signature moves since he began his political career. He completely ignores his Democratic opponents and talks directly to his voters. He\u2019s smart enough to understand that his opponents aren\u2019t really interested in actually talking to him.\n\nTrump has always understood how to get his message out, even if the message isn\u2019t all that important. It\u2019s why he never had a detailed health care plan. He wanted to talk to people who agreed with him about how terrible Obamacare was.\n\nHe\u2019s never really been interested in a detailed plan. He\u2019s interested in getting media coverage, because he knows the media will talk about what he says.\n\nTrump is constantly in search of free media coverage. The media gave him endless coverage during the campaign, and he won. And they\u2019re giving him endless coverage now, and he\u2019s still president.\n\nThe problem is that Trump is terrible at getting his message out in an organized and coherent way. He\u2019s also terrible at actually getting anything done.\n\nTrump\u2019s behavior is extremely damaging. But the damage doesn\u2019t come from the fact that he\u2019s in the news. It comes from the fact that he\u2019s so terrible at what he does.\n\nPeople who work for him never know what\u2019s going to come out of his mouth. And they\u2019re never really sure what he\u2019s going to do.\n\nBut his supporters always know where Trump stands. He\u2019s going to get rid of illegal immigrants. He\u2019s going to bring back jobs from overseas. He\u2019s going to put coal miners back to work.\n\nThat\u2019s the message he has for his supporters. And they know that even if his administration doesn\u2019t do what he says, Trump is still fighting for them.\n\nAll presidents have to use the media to get their message out. The difference with Trump is that his message is essentially nonexistent.\n\nTrump isn\u2019t president because he has an interesting policy agenda. He\u2019s president because he knows how to use the media to his advantage.\n\nThat means that we have to talk about him, whether we want to or not.]" time="1.810"><properties><property name="score" value="0.9995897561311722" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[On the evening of October 23, 2017, the celebrated actor Will Smith hosted an invitation-only event in Las Vegas, hosted by UNICEF to support vulnerable children around the world. During his remarks, Smith reflected on his recent trip to Madagascar, where he met a family who lost their three children to diarrhea. \u201cWhen I say my faith was tested, it was tested. You understand? It was like I was in the Bible.\u201d\n\nSmith is a committed Christian. In 2007, when he was named the most bankable star in Hollywood, Smith made the bold claim that he\u2019s the only top-grossing actor who is a professed Christian. At the time, a study by the Pew Forum on Religion and Public Life found that just 18 percent of Americans consider Hollywood stars to be good role models, and just 10 percent said the same for politicians. \u201cIf the group that you are the leader of is viewed that way, you have to examine what you\u2019re doing.\u201d Smith said.\n\nOn this day, Smith was concerned about children around the world who are not receiving the care they need to survive. As he read about the dire need for vaccines, Smith thought about his own two kids, and how helpless he would feel if they were sick and dying. \u201cI can\u2019t imagine what that must be like. So for me, the easiest thing to do is to support UNICEF. Because when my children get sick, it\u2019s my job to get them help, and if they die, they die. I\u2019m not gonna stop that. I\u2019m not God. But when children around the world get sick and die, it\u2019s God\u2019s responsibility. And I have to believe that God will ask him, \u2018Are you there?\u2019 And when He says, \u2018Yes, I\u2019m there,\u2019 he\u2019ll say, \u2018Are you there?\u2019 And he\u2019ll say, \u2018I\u2019m there,\u2019 and God will say, \u2018Well, why didn\u2019t you help these children?'\u201d\n\nIn one year, UNICEF vaccinates 46 million children against preventable diseases. Every dollar invested in immunization results in $44 saved in treatment. UNICEF reaches children in 190 countries and territories. Each year, with the help of its many partners, UNICEF supplies more than 50 million children under the age of five with immunizations against the six life-threatening diseases: pneumonia, tuberculosis, diarrhea, polio, measles, and tetanus. UNICEF also supports the vaccination of pregnant women in countries where the practice has been shown to prevent transmission of vaccine-preventable diseases to newborns. In addition, UNICEF is the largest distributor of antiretroviral medicines for children and pregnant women with HIV in the developing world.\n\nUNICEF also provides thousands of training opportunities for community health workers to deliver health services to children in their own communities. UNICEF, through its extensive network of child protection partners, also provides psychological support and counselling for children and adolescents affected by violence, abuse and exploitation. For millions of children who would otherwise die, it is UNICEF that saves lives.\n\nIn response to the global refugee crisis, UNICEF is at the forefront of emergency response efforts to help children and families who are fleeing conflict and violence. The agency has distributed more than 1 million emergency packages in Syria, Iraq, Lebanon and Jordan, and has provided over 1.5 million people with emergency supplies including tents, kitchen sets, heaters and sleeping mats. UNICEF has established 41 child-friendly spaces for traumatized children in the region and continues to provide medical and psychosocial support, protection and assistance for refugees.\n\nIn 2016, UNICEF\u2019s commitment to the protection of children in conflict was formally recognized with the Nobel Peace Prize. The Nobel Prize was awarded to UNICEF \u201cfor its work for children in developing countries, and for long-standing advocacy in favor of children\u2019s rights.\u201d\n\nThe U.S. and UNICEF first partnered in 1953, to help children devastated by the Korean War. From that time, their relationship has strengthened. In fiscal year 2015, the U.S. government provided UNICEF with more than $77 million to support its humanitarian efforts. UNICEF also works with the U.S. government on a number of global health initiatives. These include ensuring that every child is fully immunized and preventing HIV transmission from mother to child. UNICEF also works with the U.S. government on the President\u2019s Emergency Plan for AIDS Relief (PEPFAR).\n\nThrough PEPFAR, the U.S. government has invested more than $80 billion to combat global HIV/AIDS since 2003. PEPFAR is a cornerstone of U.S. global health efforts, and it is the largest commitment by any nation to combat a single disease. The program provides 11 million people in over 50 countries with lifesaving antiretroviral treatment, and has contributed to a 40 percent decline in AIDS-related deaths since 2004. In fiscal year 2017, the U.S. contributed $4.2 billion to PEPFAR.\n\nChildren born with HIV and AIDS in Africa, including those living in remote rural areas, can now receive treatment. PEPFAR is working to ensure that every child is tested, identified, and started on treatment. New initiatives are also helping to treat HIV-positive mothers and prevent them from passing the virus on to their unborn children.\n\nIn the five years since the Global Health Security Agenda (GHSA) was launched, more than 40 countries have invested in the development of a National Emergency and Medical Preparedness and Response Plan. To date, more than 15 countries have published plans. In 2016, the GHSA received a boost when the U.S. committed to invest $1.3 billion through 2020. This funding will support the purchase of 50 million doses of vaccines to help save lives from infectious diseases like]" time="3.109"><properties><property name="score" value="0.671850194533666" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[1. I cannot stop smiling, I am so happy.\n\n2. I saw Ellie today and it made me feel really good.\n\n3. My boyfriend\u2019s been away for a couple of days and it made me feel really bad.\n\n4. I have so much to do and so little time to do it all.\n\n5. But I am really happy with my life at the moment.\n\n6. I have been getting drunk and naked with the birds and it makes me feel like a teenager again.\n\n7. I bought a slice of delicious cherry pie today and it made me feel really guilty.\n\n8. But the guy at the deli was really cute so I might go back to buy more.\n\n9. I went out with some friends last night and it made me feel really old.\n\n10. We\u2019re all growing up too fast.\n\n11. We\u2019re all so busy.\n\n12. And our lives are changing.\n\n13. And I am really enjoying my life at the moment.\n\n14. And that makes me happy.\n\n15. And I\u2019m not gonna lie, it makes me feel really good to write all of this stuff down.]" time="0.650"><properties><property name="score" value="0.9997585415840149" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[2017.01.15. Szerz\u0151: Kir\xe1ly Zolt\xe1n\n\nAz elm\xfalt id\u0151szakban t\xf6bbsz\xf6r is m\xe1r t\xe9m\xe1t tett\xfcnk a honlapra, amelyekben a kiv\xe1ndorl\xe1s \xe9s a bev\xe1ndorl\xe1s probl\xe9m\xe1j\xe1r\xf3l, valamint a bev\xe1ndorl\xe1s befektet\xe9si lehet\u0151s\xe9geir\u0151l besz\xe9lt\xfcnk. Az itt szerepl\u0151 \xedr\xe1sokat mindenkit arra k\xe9rj\xfck, hogy ismertesse aj\xe1nl\xe1sainkat, s\u0151t: k\xf6zrem\u0171k\xf6dj\xf6n vel\xfck, ha \xe9s amennyiben kedve tartja. Hogy v\xe9g\xfcl ez az egy kicsit kev\xe9sb\xe9 szeml\xe9letes le\xedr\xe1s is az id\u0151szakon \xe1ttekinthet\u0151 legyen, t\xe9m\xe1nk megoszt\xe1s\xe1val, szeretn\xe9nk sz\xednesebb\xe9 tenni azt.\n\nBev\xe1ndorl\xe1s k\xfclf\xf6ld\xf6n: jelentkez\xe9s az online k\xe9pz\xe9sekre!\n\nEz a k\xf6zel-keleti sz\xe1rmaz\xe1s\xfa konzul bejelent\xe9se, hogy magyar diplom\xe1val is lehet amerikai \xe1llampolg\xe1r. Persze ez csak amolyan k\xf6lts\xe9ges tan\xfas\xedtv\xe1ny, amivel t\xf6bbek k\xf6z\xf6tt az amerikaiakn\xe1l \xe9s a sk\xf3t f\xfcggetlens\xe9gi referendumn\xe9l is seg\xedtette ezt az orsz\xe1got. Nem akarunk err\u0151l semmit el\xe1rulni, mert igen sokan csak ezzel a k\xfcl\xf6nleges diplom\xe1val mentek t\xf6megesen az USA-ba. Szerint\xfcnk ezt a t\xe9m\xe1t egy kicsit megk\xf6nnyebb\xfclj\xfck, ha leiratjuk a \u201eRe: \xc9rdekl\u0151d\xe9s angol nyelv\u0171 m\u0171szaki k\xe9pz\xe9s\xe9re\u201d \xe9s \u201eRe: F\xf3kuszt\xe1bor aul\xe1s k\xe9pz\xe9s, IAS vagy EC\u201d feliratokat is. Ugyanis a bev\xe1ndorl\xe1s nyitott k\xe9rd\xe9s, nyilv\xe1n nem tudni, hogy megmaradhat-e a hazai di\xe1kvizsga, vagy azt is el kell-e engedni, amihez a hatalom t\xf6bb k\xf6zt k\xe9pes \xe9s akar is. \xdagy t\u0171nik teh\xe1t, hogy az elm\xfalt id\u0151szakban nem ez a t\xe9ma kapta el a f\u0151sodor f\xf6l\xe9ny\xe9t, hanem ink\xe1bb az, hogy a diplom\xe1kat haszn\xe1lva bev\xe1ndorl\xe1sra k\xe9sz\xfclnek, mert szeretn\xe9k k\xfclf\xf6ld\xf6n folytatni tanulm\xe1nyaikat.\n\nDe mi\xe9rt is t\xf6rt\xe9nne ennyire sz\xe9lesk\xf6r\u0171 bev\xe1ndorl\xe1s, ami a diplom\xe1t kap\xf3k sz\xe1m\xe1t is mag\xe1ba foglalja? Mert a diplom\xe1val rendelkez\u0151k ugyan egyre kevesebben vannak, m\xe9gis akad b\u0151ven olyan, aki tud \xe9s ak\xe1r \xf6r\xf6m is a tanul\xe1s, a munka, de legal\xe1bbis a k\xfclf\xf6ldi \xfat. Ugyanis nemcsak a magyar fels\u0151oktat\xe1s gondot szenvedett, hogy az elm\xfalt h\xfasz \xe9vben az egyre k\xe9pzettebb munkaer\u0151 l\xe9tsz\xe1m egyre cs\xf6kkent, hanem m\xe9g a diplom\xe1val is rendelkez\u0151k sz\xe1ma is egyre kisebb. Az emberek sz\xe1m\xe1ra a k\xfclf\xf6ldi p\xe1ly\xe1zatokat is n\xe9zve, igazolni tudom, hogy nincs egy\xe1ltal\xe1n olyan k\xe9pz\xe9s, amelynek ne lenne r\xe9szben- vagy teljesen online k\xe9pz\xe9se is. Ugyanis a mai informatika, oktat\xe1s \xe9s kult\xfara nagyon nagy szerepet kap a dolgokban, ez\xe9rt \xe9rdemes elgondolkodni azon, hogy nem kell k\xe9nytelen vagy b\xe1rmi okb\xf3l kiszolg\xe1ltatott \xe1llapotban v\xe9gezni az oktat\xe1st.\n\nAz online k\xe9pz\xe9sek jelentik az \xfajat, amit ma a diplom\xe1val rendelkez\u0151k a legnagyobb ar\xe1nyban vesznek ig\xe9nybe, \xe9s szerintem nem is egy rossz d\xf6nt\xe9s, ha valaki le is teszi az online k\xe9pz\xe9st. Ami m\xe9g \xe9rdekesebb az eg\xe9szben, az az, hogy nem csak a bev\xe1ndorl\xe1sra k\xe9sz\xfclnek az online k\xe9pz\xe9sek seg\xedts\xe9g\xe9vel. Val\xf3j\xe1ban m\xe9g a saj\xe1t orsz\xe1gunkban is egyre t\xf6bb jelentkez\u0151 van olyan online k\xe9pz\xe9sekre, amelyekkel m\xe1s orsz\xe1gokban t\xf6megesen jelentkeznek. Ez teh\xe1t el\u0151ny \xe9s h\xe1tr\xe1ny is egyben, de szerint\xfcnk a legt\xf6bben ink\xe1bb a jelentkez\xe9s k\xf6vetkezt\xe9ben a szerencs\xe9j\xfckre vannak. De hogy mi a jelentkez\xe9s eredm\xe9nye? Kiv\xe1ndorl\xe1s vagy bev\xe1ndorl\xe1s, de a legt\xf6bb helyen egyel\u0151re m\xe9g azt sem lehet tudni, hogy van-e lehet\u0151s\xe9g a bev\xe1ndorl\xe1sra, vagy sem. De a fenti c\xedmen \xe9rdemes bejelentkezni, \xe9s ut\xe1naj\xe1rni, hogy van-e itt online k\xe9pz\xe9s. K\xe9rj\xfck, hogy t\xf6bbsz\xf6r is ut\xe1naj\xe1rjon, mert a keres\xe9s egy m\xe1sik c\xedmre is el\xe9rhet\u0151, \xe9s egy j\xf3 okos program lehet\u0151v\xe9 teszi azt, hogy az eredm\xe9nyek szeml\xe9ltet\xe9se legyen a lehet\u0151 legjobb.\n\nA magyar diplom\xe1val is lehet amerikai \xe1llampolg\xe1r?\n\nA t\xe9m\xe1t komolyan vegy]" time="0.029"><failure message="RuntimeError: The expanded size of the tensor (553) must match the existing size (514) at non-singleton dimension 1.  Target sizes: [1, 553].  Tensor sizes: [1, 514]">E   RuntimeError: The expanded size of the tensor (553) must match the existing size (514) at non-singleton dimension 1.  Target sizes: [1, 553].  Tensor sizes: [1, 514]</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Twitter is the perfect platform to get the word out there about anything you are trying to promote. I am a huge fan of using Twitter as a marketing tool, even more than Facebook. It\u2019s a great way to network and reach out to people who are interested in what you have to offer. Here is my list of 10 tips to help you promote anything you are selling on Twitter:\n\n1. Use Hashtags\n\nI love using hashtags on Twitter. They help to promote the things you are interested in and create a sense of community within a group. I will admit that it can be hard to remember hashtags to use for all the things you promote. One way I suggest to overcome this is to put your hashtags in a Twitter search to make sure they are not already being used. This will save you some time and help you avoid any issues down the road.\n\n2. Retweet\n\nIf you like a tweet, retweet it! I\u2019m sure most people are familiar with the concept of retweeting, but I just wanted to put it out there to make sure it was stated. Retweeting is a great way to share the things you like, and can be a great way to show your audience that you are involved in the social media community and help them feel more comfortable doing business with you.\n\n3. Mention\n\nMentioning someone on Twitter is a great way to show your audience that you are engaged and want to interact with them. It is important to mention other people within your field of interest in order to show your audience that you are well connected. I always make sure to mention people who have liked or retweeted my stuff as well. It makes me feel good when I get a mention and it shows that the people I engage with are interested in what I have to say.\n\n4. Post About Your Daily Life\n\nThere is nothing wrong with posting about your daily life and about other stuff you like. I\u2019m sure you all know about the Twitter feeds that are all business. I don\u2019t like following people like that because I feel like it takes away from the idea of the social aspect of Twitter. By being a person and talking about your life, you can show your audience that you are a real person and are interested in other people\u2019s lives as well. I\u2019m sure you have seen some tweets about people\u2019s lives that you thought were really cool, and the people that posted those tweets have engaged you because of it.\n\n5. Be Social\n\nDon\u2019t forget to be social. I\u2019m sure that when you first started out you didn\u2019t know what to do on Twitter. My suggestion to you is to find some people who are interested in the things you are and start a conversation. You will see that the conversations that start on Twitter are real and can be interesting. One of my favorite things to do on Twitter is to start conversations with people. It gives me the chance to learn from other people and also helps them learn about me.\n\n6. Use Pictures\n\nI love using pictures on Twitter. I know a lot of people like to include pictures when they post about things, and I would like to suggest that you do so as well. Including a picture with your tweet can get you more retweets and likes, and can really help to boost the effectiveness of your post. I like to use pictures from Instagram with my tweets, but you can really use any picture. If you want to make a picture of yourself that has an impact on your audience, I would suggest using a tool like Socialbakers.\n\n7. Use Hashtags That Make Sense\n\nI\u2019ve touched on this before, but I want to emphasize it again. It is important to make sure that you are using hashtags that make sense. You don\u2019t want to hashtag your tweets with irrelevant hashtags just to get a retweet or a like. You also don\u2019t want to make hashtags that are too long. I would suggest keeping your hashtags short and relevant.\n\n8. Direct Message\n\nThere are times when a DM is more appropriate than a tweet. You will be surprised at how many people DM you. I think a lot of people just use it to ask questions, but you can also use it to network. I think a lot of people don\u2019t use DMs because they don\u2019t think their audience will see it. If you are going to send a DM, make sure it is personal. Send it to someone who is following you. It\u2019s a lot less likely that they will check the people they follow to see who is sending them a DM, so if you DM a person that is following you, they are much more likely to respond.\n\n9. Use Images\n\nI want to encourage you to use images with your tweets. Images are the first thing people look at, so you want to make sure that you are using good images. One thing I like to do is to include images with my tweets. I\u2019m sure you have seen a lot of Twitter profiles that do this. It helps to keep your tweets interesting and make your tweets more unique.\n\n10. Be You\n\nThis last one is really important. You want to be yourself on Twitter. The more genuine you are, the better. People can see through a fake personality. You want to engage with your audience and show them who you are. When they see you as a real person, it will help to build a more trusting relationship between you and your audience.\n\nThere you have it! I hope these tips help you to improve your Twitter account and that your business gets better because of it. I always love getting comments from you all, so if you have any more suggestions or questions, feel free to drop me a comment below.]" time="3.537"><properties><property name="score" value="0.9981955289840698" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Auszeichnungen\n\nBuchnerstatterin des Jahres, \xd6sterreichischer Buchpreis 2017, \xd6sterreichischer Phantastikpreis 2017, Burgenl\xe4ndische Literaturpreise 2017, Internationale Literaturpreise der Wiener Buchmesse 2018, \xf6sterreichische Kurzgeschichtensammlung des Jahres 2017, Deutscher Phantastik Preis 2018\n\nAwards\n\nAustrian Book Prize 2017, Austrian Fantasy Award 2017, Burgenland Literary Prize 2017, International Literary Awards of the Vienna Book Fair 2018, Austrian Short Story Collection of the Year 2017, German Fantasy Award 2018\n\nThema des Monats]" time="0.395"><properties><property name="score" value="0.9878178238868713" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.98781782&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.98781782
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Watermelon Fried Rice\n\nPrint This\n\nIngredients:\n\n1 can chicken broth\n\n1 lb boneless chicken, cut in strips\n\n3 cups cooked rice\n\n2 tablespoons butter\n\n2 tablespoons olive oil\n\n3/4 cup fresh cilantro, chopped\n\n2 cups watermelon, seeded and chopped\n\n2 green onions, sliced\n\n3/4 cup red bell pepper, seeded and chopped\n\n1/2 cup red onion, diced\n\n1/4 cup cashews\n\n1/4 cup raisins\n\nDirections:\n\nCombine chicken broth and chicken in a bowl. Let soak in refrigerator for at least 30 minutes.\n\nWhen ready to cook, remove chicken from broth, but reserve broth for the sauce.\n\nIn a skillet, melt butter with olive oil over medium-high heat. Add chicken and saut\xe9 until chicken is done and lightly browned.\n\nReduce heat to low. Add rice, cilantro, watermelon, onions, bell pepper, cashews and raisins. Stir until well mixed.\n\nAdd reserved chicken broth, a little at a time, until it reaches desired consistency.\n\nServes 4.\n\nRecipe compliments of American Farm Bureau]" time="0.527"><properties><property name="score" value="0.9998138546943665" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Check out the info below for future tour dates, and track them to receive alerts when they tour near you\n\n&quot;The one and only&quot; as his legions of fans call him, John Lydon is the only original member of the world's most celebrated punk band, PiL.\n\n\n\n'All the very best, Johnny Rotten!' The Independent\n\n\n\n'It was awesome!' The Guardian\n\n\n\n'He's got some music to make, let him do his job' Mojo\n\n\n\n'The People's Punk. Still Rotten after all these years' Sunday Times\n\nJohn Lydon John Joseph Lydon (born 31 January 1956), also known by his former stage name Johnny Rotten, is an English singer, songwriter, and musician. He is best known as the lead singer of the punk rock band the Sex Pistols, which lasted from 1975 to 1978, and again for various revivals during the 1990s and 2000s. He is also the lead singer of post-punk band Public Image Ltd, which he founded and fronted from 1978 until 1993, and again since 2009.\n\n\n\nLydon's musical career has spanned over four decades. As a musician, he has also been a member of several bands: PiL (1978\u20131992, 2009\u2013present), the Flowers of Romance (1978), the Joneses (1979), and the Nightingales (1979\u20131980). Since 2013, he has put his PiL bandmates on hiatus to focus on his new band, the Public Image Ltd band. He has also hosted a weekly radio show on BBC Radio 6 Music, Rotten Radio.\n\nJohn Lydon tour dates]" time="0.887"><properties><property name="score" value="0.13481095433235168" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.13481095&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.13481095
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Residence No. 8,\n\nA House with a strong connection to our Imperial Russian history, is situated in a quiet suburb of St Petersburg and has every luxury one can wish for. It\u2019s a real delight to visit the beautiful private beach of your own. The property is furnished in the best Russian style of the late 19th century.\n]" time="0.221"><properties><property name="score" value="0.9995414018630981" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[\n\nARTIST: Hot Tuna\n\nALBUM: First Pull Up, Then Pull Down\n\nYEAR: 1970\n\nCD REISSUE: no\n\nSHUFFLE: no\n\n\n\nTRACKLIST\n\n01. Somewhere In Between\n\n02. Highway Blues\n\n03. Alabama Blues\n\n04. All Right, Ok, You Win\n\n05. Low Down Blues\n\n06. Man On The Street\n\n07. Up To My Neck In Blues\n\n08. Teardrops Will Fall\n\n09. Mandy's Song\n\n10. Something To Look Forward To\n\n11. Just Between You And Me\n\n12. Muddy's Blues\n\n\n\n]" time="0.354"><properties><property name="score" value="0.6680221557617188" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.66802216&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.66802216
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[We provide comprehensive, customized solutions and a wide array of products. Some of the products we offer include:\n\nActual LED\n\nApplications\n\nAviation\n\nBulk Acrylic\n\nBusiness Cards\n\nCable\n\nClip-on\n\nCord\n\nCustom\n\nDisplay\n\nExit\n\nFlush\n\nFlush Mount\n\nGlass Mount\n\nJoint\n\nLight\n\nLit\n\nLouver\n\nMail Box\n\nMedical\n\nMedical Grade\n\nMezzanine\n\nOccupancy\n\nPanels\n\nRail\n\nSign\n\nSlim\n\nSpiral\n\nStripe\n\nTaper\n\nTranslucent\n\nWindow\n\nVisit our online store at shop.ledwholesale.com to learn more about our products and to place an order.]" time="0.344"><properties><property name="score" value="0.9968246221542358" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The Ultimate One Night Stand in the Square\n\nChapter 1: Proposal\n\nAs a bridesmaid, you're required to do a number of things that I'd never consider doing voluntarily. First of all, you're expected to wear a hideous dress that you've never worn before. Second, you're required to spend your evening catering to everyone else, acting as the host of the event. Most importantly, however, you're required to smile a lot and appear happy.\n\nAt the wedding I attended, everyone was happy, especially the bride and groom. I was happy for them, but I wasn't the happiest bridesmaid in the crowd. In fact, I was the exact opposite.\n\nWhy was I so unhappy? I wasn't the one getting married; I was the maid of honor. I was so annoyed that I was even standing up for my ex. We'd broken up last month because I thought he was cheating on me with a blonde bimbo named Summer. I was at the wedding because my best friend Alicia insisted that I go, but if she knew how much I hated the groom, she would have probably asked me not to go at all.\n\nAnyway, this was the part of the wedding that was pissing me off: the proposal. I was trying my hardest to hold back my tears, but I was failing. I didn't love him anymore, and I never would again.\n\nI glanced over at my ex-boyfriend, Blake. He was really handsome, but that's not what was important right now. He was going down on one knee and was about to propose to the most beautiful girl I'd ever seen.\n\n&quot;Andi, will you marry me?&quot;\n\nI cringed. He'd asked me the same question about ten times, but I was too busy being angry at him to remember. This time was different, however, because he was now asking the blonde bimbo he was sleeping with.\n\n&quot;Yes,&quot; she said, finally accepting the proposal.\n\nAs soon as he slid the ring on her finger, everyone in the room erupted into cheers. I just stood there, my mouth wide open, in complete shock.\n\n&quot;Yes, that's my girl,&quot; said Blake, &quot;And now she's my fianc\xe9e!&quot;\n\nAs they walked out of the church and into the limo, Blake leaned over and kissed her cheek. A couple seconds later, I lost it.\n\n&quot;Bastard,&quot; I muttered. &quot;Bastard.&quot;\n\n&quot;Andrea?&quot; said Alicia.\n\n&quot;He just asked that whore to marry him,&quot; I said. &quot;It was my turn. I was the one who should have been there. I'm the one who was supposed to marry him!&quot;\n\n&quot;Yeah, well I know that,&quot; she said, rolling her eyes. &quot;But it's over now. You just have to be a big girl and accept it.&quot;\n\n&quot;He doesn't love her,&quot; I said, &quot;I know it. He was with her while he was with me, remember?&quot;\n\n&quot;Andi, I hate to say this, but it's too late now,&quot; she said. &quot;He's getting married. You can't do anything about it.&quot;\n\n&quot;I could show up at the wedding,&quot; I said, &quot;and say that I'm his girlfriend and I'm pregnant.&quot;\n\n&quot;Andi, that's insane,&quot; she said. &quot;You know it's not your baby. You're going to destroy his wedding day if you pull something like that.&quot;\n\n&quot;That's the point,&quot; I said. &quot;I want him to feel the way I feel. I want him to feel miserable for a change.&quot;\n\n&quot;Look,&quot; she said, &quot;I'll go with you, but you can't tell Blake anything. If you do, you're going to be the biggest loser in the room, and you're going to lose your best friend.&quot;\n\nI hesitated. I knew she was right. If I ruined the wedding, I'd be in a lot of trouble.\n\n&quot;Fine,&quot; I said. &quot;I won't tell him anything.&quot;\n\nA few days later, I arrived at the church. The doors were open, so I wandered in. The church was completely empty, which made it look a lot bigger. I could hear some talking coming from the altar, so I walked in that direction. The first person I saw was Blake, wearing a brand new suit and looking like the happiest man on the planet.\n\n&quot;You look so beautiful,&quot; he said to the blonde bimbo who stood next to him.\n\nI thought I would be furious when I saw her, but I wasn't. I was heartbroken and sad, but I didn't hate her. In fact, I didn't feel anything towards her. I was just jealous, and the sight of her made me remember how much I'd wanted to marry Blake myself.\n\nI walked up the aisle, ignoring the fact that I was completely alone. I sat down in the front row, and began to think about what had happened.\n\n&quot;How did it happen?&quot; I said to myself. &quot;Why did we break up?&quot;\n\nI knew that I loved him, but I was certain he didn't love me. My best friend Alicia and I had discussed it, and she agreed with me. She was convinced that Blake was cheating on me with a blonde bimbo named Summer, but I wasn't convinced.\n\n&quot;It can't be her,&quot; I said. &quot;How did he meet her? Where did they meet?&quot;\n\nJust then, a woman walked up to the front of the church. She was dressed in a black robe, and she had short, black hair.\n\n&quot;Are you Andrea?&quot; she asked.\n\n&quot;Yes,&quot; I said.\n\n&quot;I'm sorry,&quot; she said, &quot;but the wedding has been cancelled.&quot;\n\n&quot;What do you mean?&quot; I asked.\n\n&quot;Blake is getting married in Las Vegas,&quot; she said.\n\n&quot;But I'm his fianc\xe9e,&quot; I said, &quot;And I'm supposed to be here.&quot;\n\n&quot;You were never his fianc\xe9e,&quot; she said. &quot;The bimbo, Summer, is the one he's getting married to.&quot;\n\n&quot;I'm sorry,&quot; I said. &quot;I don't understand.&quot;\n\n&quot;Neither do I,&quot; she said, &quot;but I was supposed to tell you the wedding was cancelled. So, that's what I'm doing. Sorry for ruining your night.&quot;\n\n&quot;It's okay,&quot; I said. &quot;It's not your fault.&quot;\n\nI knew I had to get out of there. I could barely stand being there as it was, so I hurried out the door. I walked down the sidewalk, not knowing where I was going. I had no idea what to do next.\n\n&quot;Hey,&quot; said a voice behind me.\n\nI turned around and saw Blake running towards me.\n\n&quot;You were supposed to be at the wedding,&quot; he said. &quot;Where have you been?&quot;\n\n&quot;I'm not going to lie to you,&quot; I said. &quot;I'm really hurt, and I don't want to be in your wedding. I'm just here because your father paid for the wedding.&quot;\n\n&quot;You don't understand,&quot; he said. &quot;We're getting married in Las Vegas tonight.&quot;\n\n&quot;Really?&quot; I said.\n\n&quot;Andi, we've got to go,&quot; he said, &quot;it's going to start any second.&quot;\n\nI took one look at his face, and I knew that he was telling the truth. I felt the tears well up in my eyes.\n\n&quot;So, this is it?&quot; I said. &quot;You're really getting married?&quot;\n\n&quot;Andi, I'm sorry,&quot; he said. &quot;I didn't know how to tell you. My father forced me into this, and I couldn't say no. If I had it my way, you'd be my wife right now. But this is the way it has to be.&quot;\n\nI felt a tear escape my eye, and I quickly wiped it away.\n\n&quot;You look beautiful,&quot; he said.\n\n&quot;I'm not going,&quot; I]" time="4.205"><properties><property name="score" value="0.9917177217347282" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[IMG_2720\n\nJune 26, 2018. This evening we worked on wrapping a glider for travel to a new home. Pictured here is Matt Berger of Berger Aviation taking care to wrap the craft. (I love the old mobile home!)]" time="0.167"><properties><property name="score" value="0.9857378005981445" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[4.25 stars - Review by Trisha\n\n\n\nIn my experience, most male-on-male romance books are really hot and steamy but the storyline is lacking, but this one definitely surpassed my expectations. This book wasn't hot and steamy. Not even a little bit. There was one sexy scene between Jay and Davey but that was about it. But it was such a great story. The storyline kept me interested the whole time and I was really rooting for them to have their happily ever after.\n\n\n\nDavey just recently moved from Ireland to get away from his family and the drama that comes with them. He loves working at a video game store and is really happy there. He has a steady boyfriend but that doesn't mean that Davey is 100% out of the closet. Jay is his friend from work but Davey doesn't know that Jay is in love with him.\n\n\n\nJay is totally in love with his best friend Davey. He thinks that Davey is perfect but he can tell that Davey is hiding something. Davey has a hard time trusting people and Jay doesn't know why. But Jay can tell that Davey is hiding something so he just decides to wait it out until Davey is ready to tell him. He's hoping that one day, Davey will realize that Jay is in love with him and he will finally be able to be with the man he loves.\n\n\n\nI really liked the characters. I felt really bad for Davey and how his family was treating him. I also really liked Jay. He was such a great guy. He was just really honest about his feelings for Davey. And I loved that about him.\n\n\n\nIt was a really great story. I loved the characters, I loved the storyline, and I was really rooting for these two to be together.]" time="0.879"><properties><property name="score" value="0.9998183250427246" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[La ministre f\xe9d\xe9rale de la Sant\xe9, Maggie De Block (Open VLD), n'entend pas se mettre \xe0 genoux devant les r\xe9clamations de ceux qui r\xe9clament un d\xe9remboursement du soin m\xe9dical de base. &quot;J'entends les gens, mais je pense qu'il faut se tourner vers d'autres solutions&quot;, explique-t-elle au soir de l'anniversaire de la prise du pouvoir par le nouveau gouvernement de coalition f\xe9d\xe9ral. &quot;Je vais en d\xe9battre avec les partenaires sociaux, dans la soci\xe9t\xe9 civile, avec les \xe9lus r\xe9gionaux et les \xe9lus locaux&quot;, ajoute-t-elle.\n\nLes chiffres sont l\xe0. Selon les calculs de la Drees, le co\xfbt du maintien du syst\xe8me de soin de base en Belgique est de 2,8 milliards d'euros par an. Une somme qui n'est pas forc\xe9ment optimale dans un contexte budg\xe9taire tendu. En 2014, par exemple, des remboursements aussi massifs ont provoqu\xe9 un manque \xe0 gagner de 1,1 milliard d'euros pour le gouvernement f\xe9d\xe9ral. Un d\xe9ficit qui s'ajoute au budget annuel d'environ 21,5 milliards d'euros de la sant\xe9 publique.\n\nLa ministre se montre prudente, parce qu'il ne faut pas se tromper de combat. Il ne s'agit pas de r\xe9former le syst\xe8me de soins de base pour une raison financi\xe8re, mais pour l'am\xe9liorer, pr\xe9cise-t-elle. &quot;Je ne suis pas pour le d\xe9remboursement de la m\xe9decine de base, mais je veux prendre un nouveau d\xe9part, et que tout le monde y trouve son compte&quot;, ajoute-t-elle.\n\nMaggie De Block a d'ailleurs entam\xe9 une s\xe9rie de consultations, avant de rendre son rapport final. A ses yeux, le &quot;temps de l'improvisation est pass\xe9&quot;, car les gens attendent des r\xe9sultats concrets. Le gouvernement lib\xe9ral flamand et le PS, ainsi que des institutions nationales, mais aussi internationales, auront leur mot \xe0 dire dans ce d\xe9bat.]" time="1.599"><properties><property name="score" value="0.9759522378444672" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.97595224&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.97595224
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Hello, I hope I'm posting this in the right area, I'm not sure. But I am in a bit of a predicament. I got my first laptop (Dell Inspiron 15R) for Christmas, and when I put it on my first university assignment it worked fine, for like 3 days. Then all of a sudden it starts turning off by itself. I've tried it on multiple chargers and outlets, still the same. I then found out that the reason why it would do this was because I had the processor set on High Performance, which I've now changed to Balanced, but that's still not fixed the problem. I have had this laptop for 2 weeks, and it's really beginning to make me feel insecure in a laptop.\n\n\n\nDoes anyone know what this could be, and how to fix it? Thanks.]" time="0.456"><properties><property name="score" value="0.9991509914398193" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The beach\n\nNew Year\u2019s Eve party\n\nNew Year\u2019s Day party\n\nThe beach is the number one attraction of Rio Grande do Sul. Each day we celebrate the weather, going to the beach in the morning and afternoon. During the evenings we celebrate the seasons of the year in winter we sit around the fireplace, while in summer we dance around the bonfire. In summer we dance samba in the streets of Santa Catarina, in Rio Grande do Sul, in the middle of the city, in the suburbs, in the mountains, at the beach and on the hills. At night we can enjoy live music in one of the many bars of Santa Catarina. You can find bars and clubs that play regional music, national music, world music and, in many cases, international music. If you are an extroverted person, there is no better place than Santa Catarina.]" time="0.404"><properties><property name="score" value="0.9982410669326782" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Details\n\nPictures\n\nClubs Top Characters Jan Jan (\u30b8\u30e3\u30f3) (Saibara Jansai) Is a survivor of the Third World. His hair is red, but he later bleaches it.\n\nHe uses a slingshot, but eventually gains a bandit's gun after trading with Gorobei.\n\n\n\nVoice Actors Hayashi, Hiroko Japanese\n\n\n\n]" time="0.330"><properties><property name="score" value="0.5572925806045532" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.55729258&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.55729258
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Replies: 17 Royal Viewer 2.1.0.d7902 (with 1.7.0.d7804 Patch) Released Views: 1,625 Posted By Rick Brown Thanks, Satch, for the fast response. ...\n\n\n\nThanks, Satch, for the fast response. I have a question about that icon: Is there an option that can be set to not show it if it's not needed, and only show it when needed? Thanks, Satch, for the fast response. I have a question about that icon: Is there an option that can be set to not show it if it's not needed, and only show it when needed?]" time="0.355"><properties><property name="score" value="0.991931140422821" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Panda's a great restaurant that does a fantastic menu and even better service. You can't go wrong with anything on the menu. It's all very tasty. The veal is very tasty.\n\nWas this review helpful to you?]" time="0.175"><properties><property name="score" value="0.9996205568313599" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Top Rank Inc. has signed welterweight contender Rod Salka to an exclusive promotional agreement, Top Rank president Todd duBoef told ESPN.com on Tuesday.\n\nSalka, 29, will fight on the undercard of the Timothy Bradley-Juan Manuel Marquez pay-per-view main event on Oct. 12 at the Thomas &amp; Mack Center in Las Vegas.\n\nSalka, a former featherweight and super featherweight, will face Jorge Pimentel (23-4, 18 KOs), 32, of the Dominican Republic, in a scheduled 10-round welterweight fight. The card will be televised by HBO.\n\n&quot;I've signed with Top Rank and will be fighting in Las Vegas on Oct. 12,&quot; Salka said. &quot;I'm very excited about this opportunity. This is a great chance for me to go against one of the biggest names in boxing. I'm so thankful for the support that I've received from my family, friends and fans.&quot;\n\nSalka, who lives in Pittsburgh, is 23-3 with 13 knockouts and has not fought since July 31, 2011, when he lost a 12-round unanimous decision to Danny Garcia on the undercard of the Amir Khan-Zab Judah card at Mandalay Bay in Las Vegas.]" time="0.732"><properties><property name="score" value="0.8979974985122681" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[From Wikipedia, the free encyclopedia\n\nHarry P. Robinson\n\nHarry Payne Robinson (May 11, 1872 \u2013 May 10, 1955) was an American Democratic politician from Philadelphia.\n\nHe was born on May 11, 1872 in Pennsylvania.\n\nRobinson was a member of the Philadelphia City Council for three terms. He was elected to the 71st United States Congress and the four succeeding Congresses. He did not seek renomination in 1920. He was elected again to the House of Representatives in 1930, this time]" time="0.249"><properties><property name="score" value="0.9983464479446411" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The Adrian School District is dedicated to providing the best possible learning environment for all students. This school district is composed of approximately 5,000 students in grades K-12 and 125 full-time and part-time teachers.\n\nAdrian is a town of approximately 16,000 people, located in southwest Michigan. Adrian is best known for its variety of community events, from festivals and parades to musical concerts. A new football stadium opened in 2015, and other athletic facilities are planned for the future.\n\nSince 1873, the Adrian School District has been dedicated to providing students with the best education possible.\n\nAdrian Schools is one of the three districts within the Jackson-Adrian Area Schools, which includes Jackson and Adrian. This collaboration allows our students to take advantage of an excellent education and opportunities for athletic competition in a small, personal community.\n\nWe are very proud of our students and staff and are very excited about the future of our district.\n\n- Chris Hettinger\n\nSuperintendent]" time="0.499"><properties><property name="score" value="0.8443364202976227" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[fot. J\u0119drzej Skiba\n\nWizualna uczta dla wielbicieli Harry\u2019ego Pottera. Ju\u017c niebawem w Olsztynie!\n\nZdj\u0119cia z filmu zrealizowane przez Alfonso Cuar\xf3na, do kt\xf3rego graj\u0105 gwiazdy jak Daniel Radcliffe czy Emma Watson, zapowiadaj\u0105 si\u0119 niezwykle okaza\u0142e. Zaanga\u017cowanie mieszka\u0144c\xf3w, \u015bwiat\u0142o, charakterystyczne elementy miasta b\u0119d\u0105 w tym filmie cz\u0119\u015bci\u0105 wizualnej uczty dla ka\u017cdego fanu Harry\u2019ego Pottera.\n\nPlanowana dat\u0105 startu projektu jest pa\u017adziernik 2019 roku. Projekt b\u0119dzie realizowany przez Filmowy Ob\xf3z Gier, Wsp\xf3\u0142pracuj\u0105cych ze Stacj\u0105 Filmoznawcz\u0105 im. Andrzeja Wajdy i Stowarzyszenie Pomocy Inicjatywom Kulturalnym Miasta Olsztyna.\n\n- Miejscem realizacji filmu b\u0119dzie miasto Olsztyn, zar\xf3wno star\xf3wka, jak i inne cz\u0119\u015bci miasta, a prawdopodobnie tak\u017ce same Podlasie. Trudno jednoznacznie wskaza\u0107 przeznaczone obszary, gdy\u017c pozyskamy najlepsze miejsca do filmowania. W pobli\u017cu rzek, na szlakach turystycznych, w zabytkowych okoliczno\u015bciach, w obiektach, kt\xf3re by\u0142yby dost\u0119pne dla produkcji, przy budynkach kulturalnych, w wojskowych obiektach - m\xf3wi Mariusz \u0141azicki, prezes Filmowego Obozu Gier.\n\nTymczasem pocz\u0105tek pracy tw\xf3rc\xf3w zaplanowany jest na 21 wrze\u015bnia w Centrum Kultury w Olsztynie. Wej\u015bcie na teren projektu jest bezp\u0142atne, a pierwsze pr\xf3by rozpocznie si\u0119 w pa\u017adzierniku, r\xf3wnolegle z realizacj\u0105 obiektu.\n\nPodczas wyprawy filmowc\xf3w po Europie, prac\u0119 tw\xf3rc\xf3w poprowadzi ameryka\u0144ski re\u017cyser.\n\nW\u015br\xf3d uczestnik\xf3w b\u0119dzie mia\u0142a r\xf3wnie\u017c swoje miejsce Olszty\u0144ska Fabryka Form Filmowych - eksperymentalna scenografia - najwi\u0119kszy tego typu obiekt w Europie.\n\n- Ten nowy obszar pracy filmowej przejdzie gruntown\u0105 renowacj\u0119, w tym nowa przepustowo\u015b\u0107 dla t\u0142umu widz\xf3w. Olsztyn i teren przy Fabryce Form Filmowych wyr\xf3\u017cniaj\u0105 si\u0119 na tle innych miast w Polsce, ale i w Europie. Dzisiaj ju\u017c niebawem mo\u017cemy wystawi\u0107 szyld w kraju i na \u015bwiecie - m\xf3wi \u0141azicki.\n\nOd 2008 roku, czyli od czasu pojawienia si\u0119 pierwszych prac re\u017cyserskich w tym miejscu, powstaje kilkana\u015bcie prac, dzi\u0119ki czemu ma miejsce w Europie, wykorzystywane na potrzeby filmowe i komercyjne.\n\nObecnie w przysz\u0142ym roku zaplanowana jest realizacja kolejnego filmu o losach czarodziejki, gdzie Olsztyn b\u0119dzie g\u0142\xf3wnym miejscem akcji.\n\n- Jedna z produkcji ma za zadanie pokaza\u0107 mieszka\u0144com miasta, a za po\u015brednictwem filmu z ca\u0142ego \u015bwiata, jak Olsztyn zosta\u0142 wybrany do filmowania. W przysz\u0142o\u015bci obiekt wykorzystany jest do realizacji dalszych produkcji z unikalnym potencja\u0142em. Takie miejsce ma szans\u0119 powsta\u0107 tylko raz na kontynencie - podsumowuje prezes.\n\nW przysz\u0142o\u015bci, w obiekcie b\u0119d\u0105 realizowane kolejne filmy, nie tylko te komercyjne.\n\n- Konsekwencj\u0105 pojawienia si\u0119 Fabryki Form Filmowych, jest potrzeba dalszej rozbudowy obiektu. Tw\xf3rcy nie zapomn\u0105 o Olsztynie. Chc\u0105, aby fabryka sta\u0142a si\u0119 pr\u0119\u017cnie dzia\u0142aj\u0105c\u0105 wydarzeniem kulturalnym i edukacyjnym, daj\u0105cym prac\u0119 tysi\u0105com ludzi w Polsce. Jak i te\u017c nie zapominaj\u0105, \u017ce fabryka powsta\u0142a dzi\u0119ki Polakom, kt\xf3rzy dzia\u0142ali w tym miejscu pod nazw\u0105 Fabryka Filmowa &quot;Kolor&quot; - m\xf3wi prezes.]" time="1.266"><failure message="RuntimeError: The expanded size of the tensor (515) must match the existing size (514) at non-singleton dimension 1.  Target sizes: [1, 515].  Tensor sizes: [1, 514]">E   RuntimeError: The expanded size of the tensor (515) must match the existing size (514) at non-singleton dimension 1.  Target sizes: [1, 515].  Tensor sizes: [1, 514]</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Share Tweet Email Copy Link Copied\n\nTimeless creator Eric Kripke wants a season 4 to resolve the show's cliffhanger ending.\n\nThe time-travel series, starring Abigail Spencer, Matt Lanter, Malcolm Barrett and Goran Visnjic, was canceled after two seasons on NBC. However, after a last-minute bid from Sony Pictures Television, it was ordered to a special two-part finale. The shows were meant to resolve the cliffhanger, and answer why Rufus was shot and where Lucy and Wyatt ended up. However, when the episode aired, nothing was resolved.\n\nRelated: Timeless Fits Perfectly Into The Revival TV Movement\n\nIn an interview with Syfy Wire, Kripke revealed he's hoping for a season 4 so that the show can get a proper ending.\n\n&quot;We always conceived of [the finale] as a potential series finale, because we couldn\u2019t get canceled! So that\u2019s just what it ended up being. It\u2019s like, \u201cOK, we can\u2019t end the show, so let\u2019s end the arc.\u201d It felt really appropriate and full circle, and everybody at NBC, Sony and Warner Bros. was like, \u201cThat\u2019s a perfect place to end this.\u201d We\u2019re not ruling out a future for the show. Who knows? We still feel passionately about it. Maybe there will be a season 4. I think everyone at the studio and the network felt really good about where it ended.\u201d\n\nIf]" time="0.815"><properties><property name="score" value="0.17134180665016174" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.17134181&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.17134181
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Arnold R\xfc\xfctel\n\nArnold R\xfc\xfctel (born 3 April 1937) is an Estonian politician who was the third President of Estonia from 2001 to 2006. He was born in Rakvere, L\xe4\xe4ne-Viru County. He served as the Mayor of Tallinn from 1992 to 2001 and was the Speaker of the National Assembly from 1999 to 2001.\n\nR\xfc\xfctel served as a Colonel in the Soviet Air Defence Forces from 1967 to 1992. He joined the pro-independence party &quot;Vaba Eesti&quot; (Free Estonia) in 1988.\n\nIn the last Soviet-era election of the Supreme Soviet of the Estonian SSR, held on 17 March 1990, R\xfc\xfctel was elected to the parliament, representing the Tallinn constituency. On 20 August 1991, he was one of the speakers at the Congress of Estonia, which proclaimed independence. He was elected to the new Riigikogu (Parliament) on 18 September 1991, and became Minister of Foreign Affairs in the government of Mart Laar.\n\nR\xfc\xfctel joined the conservative Pro Patria Union party in 1993 and was elected chairman in 1996. He ran for president in 1999 and came in second to Arnold R\xfc\xfctel. He became a member of the Riigikogu again in 2003, this time for the Reform Party. R\xfc\xfctel is considered to be a pro-Western politician and a leading proponent of Estonia's membership of NATO. In February 2004, he was the first person to publicly endorse IRL politician Ene Ergma as the Reform Party's candidate for president in the scheduled elections. On 12 September 2004, he announced that he was joining the Union of Pro Patria and Res Publica party, whose candidate for president he had been in 1999.\n\nOn 23 September 2004, the Riigikogu elected R\xfc\xfctel to the office of President of Estonia. R\xfc\xfctel was sworn into office on 9 October 2004, and was re-elected in 2006.\n\nR\xfc\xfctel has held a number of positions in the Estonian government since 1991, including as Minister of Foreign Affairs, Minister of Education and Research, Speaker of the Riigikogu, and a two-year term as the Mayor of Tallinn (from 1992 to 1994).\n\nHe has also served as the Chairman of the National Defense Committee and was the leader of the European Movement in Estonia from 1996 to 2004.\n\nOn 20 February 2007, R\xfc\xfctel announced his intention not to run for re-election in the scheduled presidential election of 2007.\n\nHis first wife Tiina Urm (1942\u20132011) died on 9 May 2011, aged 69, from undisclosed causes. R\xfc\xfctel has one son, Peeter, born in 1968, from his first marriage.\n\nHis second wife is Kadri M\xe4lk, with whom he has a daughter, Triin, born in 1989.\n\n]" time="1.581"><properties><property name="score" value="0.34467093149820965" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Crescent-chested whitestart\n\nThe crescent-chested whitestart (&quot;Myioborus luteoventris&quot;) is a species of bird in the Parulidae family. It is found in the humid lowland forests of tropical South America. Its range extends from southeastern Colombia, eastern Ecuador, and northeastern Peru to northern Bolivia, southeastern Venezuela, and northern Brazil. It is a fairly common species within its small range, and the IUCN has rated it as a species of &quot;least concern&quot;.\n\nThe crescent-chested whitestart is a medium-sized songbird. The adult has a fairly long tail and a thin bill. It is mainly yellow-olive, with a white belly and vent, a yellow breast, and a yellow eyebrow. It has a black stripe over the eye and a whitish crescent on the breast. It is similar to the yellow-bellied white-start and the turquoise-fronted whitestart.\n\nThe crescent-chested whitestart is native to eastern South America, its range extending from southeastern Colombia, eastern Ecuador, and northeastern Peru to northern Bolivia, southeastern Venezuela, and northern Brazil. It inhabits humid tropical forest between sea level and about above sea level. It can be found at altitudes of , but generally below .\n\nThe crescent-chested whitestart is a fairly common species. No particular threats have been identified, and the International Union for Conservation of Nature has rated its conservation status as being of &quot;least concern&quot;.\n]" time="0.763"><properties><property name="score" value="0.9017813205718994" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Alejandro Leguizam\xf3n\n\nAlejandro Leguizam\xf3n (born October 30, 1961 in Santa Fe) is an Argentine journalist. He is a columnist of &quot;Noticias&quot; and &quot;P\xe1gina/12&quot;. Leguizam\xf3n also is a member of &quot;The Clinic&quot;, a leading group of Argentine intellectuals, and a writer of a weekly column in &quot;P\xe1gina/12&quot;.\n\nAlejandro Leguizam\xf3n is the son of Eduardo Leguizam\xf3n, founder of the &quot;C\xedrculo Militar&quot;, and brother of the TV presenter Mariana Leguizam\xf3n.\n\n]" time="0.310"><properties><property name="score" value="0.9981039762496948" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Friedrich Hofmeister\n\nFriedrich Hofmeister was a German piano and music publisher in Leipzig. His company, Friedrich Hofmeister Musikverlag, is the oldest music publishing company in Germany and one of the oldest in the world. It was founded in 1798. His catalogue contained more than 25,000 works, many published as composers' copies, without printed music.\n\nHofmeister was a friend and patron of Robert Schumann, a composer whose music was frequently published by Hofmeister, and a generous supporter of Felix Mendelssohn Bartholdy. A close friendship developed between Mendelssohn and Schumann, who lived at the same time.\n\nOne of the most important works of Schumann, his Piano Quintet in E-flat major, op. 44, is dedicated to Friedrich Hofmeister, as is Schumann's Piano Concerto in A minor, op. 54.\n\nIn his will, Schumann left his house to Mendelssohn.\n\nIn 2011, a new edition of the full catalogue was launched. The company today is a subsidiary of the German publisher B\xe4renreiter, founded in 1525.\n]" time="0.583"><properties><property name="score" value="0.9998143315315247" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Manigramam\n\nManigramam is a village in the Orathanadu taluk of Thanjavur district, Tamil Nadu, India.\n\nAs per the 2001 census, Manigramam had a total population of 2169 with 1068 males and 1101 females. The sex ratio was 994. The literacy rate was 74.41.\n]" time="0.215"><properties><property name="score" value="0.9499715566635132" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.94997156&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.94997156
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Claudia Escobar\n\nClaudia Esther Escobar Ram\xedrez (born 9 April 1995) is a Mexican professional football Midfielder who currently plays for Puebla in the Liga MX Femenil.\n\nEscobar debuted in the 2013\u201314 season in a match against Pachuca.\n\nIn July 2015, Escobar was transferred from C.D. Guadalajara to Puebla F.C..\n\nShe won the first Women's Football Mexican Championship in 2017, with Puebla F.C..\n]" time="0.298"><properties><property name="score" value="0.910408616065979" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.91040862&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.91040862
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Bucculatrix pomifoliella\n\nBucculatrix pomifoliella is a moth in the family Bucculatricidae. It is found in North America, where it has been recorded from Maryland, Ohio, Oklahoma, Texas and Maine.\n\nThe wingspan is 8\u201310\xa0mm. The forewings are white with pale brownish-grey lines. Adults have been recorded on wing in August.\n\n]" time="0.378"><properties><property name="score" value="0.8144510388374329" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Transport and Communications Workers Union\n\nThe Transport and Communications Workers Union (T&amp;CWU) is a trade union in Trinidad and Tobago, with the bulk of its membership employed by the state-owned National Petroleum Company (Petrotrin).\n\nThe union was founded in 1952 as the Oilfield Workers Trade Union, and in 1974 it became the Petroleum Workers Trade Union. In 1977, the union renamed itself as the National Oilfield Workers Trade Union, and then in 1987, the name was changed again to the National Oilfield Workers Union.\n\nThe union changed its name again in 1997 to its current form. In 2005, it merged with the Civil Service and Allied Workers Union, forming the Petroleum and Allied Workers Trade Union. The union re-split in 2007.\n\nFor many years, the union was led by Ancel Roget, and in 1997 he was convicted for soliciting a $40,000 bribe from an executive at the Shell Oil Company. In 2000, Roget was forced to retire from his position at Petrotrin due to a management investigation.\n\nIn 2006, the union was involved in a strike over the state of Petrotrin. In 2010, the union was accused of illegally striking.\n\nIn 2008, a branch of the T&amp;CWU was reported to have hired gang members to threaten striking Petrotrin workers.\n\n\n]" time="0.814"><properties><property name="score" value="0.9939083456993103" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Raymond T. McNally\n\nRaymond Thomas McNally (March 23, 1917 \u2013 March 15, 2005) was an American author, best known for his book, &quot;The Dracula Scrapbook&quot;.\n\nHe was born in Connecticut and was of Irish descent. He began work at the RKO film studios at the age of 19 as a film librarian. During the 1940s he was a freelance writer and served in the United States Army.\n\nMcNally was best known for his &quot;Dracula Scrapbook&quot;, a work which consisted of a compilation of newspaper reports, letters, affidavits, and legal depositions detailing the various efforts to suppress Bram Stoker's novel &quot;Dracula&quot;. He also wrote books on the films of Marilyn Monroe]" time="0.399"><properties><property name="score" value="0.9741377234458923" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.97413772&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.97413772
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[A/N: Before I start, I must confess. I lied. In an interview with Josh Radnor from earlier this year, he claimed he would only do a spin-off if it was centered around Ted and Barney, with the two of them going off to some far off place (New Zealand? Australia?) where they lived in a beach house and watched sports. I laughed. A lot. And now that I have come up with a sequel for The Only Person Who Didn't Run from Ted Mosby, I don't really have a plot for it...\n\n(Disclaimer: This is a work of fan fiction. Any resemblance to any other work of fiction is entirely coincidental. Not to mention a little insulting)\n\nI sat on my bed, looking down at my phone. There was only one person who could help me write this...\n\nI knew it was kind of a lame excuse. I mean, my sister works at a museum, for Pete's sake. She can't help me with a stupid paper. What was I thinking? And I had a really good excuse to tell my girlfriend when I get home. We had a date. But, in order to give me that good excuse, I have to write this stupid paper and get it done. I have to talk to my sister.\n\nI dialed the number to my sister's work, and it started ringing. My sister, Gloria, picked up.\n\n&quot;What's up, D?&quot;\n\n&quot;Glor, do you think you can help me with a paper? I really need some help.&quot;\n\n&quot;Sure. What is it about?&quot;\n\n&quot;It's a paper about JFK.&quot;\n\n&quot;Oh. Cool. Yeah, sure. What do you need?&quot;\n\n&quot;I need some help finding some good sources for this. I need to know how to start a paper.&quot;\n\n&quot;I'm glad you asked.&quot;\n\nAnd I was off to a good start. I sat at my desk, making notes. The paper was going to be really easy to write. It would probably only take a couple hours. I knew it. And I would be home before Lily.\n\nThe paper was starting to look really good. I had a bibliography in place, a couple of good paragraphs, and a nice introduction. Everything was great.\n\nI had a paper ready to turn in the next morning. It was almost done, and I was free to go home. But I had a date to get to.\n\n&quot;Bartender!&quot; I called. &quot;One more, please!&quot;\n\n&quot;You sure, man? You're not gonna make it home tonight.&quot;\n\n&quot;Nah, man. It's all good. I'll just be getting home, like, ten minutes later than I would.&quot;\n\n&quot;Okay, man. Whatever you say.&quot;\n\nI started walking out the door, but I took the bar stool with me. I had to carry it around to make it look like I was in a drunken stupor. If my girlfriend saw that I was drunk when I got home, I would never be able to keep my secret.\n\nAs I made my way through the train station, I heard a strange beeping sound. I turned my head to the side, trying to find the source of the noise. It was coming from my phone. There was a text message. I opened my phone, and read it.\n\nI was starting to feel a little tipsy. It was a good thing I was carrying a bar stool around. It made me look like I had a reason for not walking right.\n\nThe message was from Lily. I read it.\n\nThe good thing about living in New York was the night life. I was lucky enough to be able to spend an evening at a nightclub, so I would be home in no time.\n\n&quot;Yeah, man,&quot; I said. &quot;I don't know why I even did this, dude.&quot;\n\nI walked out of the nightclub and back into the train station. I knew it was late, but the train wasn't too crowded. I sat down in a seat, trying to get some sleep. I would be home soon enough.\n\nI woke up as the train pulled into the station. I was close to home now. It was getting a little crowded on the train, but I managed to squeeze myself into the seat. I was a little more aware of the fact that I was still carrying my bar stool, but I decided to go on anyway.\n\nI stumbled out of the train station, still carrying the bar stool. I made my way through the crowd, out into the parking lot, and across the street to my car. I put the bar stool in the trunk, and got in the car. I turned on the engine, and drove off.\n\n&quot;Hey, Lily. Yeah, I'm on my way home. I'm sorry, but I ran into some trouble. It shouldn't take me long to get home.&quot;\n\n&quot;You said that you were just going to be a few minutes late. You said you were on your way home, and then you don't show up for three hours. You call me and say you're sorry? That doesn't even make sense!&quot;\n\n&quot;I know, I know. It's just that it took me a little longer to get home than I thought.&quot;\n\n&quot;You lied to me! You said you were going to be home at 8:30! I waited until midnight! You were just trying to get out of our date!&quot;\n\n&quot;It wasn't like that, Lily. I had some trouble with a paper, and then I was at a nightclub. I had a lot of stuff on my mind.&quot;\n\n&quot;So, you didn't want to come home. You just didn't want to spend time with me. And now you're calling me after it's all over and trying to explain yourself. And I don't want to hear it! Goodbye!&quot;\n\nI was standing at the door to my apartment, keys in my hand, when I heard my phone ring. It was Lily. I stood there for a moment, trying to decide whether or not I should answer. I looked down at the floor, and then back at my phone. I answered.\n\n&quot;Lily, I-&quot;\n\n&quot;I don't want to hear it, Ted. I told you that I don't want to hear it.&quot;\n\n&quot;No, no, it's just that I really have to tell you something. I can't wait until tomorrow. I really need to tell you right now.&quot;\n\n&quot;Well, fine. What do you need to tell me? Spit it out.&quot;\n\n&quot;It's just that...I think that...&quot;\n\n&quot;Ted, you're babbling. What is it?&quot;\n\n&quot;I...I...well...&quot;\n\n&quot;Ted? Ted! You're breaking up. I think the connection is getting worse. What was it?&quot;\n\n&quot;I think we should break up.&quot;\n\n&quot;What?&quot;\n\n&quot;I think we should break up.&quot;\n\n&quot;What?&quot;\n\n&quot;I...I...you were right.&quot;\n\n&quot;What?&quot;\n\n&quot;I didn't want to spend the night with you. I didn't want to be around you anymore. I wanted to get out of the date. I know it's not fair. I know I was a jerk. I know I lied to you. I just...I don't think we should be together anymore.&quot;\n\n&quot;What? Why?&quot;\n\n&quot;I don't know. I just don't.&quot;\n\n&quot;Ted, I love you. You can't just say that we should break up like this.&quot;\n\n&quot;You were right. You shouldn't have to date me.&quot;\n\n&quot;Ted...&quot;\n\n&quot;I'm sorry. I really am. I didn't mean to hurt you. You're a wonderful girl. I just...I don't know. I'm sorry.&quot;\n\n&quot;Ted...&quot;\n\n&quot;Goodbye, Lily.&quot;\n\nI hung up the phone, and dropped it to the floor. I was standing there for a moment, trying to get my emotions in check.\n\nI had done it. I broke up with my girlfriend of three years. I had broken up with the girl I loved. I had no idea why I did it, but I was doing it.\n\nI opened my door, and made my way to my bedroom. I dropped my keys on the dresser, and sat down on my bed. I leaned against the wall, and closed my eyes.\n\n&quot;Hey, Ted. What are you doing home so early?&quot;\n\n&quot;Lily, I...&quot;\n\n&quot;Ted, I know I was mad at you before, but I've calmed down a lot. You know I love you. I can't live without you.&quot;\n\n&quot;Lily, I...&quot;\n\n&quot;Ted, you just scared me, is all. I thought I lost you. I thought you didn't want to be with me anymore.&quot;\n\n&quot;Lily, I...&quot;\n\n&quot;I know it was wrong for me to keep pushing you to make up your mind about me, but you couldn't see that.&quot;\n\n&quot;Lily...&quot;\n\n&quot;And I know that I can be a little difficult to get along with sometimes, but I know you'll come around.&quot;\n\n&quot;Lily, I...&quot;\n\n&quot;We're supposed to be together, Ted. We always knew that. You and me, we're meant to be. You can't just do something like this, Ted.&quot;\n\n&quot;Lily...&quot;\n\n&quot;I love you, Ted. You can't just throw that away. I can't just throw that away. I can't...&quot;\n\nI opened my eyes, and saw Lily]" time="5.141"><properties><property name="score" value="0.9446615055203438" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Former Arizona Sen. John McCain died after a yearlong battle with brain cancer.\n\nMy fellow Republicans, John McCain was a true American hero. https://t.co/YlNEJ2KRAS \u2014 Jake Tapper (@jaketapper) August 26, 2018\n\nArizona Sen. John McCain, the maverick Republican whose war record and outspokenness against abuses of power made him an American icon, died on Saturday. He was 81.In a statement, McCain\u2019s office said he died at 4:28 pm local time at the family\u2019s home in Arizona surrounded by family members. The cause of death was not immediately known.The former Navy pilot, who was imprisoned and tortured during the Vietnam War, rose to the top of American politics after emerging as a prominent critic of President Donald Trump.An upstart presidential bid in 2000 did not succeed but after he won the Republican nomination in 2008, he lost to Democrat Barack Obama.Despite the loss, his image soared and he emerged as one of the most influential members of the Senate, willing to work with Democrats on issues like campaign finance reform.By the time he ran for president again in 2016, he seemed like a throwback to a bygone era of civility and bipartisanship. McCain ran against Obama again in 2008.He continued to play a pivotal role in American politics, becoming chairman of the powerful Senate Armed Services Committee after the Republicans regained control of the Senate in 2015.McCain in the past had been a thorn in the side of Republican presidents \u2014 George W. Bush and Trump.He was also well known for reaching across the aisle.He took a secret trip to visit Syrian dictator Bashar al-Assad during the early years of the country\u2019s civil war, and was one of the final Republicans to speak with the president before his firing last year.He was a long-time critic of the U.S. use of torture against detainees.He held out against Republicans\u2019 efforts to dismantle Obama\u2019s signature healthcare legislation and refused to support the Graham-Cassidy bill last year.He was a harsh critic of Russia, and one of the earliest supporters of sanctions against Russian officials. He played a role in seeking to keep the United States from launching a pre-emptive attack against North Korea in the face of its nuclear provocations.McCain was one of the first Republicans to support Trump\u2019s presidential campaign in 2016. But when the \u201cAccess Hollywood\u201d tape emerged, he withdrew his support, and famously said in his 2017 memoir \u201cThe Restless Wave,\u201d \u201cI do not know how I could have spoken so disrespectfully about the commander in chief of the Armed Forces.\u201dHe also repeatedly denounced what he called Trump\u2019s \u201ccheap rhetoric and insults\u201d against U.S. allies, and after Trump\u2019s summit with Russian President Vladimir Putin in July 2018, called it a \u201ctragic mistake.\u201d He was one of the loudest Republican voices criticizing the president for his performance at a summit with Putin in Helsinki.Trump mocked him as \u201ca clown,\u201d and even at the end of his life, he was still a target for the president, who reportedly resisted hosting a state funeral.On Saturday, Trump said he sent his \u201cdeepest sympathies and respect\u201d to McCain\u2019s family.\u201cThe nation mourns the loss of a great American patriot, a statesman who put his country first and enriched this nation every day he served.\u201dFormer President Barack Obama called him a \u201cgiant of American politics,\u201d and his one-time opponent in the 2008 presidential race said he was \u201ca man of deep conviction and a patriot of the highest order.\u201dPresidential candidate Sen. Bernie Sanders, I-Vermont, called McCain \u201can American hero\u201d on Twitter.McCain\u2019s funeral service will be held at the Washington National Cathedral and the ceremony will include eulogies from former President George W. Bush, former President Barack Obama, former Secretary of State Henry Kissinger and former Sen. Joe Lieberman, the former vice-presidential candidate who was a close friend.After the service, he will be buried in Annapolis, Md., at the U.S. Naval Academy.McCain was born at a naval station in the Panama Canal Zone, where his father was stationed. After his father\u2019s career in the Navy, the family settled in the Arizona desert, where he attended high school.McCain studied at the Naval Academy and became a pilot in the Navy, serving on a ship that was shot down in Vietnam. He spent five-and-a-half years in captivity, which included torture, before he was released in 1973.McCain returned to the U.S. and was assigned to the Senate as a liaison to the Navy. He became a member of the House of Representatives in 1983 and eventually the chairman of the Senate Committee on Commerce.During his 2008 presidential bid, he secured the GOP nomination in the face of much skepticism. At a dinner in 2008, then-Sen. Barack Obama joked that after losing to Obama, McCain would be on a beach sipping a drink with \u201can umbrella in it.\u201dThe Republican-led Senate, which was much closer to McCain than to the president, overwhelmingly passed a resolution earlier this year that censured Trump for his \u201cbigoted and racist\u201d remarks about four members of Congress.In his memoir, McCain said that he regretted that his choice of Sarah Palin as a running mate did not propel him to the White House. He said he made the choice based on who he thought could become president, but that was wrong.\n\nJoin Jerusalem Post Premium Plus now for just $5 and upgrade your experience with an ads-free website and exclusive content. Click here&gt;&gt;\n\n\n\n]" time="2.997"><properties><property name="score" value="0.6243777374426523" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Avolites V-Color 2\n\nV-Color is an all-in-one visualisation system for lighting professionals which empowers operators with unprecedented control and flexibility, while also offering total ease-of-use. A complete visualisation solution for the modern production environment.\n\n\n\nNew V-Color 2 Software\n\n\n\nUsing the same hardware, V-Color 2 offers full compatibility with existing shows, but with brand new software including some great new features and exciting innovations.\n\n\n\nMore Detailed, Better Feedback\n\nThe addition of 3D colour bars gives a better representation of colour than before. The inclusion of warm/cool &amp; complementary information makes diagnosing colour temperature easier than ever. An inbuilt screen lock function and scene copy capability further streamlines the user experience.\n\n\n\nEasier Control of Everything\n\nWith a large, backlit LCD screen, V-Color 2 gives the operator total control of the system from a distance. The touchscreen interface makes everything intuitive and gives instant access to show information and control options.\n\n\n\nSeamless Integration into Live Production Workflow\n\nPowered by ProfilerX (firmware version 2.3 and above), V-Color 2 gives the operator seamless and continuous control over their show using the same industry standard protocol as their other devices. This means that they can operate their visualisation system in parallel with their lighting console, creating the ideal platform for shows involving complex automated and DMX control.\n\n\n\nBigger, Better, More Powerful\n\nThe increased size of V-Color 2 gives it the space to fit a larger screen with a better resolution. The dual menu and page system makes navigation easier than ever and the addition of a USB port and SD card slot allows the operator to save and play back their show in a simple and fast manner.\n\n\n\nFeatures:\n\n\u2022 Upgradeable for full ProfilerX support.\n\n\u2022 Highly responsive and intuitive touchscreen interface.\n\n\u2022 Fits into any lighting desk.\n\n\u2022 Dual menu system.\n\n\u2022 USB &amp; SD card slot for show playback.\n\n\u2022 Upgraded 3D colour bars.\n\n\u2022 Larger display for better visibility.\n\n\u2022 ProfilerX compatible.\n\n\u2022 Industry standard OSD protocol.\n\n\u2022 Relay channel output for all screens.\n\n\u2022 Separate input and output cables.\n\n\u2022 Quick calibration system.\n\n\u2022 Assignable RS232 port.\n\n\n\nSpecifications:\n\n\u2022 VGA &amp; HDMI Input.\n\n\u2022 TFT screen.\n\n\u2022 WVGA screen resolution.\n\n\u2022 USB &amp; SD card slot.\n\n\u2022 Can be fitted to V-Pad stand.\n\n\u2022 Size: 10.1&quot; x 5.9&quot; x 1.7&quot;.\n\n\u2022 Weight: 1.2kg.]" time="1.340"><properties><property name="score" value="0.3330735961596171" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.3330736&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.3330736
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Video\n\nIt was the final episode of the ITV sitcom series Night and Day.\n\nSet in a TV station, the programme starred Barbara Flynn and William Gaunt as the heads of rival departments - the news and the features.\n\nAs BBC reporter Paul Collard watched the final episode in 1989, he wondered what might have happened to some of the characters.\n\nProducer: Bill Powell]" time="0.244"><properties><property name="score" value="0.7977500557899475" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.79775006&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.79775006
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Jason Kreis became a bit of a legend at the University of Pennsylvania. Not only did he make two trips to the College Cup, leading the Quakers to their first ever national championship in 1998, but he also played a critical role in the Penn men's soccer team's decision to make the historic transition from the NCAA Division I to Division III. While he was at the helm of the Red and Blue, he produced such a respectable record (70-32-16) that Penn was actually invited to compete in the Northeast Conference, a Division I conference at the time. Although he is now best remembered for that phenomenal campaign in 1998, he remains a very active member of the Penn community.\n\n\n\n&quot;It's pretty interesting that my last name is 'Kreis' because in Germany, that means 'square' or 'squared,' but]" time="0.399"><properties><property name="score" value="0.9994845390319824" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Power Consumption\n\nIdle: Windows Vista Aero sitting at the desktop (1280x1024 32-bit) all windows closed, drivers installed. Card left to warm up in idle mode until power draw was stable.\n\nMulti-monitor: Two monitors connected to the tested card, both using different display timings. Windows Vista Aero sitting at the desktop (1280x1024 32-bit) all windows closed, drivers installed. Card left to warm up in idle mode until power draw was stable.\n\nAverage: Crysis 2 at 1920x1200, Extreme profile, representing a typical gaming power draw. Average of all readings (12 per second) while the benchmark was rendering (no title/loading screen).\n\nPeak: Crysis 2 at 1920x1200, Extreme profile, representing a typical gaming power draw. Highest single reading during the test.\n\nMaximum: Furmark Stability Test at 1280x1024, 0xAA. This results in a very high non-game power-consumption that can typically be reached only with stress-testing applications. Card left running stress-test until power draw converged to a stable value.\n\nBlu-ray Playback: Power DVD 9 Ultra was used at a resolution of 1920x1200 to play back the Batman: The Dark Knight disc with GPU acceleration turned on. Playback started around timecode 1:19, which has the highest data rates on the BD with up to 40 Mb/s. Playback was left running until power draw converged to a stable value.\n\nCooling modern video cards is becoming more and more difficult, especially when users are asking for quiet cooling solutions. That's why the engineers are now paying much more attention to power consumption of new video-card designs.For this test, we measured the power consumption of the graphics card only via the PCI-Express power connector(s) and PCI-Express bus slot. A Keithley Integra 2700 digital multimeter with 6.5-digit resolution was used for all measurements. Again, the values here reflect only the power consumption of the card measured at DC VGA card inputs, not of the whole system.We choseas a standard test representing typical 3D gaming usage because it offers the following: very high power draw; high repeatability; is a current game that is supported on all cards because of its DirectX 9 roots; drivers are actively tested and optimized for it; supports all multi-GPU configurations; test runs in a relatively short time and renders a non-static scene with variable complexity.To measure the power consumption of our testbed we used a digital power meter (Keithley Integra 2700) connected before the power supply. The power consumption measurement of the system is based on the power consumption of the drives, not of the graphics cards. All cards were tested with a total of four SSDs: one in the system, one used as a primary test drive in a slot without a powered connector, and two additional drives in the system (not used for testing). The SSDs are Western Digital WD360GD (362GB), WD1500HLHX (150GB), Samsung SSD 830 (128GB) and OCZ Agility 3 (240GB).Idle power consumption of the whole test system (at the desktop) is measured after sitting idle at the desktop for 30 minutes. The cards are not closed. The result is presented as the average value of all four cards.Please note that after extensive testing, we found that measuring idle power consumption only, as seen in many other reviews, can be misleading. Due to different testing methodology and environments, this number can be significantly lower than the real power consumption. Also, in many cases, the card does not reach the desired idle temperature, so it will draw more power. In our test setup, the cards reach their idle temperature very quickly (in a few minutes) and in a climate controlled room (22\xb0C).Therefore, we use this only as a rough indication of the cards' power consumption.Please note we are using different adjustements to test the performance of our graphics cards. Our testing methodology is detailed here . It's worth mentioning that our method allows us to perform a more thorough test that's free from GPU bottlenecks. It's also free from variations caused by dual-gpu/hybrid configurations.The following two values are the peak power consumption and the maximum power consumption of our testbed (without speakers/headphones) measured after long periods of use (at least 15 minutes) with Windows Aero enabled. The peak power consumption is the maximum power draw of the graphics card only, measured in the most stressful situation (generally, when it's overclocked). The maximum power consumption is the maximum power draw of the complete testbed, including the monitor, speakers, hard-disk, SSD, etc.The results are impressive. Idle power consumption is a tiny bit higher than that of the GTX 580. All cards operate at fairly low power levels in idle. NVIDIA has done a good job reducing idle power consumption of its GTX 680. Under full load, the GTX 680 is only slightly more power hungry than the GTX 580. Compared to AMD, the GTX 680 fares even better as it uses slightly less power than the HD 7970.All graphics cards tested consume quite a lot of power during Blu-ray playback. Interestingly, this is one area where NVIDIA can't keep up with AMD.]" time="2.669"><properties><property name="score" value="0.6072969913482666" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.60729699&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.60729699
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[This is an experiment in one-shot plays. I've been inspired by the weekly news and what it tells us about the human condition. I call this &quot;Instructional Theatre&quot; because these plays are intended to instruct the audience in a point of view. And they are performed in one sitting. Here is the first one, 'If I Only Had A Wife' by Henrik Ibsen.]" time="0.238"><properties><property name="score" value="0.9996129870414734" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[A bi-partisan bill that would give states the option to spend welfare funding on faith-based charities is making its way through the U.S. Senate.\n\nThe legislation is called the Combating and Preventing Crimes of Violence Act, and it would give states the ability to use federal funds on programs that promote &quot;the traditional values of marriage, family, religion, and morality.&quot; The bill would also authorize $5 million annually to states to promote chastity and healthy marriage.\n\nState governments would be permitted to use federal welfare funding to teach young people &quot;the social, psychological, and health gains to be realized by abstaining from sexual activity,&quot; according to the text of the bill, which was introduced by Sen. Orrin Hatch (R-Utah) and 23 other co-sponsors.\n\nSen. Hatch said the legislation would reinforce &quot;state efforts to strengthen families and communities by increasing opportunities for the social and economic advancement of young people.&quot;\n\nSen. Hatch's home state of Utah passed a similar law in 2013, and it was successfully implemented, according to a statement from the senator's office.\n\n&quot;This is not an effort to allow taxpayer funding for religious groups, but to give states the option to partner with groups to promote healthy relationships,&quot; the senator's statement said.\n\nHowever, the American Civil Liberties Union (ACLU) criticized the legislation as &quot;misguided&quot; and &quot;misinformed,&quot; saying that it would restrict the freedom of organizations that provide critical services to women and children.\n\n&quot;We have grave concerns about this bill because it's divisive, discriminatory, and fiscally irresponsible,&quot; Laura Murphy, director of ACLU's Washington Legislative Office, said in a statement. &quot;It would allow states to use federal funds to provide child welfare services in a way that would exclude those who are not religious or who don't subscribe to traditional family values. This means that women and children who need assistance would be turned away, and those who provide critical services would be fired or not hired.&quot;\n\nThe bill, which was approved by the Senate Committee on Finance on Wednesday, now heads to the Senate floor.]" time="1.025"><properties><property name="score" value="0.9773589173952738" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[HUNDREDS OF people are expected to attend a public meeting in Clonmel this evening which will focus on the recent developments in the Irish Water saga.\n\nClonmel Borough Council has granted permission for the meeting to be held in the Tower Hotel at 7pm.\n\nGarda\xed have also been notified of the meeting but a spokesperson said they are not expecting any trouble.\n\nThe meeting comes in the wake of a massive turnout at a public meeting on Sunday last where thousands of people flocked to a hall in Carrick-on-Suir, Co Tipperary to hear speakers on the Irish Water issue.\n\nThe Tipperary Star reported today that a Fine Gael representative has complained to the party\u2019s headquarters about a leaflet produced by the Carrick on Suir branch of the party.\n\nThe leaflet called on people to attend the meeting on Sunday last in order to oppose Irish Water and the installation of water meters.\n\nSpeaking to the Star, Fine Gael TD for Tipperary South Noel Coonan said the leaflet had damaged the image of Fine Gael in the area.\n\nHe also described it as a \u201cdistortion of facts\u201d.\n\n\u201cIf this is being done in the name of Fine Gael, it is very serious,\u201d he said.\n\n\u201cI was very disappointed to see this leaflet which was distorting the facts. I\u2019m very annoyed about this and I will be taking it further.\u201d\n\nCouncillor Coonan also questioned the language used in the leaflet and the manner in which the information was presented.\n\nFine Gael councillor in Carrick on Suir, Councillor Martin Browne defended the leaflet, saying that it was a statement from the party\u2019s national executive and was a totally legitimate use of the party\u2019s logo.\n\n\u201cIt was used to promote a public meeting which was organised by a totally independent group,\u201d he said.\n\nThe group was not aligned to the party in any way and was not part]" time="0.945"><properties><property name="score" value="0.9931741058826447" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Horsehead\u2013Saxonburg Bridge\n\nThe Horsehead\u2013Saxonburg Bridge was a single-track, long railroad bridge in Armstrong County in the U.S. state of Pennsylvania. It carried the Bessemer and Lake Erie Railroad (B&amp;LE) main line across the Allegheny River and was the highest rail line on the B&amp;LE, with a maximum elevation of .\n\nThe bridge connected the small hamlet of Saxonburg on the west side of the river with Horsehead Landing on the east side.\n\nIn 2016, the Pennsylvania Department of Transportation announced plans to replace the bridge with a new one. The current bridge will be removed in the summer of 2019.\n\n]" time="0.347"><properties><property name="score" value="0.9989650249481201" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[John Astley (diplomat)\n\nJohn Francis Melhuish Astley, CMG (born 2 January 1950) is a former British diplomat.\n\nAstley was educated at St Paul's School, London, and the University of Cambridge, where he was a member of St John's College.\n\nAstley joined the Foreign and Commonwealth Office (FCO) in 1973. He served as Consul in Santa Cruz, Bolivia, 1973\u20131975, Deputy Consul General in San Francisco, California, 1981\u20131984, Deputy Head of the Middle East Department, FCO, 1985\u20131987, Private Secretary to the Secretary of State for Foreign and Commonwealth Affairs, 1987\u20131989, and Ambassador to Bahrain, 1990\u20131993.\n\nIn 1993 Astley became the first Permanent Under-Secretary of State at the FCO since the Second World War. He was appointed Companion of the Order of St Michael and St George (CMG) in the 1995 New Year Honours. From 1996 he was Ambassador to the Netherlands.\n\nIn 1999, during his tenure as Ambassador to the Netherlands, he was informed that his name appeared on a list of paedophiles. He had no connection to the accusations, but he resigned anyway, rather than face public humiliation.\n\nFrom 2000 to 2006 Astley was Chair of the Northern Ireland Audit Committee, an all-party group of the Northern Ireland Assembly, reporting to the Assembly and the Northern Ireland Audit Office.\n\nAstley was a director of Intraco Management Services Ltd, a recruitment firm.\n\nIn 2006 he was appointed a non-executive director of the Ministry of Defence's Defence Business Services organisation, where he remained until 2010.\n\nAstley is a member of the Athenaeum Club.\n]" time="0.851"><properties><property name="score" value="0.0522041916847229" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05220419&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.05220419
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Beata D\u0105browska\n\nBeata D\u0105browska (born 8 January 1967) is a Polish fencer. She competed in the women's individual and team \xe9p\xe9e events at the 1988 Summer Olympics.\n]" time="0.177"><properties><property name="score" value="0.9971202611923218" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99712026&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99712026
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Wilfred Harris (cricketer)\n\nWilfred James Harris (1884 \u2013 16 January 1935) was an English cricketer. He played two first-class matches for Cambridge University Cricket Club in 1905.\n\n]" time="0.168"><properties><property name="score" value="0.6331624984741211" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Stephen W. Kearney\n\nStephen Watts Kearney (June 17, 1794 \u2013 October 31, 1848) was an officer in the United States Army who is best known for his leadership in the Mexican\u2013American War, in particular the Battle of Monterrey and the Battle of Buena Vista.\n\nKearney was born in Newark, New Jersey. He was a descendant of John Kearney (1635\u20131701), a native of Ireland who was the first of the family to settle in what became the United States. John's son Richard (1668\u20131734) was born in Bristol, England, but became a fur trader in what was then the English colonies and later became New Jersey. Richard was commissioned as an ensign by British colonial governor Jonathan Belcher in 1709.\n\nRichard's son James (1703\u20131785) was a general in the New Jersey colonial militia during the French and Indian War. James's son Stephen Watts Kearny, the namesake of the general, was a captain in the War of Independence and was at the surrender of General John Burgoyne. His home, the Kearney Mansion, has since been preserved in Newark.\n\nIn 1794 he moved with his parents to Louisville, Kentucky. His father, Alexander Morris Kearney, a native of South Carolina, was a captain in the American Revolutionary War. In the War of 1812, Kearney participated in the Battle of Lundy's Lane, where he was wounded and captured by the British. After being released, he joined the Missouri Rangers and, in 1813, married Mary Radcliffe Matthews. In 1814 he joined the 5th United States Infantry as its sixth lieutenant and served in the war against the Creek nation. In 1819 he was promoted to captain and assigned to the 1st U.S. Infantry. Kearney resigned from the Army in 1822 and became superintendent of Indian affairs at St. Louis, Missouri.\n\nIn 1832 Kearney was appointed as governor of the New Mexico Territory, where he was active in promoting settlement and negotiating treaties with the Native Americans. Kearney participated in the Gila Expedition of 1837 and then commanded the Army's Third Dragoon Regiment in Florida in 1837 and 1838.\n\nDuring the Mexican\u2013American War, Kearney was appointed colonel of the 1st New York Dragoons on June 30, 1846. Kearney's new regiment was sent to Santa Fe, New Mexico, to reinforce the Army of New Mexico under Col. Stephen Watts Kearny (no relation). There they joined the march on the capital of Mexico City. On the way, Kearney's regiment carried out several successful raids, including the Battle of La Mesa, where he captured a large amount of ammunition and supplies. He took the Puebla de Los Angeles without opposition on December 7, 1846. There he received word that his promotion to brigadier general had been approved.\n\nKearny, reinforced by the 3rd Infantry and Mormon Battalion, took the city of Monterrey on September 25, 1846. During the battle, a cannonball tore away part of Kearny's jaw and neck.\n\nFollowing the battle, Kearney was carried to Saltillo for medical treatment. While he was recovering, Lt. Colonel John Garland was given command of the 1st Dragoons. In February 1847, Kearney returned to the army and resumed command of the dragoons. Kearney served in the Battle of Buena Vista, and later he fought in the Taos Revolt. He returned to the United States in October 1847 and was assigned to Fort Leavenworth in the Kansas Territory.\n\nKearney, a bachelor, died of cholera at Fort Leavenworth at the age of 54. He was buried in the Fort Leavenworth National Cemetery.\n\nSeveral U.S. Army installations are named in honor of Kearney, including Kearney Barracks in Wilhelmshaven, Germany, the Kearney Building at Fort Leavenworth, and Fort Kearny, Nebraska, which is now a state historic site.\n\n\n]" time="2.023"><properties><property name="score" value="0.9936164915561676" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Los R\xedos Province, Chile\n\nLos R\xedos (Spanish for &quot;The Rivers&quot;) is one of four provinces in the southern Chilean region of Los R\xedos (IV). The capital is Valdivia.\n\nAs a province, Los R\xedos is a second-level administrative division of Chile, governed by a provincial governor who is appointed by the president.\n\nThe province comprises eleven communes (Spanish: &quot;comunas&quot;), each governed by a municipality consisting of an alcalde and municipal council:\n\nThe province spans an area of , the second largest in the region. According to the 2002 census, Los R\xedos Province had a population of 253,331 inhabitants (133,315 men and 120,016 women), giving it a population density of . Of these, 119,262 (46.4%) lived in urban areas and 134,069 (53.6%) in rural areas. Between the 1992 and 2002 censuses, the population grew by 1.9% (4,243 persons).\n\nAs a province, Los R\xedos is a second-level administrative division of Chile, which is further divided into eleven communes.\n\nWithin this province are the most important archaeological ruins in all of southern Chile. The ancient city of &quot;Puquios&quot;, capital of the &quot;Lentas&quot; nation, is located in the commune of Valdivia. Other ruins include those at Chep\xe9n, Chichihualco and the ancient fortifications of Valdivia. Also of interest is the &quot;Cueva del Milod\xf3n&quot;, which features the remains of an extinct species of giant sloth.\n\nThe native tree &quot;Fitzroya cupressoides&quot;, locally called &quot;lauca&quot; or &quot;lauco&quot;, is a common tree in the forests of the coastal mountains.\n\nThe Rupanco and Quinchao Islands in the Chonos Archipelago are part of Los R\xedos.\n\nLike other southern regions of Chile, Los R\xedos experiences a moderate climate. Its weather is affected by the cold Humboldt Current which flows along the west coast of South America from Antarctica. During the summer months, temperatures can reach , while in the winter temperatures hover around .\n\nThe province is one of the most forested ones in Chile. Its economy is based on timber, salmon and tourism.\n\nRupanco and Quinchao Islands are part of the commune of Quinchao. The latter, together with a few minor islands, forms the &quot;Ciudad Insular&quot; (Inhabited Island) of Quinchao, with approximately 3,500 inhabitants, and is administered as part of the &quot;Comuna&quot; of Quinchao, in the province of &quot;Llanquihue&quot;.\n\n]" time="1.409"><properties><property name="score" value="0.29102055231730145" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[In many places there are no longer any young men at all. According to the World Health Organization, &quot;in countries with very low fertility, the proportion of young people may halve in the next 40 years&quot;. At this rate, by 2100 a majority of the world's population will be over 50. We will then have to find a way to accommodate the needs and aspirations of an unprecedented number of very old people in a society that has been moulded for decades around the needs of youth.\n\nNowhere is this more true than in Japan, the world's fastest-ageing country. The greying of Japan \u2013 &quot;silver tsunami&quot; or &quot;grey goo&quot; as it is known \u2013 is by now well known. There are a number of reasons for this. Japanese have one of the longest life expectancies in the world, with women expected to live to 87 and men to 80. Meanwhile, a long-standing national emphasis on having small families and long working hours \u2013 for both men and women \u2013 have kept the birth rate below replacement levels. And Japan's postwar economic miracle has not been kind to the young. The country's population is shrinking, partly because of a low birth rate and partly because of a high level of immigration.\n\nBut what is less well known is that this graying of Japan \u2013 and other nations such as Germany, China, South Korea and Thailand \u2013 is not just due to longer life expectancy and lower birth rates. There is another factor in play, and one that could have dramatic implications: very low fertility rates, particularly among younger women. It's estimated that around a quarter of the population will be aged over 65 by 2050.\n\nIn Japan, the &quot;total fertility rate&quot; (the number of children a woman is likely to have in her lifetime) is just 1.4 \u2013 and has been for the past 20 years. Among women aged between 15 and 24, the rate is even lower at 0.8, lower even than the famously low rate in Germany (which is 1.38). And in Tokyo, that rate is even lower at 0.7. By contrast, in 2010, the average rate in Europe was 1.6. The fertility rate in England and Wales in 2010 was 1.9.\n\nThe world's media are full of breathless stories about Japan's crisis of depopulation. Last year, an article in the International Business Times said that &quot;Japanese women are rejecting marriage and motherhood at a staggering rate&quot;. In the New York Times, another piece claimed that Japan is &quot;exporting its young women&quot; and is now &quot;facing a demographic crisis of epic proportions&quot;. More soberly, the BBC's business editor Robert Peston has written: &quot;Without a dramatic increase in its birth rate, Japan will face economic and social ruin.&quot;\n\nThe sense of panic is understandable. &quot;The number of young people will go down by two-thirds by 2035,&quot; says demographer Toshio Kanazawa. &quot;That means not only a decline in the size of the workforce but a dramatic increase in the number of retirees. Who will pay for them?&quot;\n\nThe answer is &quot;women&quot; \u2013 or more specifically, their wombs. And in this, there is something not quite right.\n\nIn Tokyo, I meet Nana, a 29-year-old who is employed in the financial sector. She is smartly dressed in a suit and speaks good English. She tells me she has a boyfriend and they are very happy. She is, she says, the typical &quot;She Economy&quot; \u2013 a young woman working in the city and her parents are very pleased. &quot;They want me to settle down,&quot; she says, &quot;and have children.&quot; But she's not so sure.\n\n&quot;It's too late,&quot; she says. &quot;I've got to think of my future. And I just don't know if I could afford it \u2013 having children. I'd need a bigger flat, and pay for childcare, and a new car.&quot;\n\nFor Nana, and for many Japanese women like her, the economic reality of having children is daunting. &quot;I have two brothers,&quot; she says, &quot;and they have wives and children. The wives don't work, and they have one or two children. One of my sisters-in-law doesn't even work outside the house. They don't have much money, but my parents love them. They are very happy. So I know what my parents want for me. But it's impossible. I can't do it.&quot;\n\nThis is a scenario that is playing out in homes across the country. The situation is, in many respects, rather unique. As Peston points out, Japanese women who stay in employment tend to have children later. As a result, while there is a generation of Japanese in their early 30s who have given up on having children, there is a generation of younger women who are still trying to get pregnant.\n\n&quot;I think it's pretty much impossible to have two or three children in Tokyo,&quot; says Seiko, a 35-year-old marketing executive. &quot;I don't think I could afford it. You have to pay so much for childcare. And then, you've got to give up work. I just couldn't do it.\n\n&quot;But, still, my mother is pressuring me to have children. If you have a boyfriend, you can't get married. It's impossible. You have to marry first, then have children. So, if you are a working woman, you have to get married, have children, and then stay at home with your children. That's it. And that's what everyone wants. That's the ideal.&quot;\n\nAt the root of this ideal is Japan's traditional gender-role society \u2013 one that has stayed remarkably the same since the end of the second world war. It is not, perhaps, surprising that such a rigid society is under strain. And not only for women. It is also under strain for men.\n\n&quot;Women in Japan are not treated well,&quot; says Toshio Kanazawa. &quot;They don't earn a lot of money. They are discriminated against in many ways. That's why there aren't enough babies.&quot;\n\nNot everyone agrees. &quot;I don't think it's true,&quot; says Nana, who, I've noticed, has referred to herself as &quot;boku&quot; (which means &quot;I&quot;) all evening, rather than the more feminine &quot;watashi&quot; (which means &quot;I&quot;). &quot;Women in Japan are strong. They do well at work and they earn a lot of money. They can do everything. So why should they have children?&quot;\n\nIn some respects, Nana has a point. Japan is a country in which young women are breaking glass ceilings. The number of women in senior management in Japan's 3,000 largest companies is at a record high. And in 2013, the World Economic Forum rated Japan the fourth most &quot;gender-equal&quot; nation in the world.\n\nThe Japanese government has encouraged this development. Ever since the Second World War, successive prime ministers have tried to get more women into the workplace and have even gone so far as to tell men to &quot;say goodbye to their bonuses if they refuse to let their wives and daughters work outside the home&quot;.\n\nBut while some Japanese women are finding their way into senior positions, many are also facing a harsh reality: most large Japanese corporations have an unofficial &quot;marriage bar&quot; that prohibits the promotion of women who are married and have children. This is particularly true of men. According to a study by the Japanese government, 61.5 per cent of male managers are married. By contrast, only 42.2 per cent of female managers are married. This means that women with children can be in a very difficult position, because \u2013 as Shihoko Goto of the Wilson Center points out \u2013 they face a choice between marriage and career. And that is not a choice that men face.\n\n&quot;If you look at people in their 30s,&quot; says Seiko, &quot;I think many of them are not getting married and having children. If you don't have children, you have more time and money to enjoy your life. So why not?\n\n&quot;I don't think that women are rejecting marriage and family,&quot; she says. &quot;It's just that if you want children and you are a career woman, it's impossible. There is no way. You can't do both. And so you have to make a choice. Either you go for your career or you go for marriage and family. That's what I think is happening.&quot;\n\n&quot;In Japan,&quot; says Goto, &quot;it's very difficult for men and women to be able to balance work and family. So, if you are a career woman, you will have to sacrifice family. That's the problem. That's why so many women don't have children. Because they want to focus on their career.&quot;\n\nThis is a view echoed by Ai Kato, a 29-year-old tax accountant. &quot;It's true that if you work in a large Japanese company, you cannot get married or have children if you want to keep your job,&quot; she says. &quot;That's the reality. It's a matter of the company culture. It's difficult to balance both. And so people are having fewer children.&quot;\n\nJapan has never really known anything else. &quot;Since the Meiji era [from 1868 to 1912] Japan's population has been very stable,&quot; says Kanazawa. &quot;So they think that is normal. But it's not. If you look at Germany, their population is falling too, but they are getting immigrants. And in the US and the UK, the population is stable, but people are having more children. So there are many different types of demographic change, and Japan is one of the most extreme.&quot;\n\nWhat is less clear is why this is happening. The popular explanation is that Japan is a victim of the phenomenon known as the &quot;demographic time bomb&quot; or &quot;]" time="4.968"><properties><property name="score" value="0.9821796351008945" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Australian Computer Emergency Response Team\n\nThe Australian Computer Emergency Response Team (AusCERT) is the recognised national computer security incident response team of the Australian Government. AusCERT works closely with the Australian Communications and Media Authority (ACMA), the Australian Federal Police (AFP), the Department of Defence (Defence), the Australian Taxation Office (ATO), and other government departments and agencies to deal with computer security incidents on an operational level. The organisation also provides specialist security incident response advice and services to the private sector.\n\nAusCERT was established in 1987 under the auspices of the Australian Computer Society (ACS) and is funded by the Australian Government through the Department of Communications and the Australian Government Information Management Office (AGIMO).\n\nAusCERT provides services to the public and private sectors in the following areas:\n\nThe headquarters for AusCERT are in the inner-city Melbourne suburb of Fitzroy, and are situated at the ACS Victoria. Other AusCERT team members are located around the country, including in Sydney, Brisbane, Adelaide and Perth.\n\nIn the early 1980s, the Australian federal government recognised the need for a computer emergency response capability and developed the Computer Emergency Response Team (CERT) concept. The first CERT team was established in December 1987 as the Computer Incident Advisory Capability (CIAC) under the auspices of the Australian Computer Society (ACS) at the University of Melbourne. In the mid-1990s, CIAC changed its name to AusCERT and moved its headquarters to the Department of Communications in Canberra. Since then, it has moved to the AGIMO building in Belconnen, ACT. In January 2013, AusCERT relocated to the City of Melbourne.\n\nAusCERT is a specialist organisation that provides advice and assistance in computer security incidents to its constituent members. As an operational unit, AusCERT:\n\n\nAusCERT's incident management process incorporates the Australian Signals Directorate's Incident Handling Working Model, an approved Australian Standard, AS 4801-2005.\n\nAusCERT is the designated lead agency for computer security incidents within Australia's Critical Infrastructure community, as defined by the Communications and the Arts portfolio of the Commonwealth Government.\n\nSince 2001, AusCERT has been responsible for issuing national warnings and alerts on a number of security issues, as directed by the Australian Communications and Media Authority (ACMA).\n\n\nAusCERT is active in a number of national and international computer security forums. AusCERT regularly participates in the Cyber Storm exercises conducted by the US Department of Homeland Security, as well as the Cyber Storm exercises hosted by the Singapore Infocomm Development Authority. AusCERT has also participated in the Cyber Security Exercise Series (CSES) developed by the Government of New Zealand. In 2003, AusCERT attended the UK's National Infrastructure Security Coordination Centre (NISCC) / British Computer Emergency Response Team (BCERT) course, and was the first international agency to receive the NISCC's Advanced Incident Handling course.\n\nOn 8 September 2005, AusCERT signed a Memorandum of Understanding with the US Department of Homeland Security's United States Computer Emergency Readiness Team (US-CERT) to formalise cooperation in the areas of:\n\n\n]" time="1.554"><properties><property name="score" value="0.8832996785640717" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Dick Carr\n\nDick Carr (born March 23, 1941) is a former American football player and coach. He served as the head football coach at Washington &amp; Jefferson College from 1978 to 1980 and at Allegheny College from 1990 to 1999, compiling a career college football record of 55\u201383\u20132.\n]" time="0.218"><properties><property name="score" value="0.9197072982788086" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.9197073&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.9197073
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Jandiala Guru railway station\n\nJandiala Guru railway station is a railway station in Sangrur district, Punjab. Its code is JG. It serves Jandiala Guru city. The station consists of 3 platforms.\n\nThe following trains stops at Jandiala Guru station :\n\n]" time="0.198"><properties><property name="score" value="0.6137322783470154" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.61373228&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.61373228
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Balantiocheilos melanopterus\n\nBalantiocheilos melanopterus is a species of fish in the family Cyprinidae.\n\nIn the wild they are found in southern Borneo, the lower portions of the Mekong River basin, and in the Chao Phraya and Mae Klong River basins. They are usually found in large, slow-flowing, and often turbid rivers. They live near the bottom, sometimes swimming over sandy areas, with many aquatic plants. They are often found in medium-sized shoals. They are mostly omnivorous. They are one of the smallest cyprinids. They have a maximum length of .\n\nThey are well-known for their remarkable abilities to tolerate very low water temperatures. They are known to survive long periods at temperatures as low as . However, when exposed to high water temperatures of , their metabolism starts to fail, which eventually leads to death. For this reason they are not recommended for tropical aquariums, but can make an excellent and hardy aquarium fish for more temperate regions.\n\n]" time="0.474"><properties><property name="score" value="0.9998251795768738" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Ahmed Harkat\n\nAhmed Harkat (, ; born April 27, 1971 in Ottawa, Ontario, Canada) is a Pakistani Canadian and Muslim who was detained as a threat to national security and imprisoned for two years in Canada.\n\nOn May 1, 2005, a Canada-wide arrest warrant was issued for Harkat by the Federal Bureau of Investigation (FBI) of the United States. On July 18, 2005, Ahmed Harkat was arrested under a Security Certificate by the Government of Canada, at his home in Ottawa, on the grounds that he was involved in terrorist activities, in particular in association with the al-Qaeda organization. On November 23, 2005, Ahmed Harkat was detained indefinitely by the Federal Court. He was released on strict house arrest conditions on February 1, 2006, after the Supreme Court of Canada ruled that his detention was unconstitutional. On December 13, 2007, the Court of Appeal for Ontario found that Harkat was detained lawfully. However, the Court ordered a new detention review hearing to be held before the end of June 2008. The Supreme Court of Canada refused to hear his appeal on February 15, 2008.\n\nHarkat applied for a judicial review of the decision of the Minister of Public Safety to order his detention. The Federal Court of Canada dismissed his application on December 14, 2006. Harkat has appealed to the Supreme Court of Canada.\n\nThe Federal Court decision came a week after the United States Senate rejected a bill to close the Guantanamo Bay detention camp.\n\nIn August 2008, Harkat was living under 24-hour house arrest, with some restrictions.\n\nOn June 17, 2009, the Supreme Court of Canada ruled that the security certificate under which Harkat was detained was valid and that he must be deported.\n\nOn July 29, 2015, the Federal Court of Appeal upheld a lower court decision and ruled that Harkat could stay in Canada. The ruling meant that Harkat could not be removed from Canada and remained under a strict form of house arrest.\n\nOn January 12, 2017, the Supreme Court of Canada refused to hear his appeal.\n\nAhmed Harkat was born in Pakistan in 1971, and he came to Canada in 1995. He speaks some English and has a wife and three daughters. He is a pizza delivery driver in Ottawa, Ontario, Canada.\n\nHarkat worked in Peshawar, Pakistan as a relief coordinator for Care International Pakistan, an international humanitarian organization, until 1999. He then went on to work as an accountant and cashier at Afghan Relief in Peshawar, a project of the National Council for Voluntary Social Services (NCVSS).\n\nHarkat came to the attention of Canadian intelligence officials after being deported from Dubai in April 2000, after he failed to obtain a visa for the United States. He has never visited Afghanistan, a key requirement of a security certificate. Harkat also states that he was not involved in any terrorist activity, and has never supported or associated with a terrorist organization.\n\nHarkat was arrested on December 10, 2002. He was charged with knowingly facilitating a terrorist activity and knowingly participating in the activities of a terrorist group.\n\nOn January 23, 2003, a judge granted the Royal Canadian Mounted Police (RCMP) an order to monitor his activities and to search his home. On March 20, 2003, Harkat was released under a strict curfew. The surveillance and house arrest was lifted on September 11, 2003.\n\nOn February 10, 2004, the judge in the case ruled that Harkat would not be allowed to work as a relief coordinator in his community or to interact with a list of specified persons, including the media. On March 10, 2004, Harkat's curfew was changed from 11 p.m. to 6 a.m., and the list of people he was not allowed to contact was expanded to include more than 100 people, including several Members of Parliament.\n\nOn May 11, 2004, the judge in the case ruled that Harkat would not be allowed to attend community events, and that he would be required to meet regularly with RCMP officials.\n\nOn August 16, 2004, Harkat was accused of writing love letters to two married women, one of whom was an RCMP officer.\n\nOn September 15, 2004, the Attorney General of Canada, under Paul Martin, asked that Harkat be detained for an additional six months, which was approved by a judge on October 12.\n\nOn September 30, 2004, Harkat's wife, Sophie, appeared on a news show, the &quot;National Magazine&quot; on CBC.\n\nOn January 21, 2005, the Federal Court of Canada ordered Harkat to remain under a 24-hour surveillance and under house arrest. On January 23, 2005, he was required to surrender his passport.\n\nOn May 1, 2005, a Canada-wide arrest warrant was issued by the FBI for Hark]" time="2.453"><properties><property name="score" value="0.9997739553451538" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[By DontHassleDaHoff\n\nWe're heating things up with PROJECT: Overcharge in the upcoming Patch 7.24b!\n\nPROJECT: Overcharge\n\nAssault up, minions! Project: Overcharge replaces Stormrazor and Stormguard Banner, empowering your entire team with 10% increased attack speed and 8% increased movement speed for 15 seconds. In addition, you'll gain Overcharge stacks while attacking enemy champions and playing as Summoner's Rift objectives. The more you participate, the faster you charge!\n\nWith the PROJECT pre-season approaching, we're giving Overcharge a test run on the PBE!\n\nHow do I get Overcharge?\n\nYour team starts with a fixed number of Overcharge tokens that reset when you recall. You earn more Overcharge by killing enemy champions and playing as objectives.\n\nObjective Total tokens Rift Herald 2 Scuttle Crab 2.5 Dragon 4\n\nKilling an enemy champion generates 1 stack, but if you earn more than 5 stacks, you'll go into &quot;overdrive&quot; and get a short burst of additional movement speed and attack speed!\n\n1-5 stacks: +8% AS / 10% MS\n\n6-10 stacks: +16% AS / 20% MS\n\n11+ stacks: +24% AS / 30% MS\n\nOverdrive lasts 1.5 seconds and grants up to +120% bonus AS and +180% bonus MS!\n\nOverdrive ends immediately if you're dealt damage greater than 4% of your maximum health or if you activate Stormguard Banner.\n\nPROJECT: Hunters\n\nPROJECT: Hunters rewards map objectives and creating map pressure for your team. If you slay a jungle monster or help your allies take down an enemy Rift Herald, Dragon, or Baron, you'll earn a stack of PROJECT: Hunters.\n\n1 stack for Rift Herald\n\n2 stacks for Dragon\n\n4 stacks for Baron\n\n3 stacks for a secured Baron kill\n\nWhen you earn a stack of PROJECT: Hunters, you and your teammates are empowered for 15 seconds with +10% bonus movement speed and +10% bonus attack speed.\n\nPlay the objective and rack up those stacks to hunt for the enemy team!\n\nExhaust changes\n\nWe're making some changes to Exhaust that are designed to let it work better with aggressive crowd control effects.\n\nWe want to remove Exhaust's bonus AD ratio because it makes it really difficult to fight champions with both Exhaust and high attack speed. However, Exhaust should feel more effective against high attack speed targets because its slow should be more powerful.\n\nWe're doing this by reducing the duration of the slow by 10% but increasing the duration of the debuff that prevents the target from autoattacking by the same amount. Overall, this means that it'll be easier to reach the minimum Attack Speed slow but harder to reach the maximum Attack Speed slow.\n\nAlso, this means Exhaust won't affect attack speed-based champs as much because it'll take more time for the debuff to tick down.\n\nWe'll continue to assess Exhaust and other individual adjustments to hit our goals in the pre-season.]" time="1.586"><properties><property name="score" value="0.3338683644930522" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.33386836&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.33386836
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Unemployed people who refuse a job offer will have their benefits cut under Government plans.\n\nThe Government will tighten the conditions which allow unemployed people to refuse work offers.\n\nWork and Income (WINZ) has been told it must provide greater support for people to get into work.\n\nBeneficiaries who refuse a suitable job offer will have their benefits reduced under new measures announced by Social Development Minister Paula Bennett.\n\nREAD MORE:\n\n* Government plans crackdown on parents who won't work\n\n* PM takes aim at beneficiaries who refuse to work\n\n* Thousands of solo parents' benefit slashed under 'tough love' plan\n\n* More sanctions on beneficiaries as Government claims welfare reforms are working\n\n* Benefit sanctions have forced hundreds to quit job search\n\n* Beneficiaries told 'get a job' in get tough letter\n\n* Benefit sanctions 'designed to make you feel like you're doing something wrong'\n\n* No Job? No Benefit.\n\nThere will be exceptions for those who cannot work due to illness or disability, and those in serious financial hardship.\n\nThose who had been receiving a sickness or invalids benefit, such as the invalids benefit, the sickness benefit, or the severe disabilities benefit for at least six months in the past two years, will be exempt from the tougher requirements.\n\nHowever, all job seekers who had been on a sickness benefit for six months or more would be required to actively look for a job, even if they were on a sickness benefit.\n\nThe measures, contained in a Ministry of Social Development discussion paper, would apply to new and existing beneficiaries.\n\nThey would also apply to beneficiaries who are looking after a child under six months old, or a child who is in the &quot;prescribed age range&quot; of six to 12 months, or is aged under two years old.\n\nPeople with children over two years old would be expected to look for work once the youngest child turned two.\n\nBennett said jobseekers needed to be active and look for work, rather than wait to be told what to do.\n\n&quot;We don't think the system is working very well at the moment. The policy is asking for more people to go into paid employment.&quot;\n\nThe Government is under pressure to get more people into work, particularly with a labour shortage.\n\nJob vacancies were at a 10-year high, but there were not enough skilled workers to fill them, Work and Income chief executive Glenys Coughlan said.\n\nThere were also not enough job offers to meet the needs of unemployed people, Bennett said.\n\n&quot;We know there are around 28,000 job seekers who have been unemployed for a year or more, and nearly 30,000 unemployed who have been unemployed for at least two years,&quot; Bennett said.\n\n&quot;We want to get these people into the workforce.&quot;\n\nThe Government's target is for 55 per cent of people aged between 18 and 64 to be in paid work by the end of this year.\n\n&quot;This is a national problem. It's going to require a lot of hard work from all of us to solve it.&quot;\n\nThe discussion document is out for public consultation, and the measures are likely to be introduced in September.\n\nIf the measures are adopted, there will be increased reporting requirements, and consequences for not meeting work obligations.\n\nThere will be four levels of compliance measures.\n\n&quot;Where a person is placed at the highest level of compliance with their obligations, it would mean that a person may be subject to reduced or no benefit payments if they are not meeting their obligations,&quot; the discussion paper said.\n\nPeople on a benefit would be expected to have regular interviews with a case manager, and have mandatory obligations to get training and development.\n\n&quot;The expectation is that the person will be actively looking for work, and not sitting on the benefit,&quot; Bennett said.\n\n&quot;They can say no to jobs, but they will still need to look.&quot;\n\nThose who refused a job offer three times would be placed on the work preparation programme. They would be expected to actively look for work, and attend job interviews, if they were suitable for the position.\n\nIf the person did not have a suitable job offer, they would be expected to participate in the work preparation programme.\n\nThose on a benefit who refused to participate in the programme would have their benefit reduced.\n\nWork and Income would also actively seek out job opportunities, and encourage beneficiaries to apply for the jobs.\n\nBennett said the work preparation programme would not be mandatory.\n\nBut &quot;if you have two years of not being in work, you should probably have a little bit of pressure on you to do a little bit more&quot;.\n\n&quot;I wouldn't want to see a situation where there was lots of sanctions. I think there will be exceptions.&quot;\n\nThe work programme already had a requirement for those who were work-ready to actively seek out work.\n\nThe changes were &quot;very targeted&quot;, she said.\n\n&quot;If you've been on a benefit for a long time, we want to make sure you are taking steps to get into work.\n\n&quot;It's about saying to people 'get with it or we'll move you along'.&quot;\n\nThe tighter requirements for beneficiaries were part of a wider review of the welfare system.\n\nThe changes could see beneficiaries' obligations reduced if they were taking care of a seriously ill or dying family member.\n\nBut if the sickness was self-inflicted, such as drug and alcohol abuse, there could be tougher requirements.\n\nCoughlan said there was already a work-ready requirement for job seekers, but that it was flexible.\n\n&quot;I think we're not as prescriptive as Australia. I think they're much more prescriptive.&quot;\n\nShe said it was important to keep the labour market up to date.\n\n&quot;That's part of our focus, to make sure that we've got the market that matches up with the right skills.&quot;\n\nThere had already been a move towards a more results-based approach, she said.\n\n&quot;We're already in the business of 'what are the outcomes for the people we serve'?&quot;\n\nThe discussion document said the proposed measures would help reduce the number of beneficiaries, while supporting those who were genuinely unable to work.\n\n&quot;It is proposed that these policy changes will provide a tighter fit between the requirements for beneficiaries to engage in paid work, and the exceptions to those requirements,&quot; it said.\n\nThe new measures are part of the Government's welfare reforms.\n\n&quot;There is a real concern that people are falling out of the benefit system. We]" time="3.271"><properties><property name="score" value="0.753700145653316" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Filth was first made into a play for the BBC\u2019s series of television plays in 2003, featuring Ewan McGregor, as a detective called Bruce Robertson. I don\u2019t think it\u2019s too far fetched to say that Bruce is the Scottish version of Neil Buchanan in American Psycho. Not necessarily in terms of the content of the story, but in the sheer fact that he is a very messed up guy, one who people can\u2019t stand, but at the same time can\u2019t stop looking at him.\n\nThis character is played to perfection by Jonny Lee Miller in the new film adaptation of Filth. The novel follows Bruce Robertson (Miller) a detective who, after starting the story off as a drunken, bitter asshole, finds out that he has cancer and so takes leave to get himself checked out. This action in turn leads to Bruce discovering that he is actually a massive pervert. He becomes obsessed with performing the most unspeakable of actions in order to be like the perverts that he so despises. The film follows Bruce as he conducts an investigation to find the person responsible for him being a terrible person and while he is still struggling with his illness.\n\nWhat is great about this adaptation is that it keeps the nastiness and the gritty, unpleasant content that is in the book. That might be what stops this film from being a complete success because this film is a difficult watch. It\u2019s not so difficult because of the content but because you will be so incredibly frustrated by the main character. He is an unpleasant person who you don\u2019t want to spend time with.\n\nIt\u2019s a testament to Miller\u2019s performance that he makes you want to keep watching even though you can\u2019t stand him. It\u2019s this nastiness that he exudes that makes him so engaging to watch. You will want to know how his investigation turns out and what his next move is, because he is truly awful. It also helps that he looks incredibly creepy, like he would kill someone at the slightest provocation.\n\nBruce is helped by a fantastic supporting cast. Jon Skellern plays Bruce\u2019s best friend Simon. You can tell that Simon has his concerns for Bruce and that his concern for him will only continue to grow as Bruce\u2019s behaviour becomes worse. Simon acts as the voice of reason when it comes to Bruce and does his best to help him, only to be shut down by Bruce\u2019s barbs and by Bruce himself. Kelly Macdonald plays Bruce\u2019s wife Susan. While Bruce gets into scrapes with his fellow detectives, Susan is struggling to cope with him and his illness. It\u2019s obvious that Susan is struggling with her husband\u2019s condition, and you will feel for her throughout the film.\n\nThere are a lot of big names in this film that will be recognisable to most people. Jonny Lee Miller, so good as Sherlock Holmes, plays Bruce with a brilliant blend of ferocity and uncertainty. While he is fully committed to the disgusting nature of his character, he also seems like he could completely lose it at any minute. Stephen Fry is outstanding as Inspector Siobhan Clarke, who is one of the few people who is able to make Bruce think twice about the things he does. Everyone is perfectly cast and help to make this film feel incredibly authentic.\n\nThe film does do a good job of adapting the book. Although the source material is sometimes over the top, the film takes that feeling and amplifies it to the next level. What is great about this film is that you do believe that all the characters are real people. You get a real sense of the community that Bruce has around him and that is what helps to make the film so engaging. You want to know what is going to happen to Bruce and how he is going to get out of the situations that he has put himself into.\n\nFilth is a good adaptation of the novel by Irvine Welsh. It does have the same nastiness that is in the novel, which is great for those of you who enjoy watching people doing things that you shouldn\u2019t be watching. The acting is excellent, with a standout performance from Jonny Lee Miller, who takes the idea of a horrible person to a whole new level. It\u2019s not the most pleasant of films to watch, but it\u2019s certainly one of the most engaging.]" time="2.087"><properties><property name="score" value="0.9997565001249313" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[This board was created for the discussions of Nintendo's two major flagship gaming systems, the Nintendo DS and the Nintendo Wii. Some discussion about previous Nintendo systems is also welcome, but as always please try to keep the conversation in the current console's area.\n\n\n\nNext Display topics from previous: All Topics 1 day 7 days 2 weeks 1 month 3 months 6 months 1 year Sort by Author Post time Replies Subject Views Ascending Descending\n\nReturn to Board index\n\nJump to: Select a forum ------------------ Forums Info &amp; Announcements Read me first! Hints &amp; Tips Gaming Marketplace Recommended Source Books Role Playing - Tales of The Mind Forums Games &amp; RPG Rules The Sinister Spire Promethean: The Created The Dreaming Stone Vampire: The Masquerade Werewolf: The Apocalypse Exalted: The Emerald Dawn Wraith: The Oblivion Promethean: The Created Wraith: The Oblivion Promethean: The Created 3rd Edition Promethean: The Created 2nd Edition Werewolf: The Apocalypse Vampire: The Masquerade Werewolf: The Apocalypse Exalted: The Emerald Dawn Wraith: The Oblivion Wraith: The Oblivion Promethean: The Created Wraith: The Oblivion Vampire: The Masquerade Wraith: The Oblivion Vampire: The Masquerade Werewolf: The Apocalypse Vampire: The Masquerade Exalted: The Emerald Dawn Werewolf: The Apocalypse Exalted: The Emerald Dawn Vampire: The Masquerade Exalted: The Emerald Dawn Wraith: The Oblivion Wraith: The Oblivion Promethean: The Created Vampire: The Masquerade Werewolf: The Apocalypse Exalted: The Emerald Dawn Vampire: The Masquerade Promethean: The Created Werewolf: The Apocalypse Promethean: The Created 2nd Edition Werewolf: The Apocalypse Exalted: The Emerald Dawn Wraith: The Oblivion Promethean: The Created 3rd Edition Werewolf: The Apocalypse Promethean: The Created Werewolf: The Apocalypse Exalted: The Emerald Dawn Vampire: The Masquerade Promethean: The Created 2nd Edition Vampire: The Masquerade Promethean: The Created 3rd Edition Vampire: The Masquerade Exalted: The Emerald Dawn Werewolf: The Apocalypse Exalted: The Emerald Dawn Werewolf: The Apocalypse Exalted: The Emerald Dawn Wraith: The Oblivion Wraith: The Oblivion Wraith: The Oblivion Vampire: The Masquerade Vampire: The Masquerade Wraith: The Oblivion Wraith: The Oblivion Promethean: The Created Vampire: The Masquerade Promethean: The Created Werewolf: The Apocalypse Werewolf: The Apocalypse Promethean: The Created Exalted: The Emerald Dawn Wraith: The Oblivion Vampire: The Masquerade Promethean: The Created 3rd Edition Vampire: The Masquerade Vampire: The Masquerade Exalted: The Emerald Dawn Promethean: The Created 2nd Edition Werewolf: The Apocalypse Promethean: The Created Werewolf: The Apocalypse Exalted: The Emerald Dawn Vampire: The Masquerade Vampire: The Masquerade Exalted: The Emerald Dawn Promethean: The Created Wraith: The Oblivion Wraith: The Oblivion Vampire: The Masquerade Promethean: The Created Exalted: The Emerald Dawn Werewolf: The Apocalypse Vampire: The Masquerade Vampire: The Masquerade Exalted: The Emerald Dawn Promethean: The Created 2nd Edition Werewolf: The Apocalypse Vampire: The Masquerade Wraith: The Oblivion Werewolf: The Apocalypse Vampire: The Masquerade Wraith: The Oblivion Promethean: The Created Exalted: The Emerald Dawn Werewolf: The Apocalypse Vampire: The Masquerade Promethean: The Created 3rd Edition Wraith: The Oblivion Werewolf: The Apocalypse Vampire: The Masquerade Promethean: The Created Wraith: The Oblivion Vampire: The Masquerade Exalted: The Emerald Dawn Vampire: The Masquerade Exalted: The Emerald Dawn Wraith: The Oblivion Wraith: The Oblivion Wraith: The Oblivion Vampire: The Masquerade Vampire: The Masquerade Exalted: The Emerald Dawn Werewolf: The Apocalypse Vampire: The Masquerade Wraith: The Oblivion Wraith: The Oblivion Vampire: The Masquerade Promethean: The Created Exalted: The Emerald Dawn Werewolf: The Apocalypse Wraith: The Oblivion Promethean: The Created Werewolf: The Apocalypse Vampire: The Masquerade Vampire: The Masquerade Promethean: The Created 3rd Edition Werewolf: The Apocalypse Vampire: The Masquerade Exalted: The Emerald Dawn Wraith: The Oblivion Wraith: The Oblivion Vampire: The Masquerade Promethean: The Created 2nd Edition Vampire: The Masquerade Promethean: The Created Wraith: The Oblivion Wraith: The Oblivion Vampire: The Masquerade Vampire: The Masquerade Exalted: The Emerald Dawn Wraith: The Oblivion Wraith: The Oblivion Vampire: The Masquerade Exalted: The Emerald Dawn Wraith: The Oblivion Wraith: The Oblivion Promethean: The Created Wraith: The Oblivion Vampire: The Masquerade Wraith: The Oblivion Wraith: The Oblivion Wraith: The Oblivion Vampire: The Masquerade Vampire: The Masquerade Wraith: The Oblivion Vampire: The Masquerade Vampire: The Masquerade Vampire: The Masquerade Exalted: The Emerald Dawn Werewolf: The Apocalypse Werewolf: The Apocalypse Vampire: The Masquerade Exalted: The Emerald Dawn Promethean: The Created 2nd Edition Vampire: The Masquerade Promethean: The Created Vampire: The Masquerade Exalted: The Emerald Dawn Wraith: The Oblivion Wraith: The Oblivion Vampire: The Masquerade Promethean: The Created 3rd Edition Werewolf: The Apocalypse Vampire: The Masquerade Wraith: The Oblivion Wraith: The Oblivion Vampire: The Masquerade Wraith: The Oblivion Wraith: The Oblivion Vampire: The Masquerade Vampire: The Masquerade Vampire: The Masquerade Vampire: The Masquerade Wraith: The Oblivion Wraith: The Oblivion Wraith: The Oblivion Wraith: The Oblivion Vampire: The Masquerade Vampire: The Masquerade Vampire: The Masquerade Wraith: The Oblivion Wraith: The Oblivion Wraith: The Oblivion Vampire: The Masquerade Vampire: The Masquerade Wraith: The Oblivion Wraith: The Oblivion Wraith: The Oblivion Wraith: The Oblivion Vampire]" time="2.846"><properties><property name="score" value="0.946131189664205" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Each year, the Duluth News Tribune names a Great Duluthian. This year's Great Duluthian is Bernie Betz, founder and president of Betz Equipment.\n\nBetz, an avid hunter and fisherman, is not only a successful business man, he is a giver. His Betz Family Foundation has donated over $5 million to various organizations and projects, including Duluth's Discovery Center and the Betz Sportsman's Club, which has operated since 1948.\n\n&quot;It has been a real honor to know Bernie, work with him, and to benefit from his support,&quot; says John Stowe, president and CEO of the Duluth Entertainment Convention Center (DECC).\n\nIn his acceptance speech, Betz shared a quote from Henry Ford: &quot;If you think you can do a thing, or if you think you can't do a thing, you're right.&quot; Betz says he tries to follow this philosophy in all aspects of his life.\n\nBetz said that there are too many good people in Duluth to name one &quot;Great Duluthian,&quot; but he humbly accepted the award. &quot;This is just a thank you,&quot; Betz said. &quot;I thank you for this.&quot;]" time="0.583"><properties><property name="score" value="0.9995373487472534" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Beauty\n\nI came across this article in Style Magazine about Kim Kardashian being named a rising star by Teen Vogue. Really?! Kim Kardashian is being compared to Audrey Hepburn, and Grace Kelly? I am not quite sure what Teen Vogue was thinking, but this makes me think they are more concerned about the bottom line of a magazine than really focusing on their readers and what they want to read about. What kind of beauty standards does Teen Vogue endorse?]" time="0.230"><properties><property name="score" value="0.9807391166687012" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[This is the first time I\u2019ve made this recipe and I absolutely loved it! It is now one of my favorites. It had a lot of flavor and the ginger and garlic added great zing. I used 1 tsp of ginger and 1/2 tsp of garlic. This was great served over brown rice. My husband raved about this meal. Thanks for a great recipe!\n\nI have made this for years but never reviewed it. The longer it sits the better it tastes. I do make it on a regular basis. I often do half coconut milk and half veggie broth. One of my family's favorite meals. I am thinking of using the coconut milk and broth when I cook it instead of water and bullion cube.]" time="0.353"><properties><property name="score" value="0.9997527003288269" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Le lancement de Canon EOS M5 est imminent et pour ce faire, la soci\xe9t\xe9 nippone lance sur les r\xe9seaux sociaux une campagne marketing ax\xe9e autour de la qualit\xe9 des vid\xe9os. Un teaser se prom\xe8ne donc sur la toile d\u2019apr\xe8s laquelle le bo\xeetier serait capable de filmer dans une r\xe9solution 4K (passe de 2,7K).\n\nLes qualit\xe9s vid\xe9o d\u2019un bo\xeetier d\xe9termin\xe9es par la r\xe9solution de l\u2019enregistrement de la vid\xe9o. La baisse de la r\xe9solution ne signifie pas pour autant que la qualit\xe9 diminue : on peut avoir une r\xe9solution 4K pour une vid\xe9o en HD, ou inversement.\n\nEt de toute \xe9vidence, la r\xe9solution 4K de la vid\xe9o de Canon EOS M5 semble plus qu\u2019une simple rumeur. Et c\u2019est un scoop ? Pas vraiment puisque le constructeur avait annonc\xe9 que son prochain compact avait la capacit\xe9 de filmer en 4K il y a un an. Et pour une raison qui n\u2019est pas claire, la marque nippone a tout simplement gard\xe9 ce secret.\n\nUne r\xe9solution 4K a de quoi attirer les foules, et surtout les utilisateurs d\u2019appareils photo. En effet, la 4K constitue aujourd\u2019hui un format de r\xe9f\xe9rence, \xe0 l\u2019instar de la r\xe9solution Full HD. Cela dit, la 4K n\u2019est pas forc\xe9ment un avantage pour un bo\xeetier de ce type, mais plut\xf4t pour les bo\xeetiers d\u2019entr\xe9e de gamme ou de milieu de gamme.\n\nCanon EOS M5 : plusieurs versions ?\n\nLa r\xe9solution 4K, comme l\u2019explique Dpreview, est un atout dans le cas des fabricants de bo\xeetiers car ceux-ci peuvent d\xe8s lors proposer des machines \xe0 un prix bas et vendre ensuite des objectifs (propres \xe0 l\u2019appareil ou non) \xe0 prix plus \xe9lev\xe9.\n\nCanon EOS M5 a beau \xeatre attendu au tournant, il ne faudra toutefois pas s\u2019attendre \xe0 ce que l\u2019appareil soit dot\xe9 d\u2019une stabilisation des images. Le bo\xeetier ne semble en effet pas disposer d\u2019une stabilisation optique d\u2019image, alors que la concurrence (compacts compris) poss\xe8de d\xe9sormais un syst\xe8me de stabilisation optique.]" time="1.618"><properties><property name="score" value="0.9997052550315857" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99970526&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99970526
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[// Use of this source code is governed by a BSD-style license that can be\n\n// Check if the field is initialized or not\n\n// Determine the kind of path to use\n\n// Generate the path from the module name, directory and field name.\n\n// Module name comes first, to keep a consistant ordering between native\n\n// native method code and java/native method code, to allow for more readable\n\n// Remove 'this' parameter from the function declaration\n\n// If the arguments are references (ie. the passed in argument is a local\n\n// variable), mark the parameter as read only.\n\n// If there is a non-factory constructor, use it\n\n// Always use the default constructor\n\n// A non-null `type` value, or `value` with a type annotation,\n\n// means that we are constructing a type.\n\n// Use the native constructor if it exists, otherwise fall back to\n\n// Use the name of the class as the default constructor name.\n\n// Check if a constructor exists on the type\n\n// If the value is a reference, use the name of the field.\n\n// Use the default constructor name for a value.\n\n// If the value is a reference, use the name of the field.\n\n// If a type is specified (name or value), it takes precedence.\n\n// If we have a type annotation on the parameter, it must match the type\n\n// If we have a name annotation on the parameter, it must match the name\n\n// Use the name of the class as the constructor name\n\n// If we have a type annotation on the parameter, it must match the type\n\n// If we have a name annotation on the parameter, it must match the name\n\n// Don't check for annotation matches in constructors.\n\n// Try to use the field name as a constructor name\n\n// If there is a non-factory method, use it\n\n// Always use the default constructor\n\n// A non-null `type` value, or `value` with a type annotation,\n\n// means that we are constructing a type.\n\n// Use the native constructor if it exists, otherwise fall back to\n\n// Use the name of the class as the default constructor name.\n\n// Check if a constructor exists on the type\n\n// If the value is a reference, use the name of the field.\n\n// Use the default constructor name for a value.\n\n// If the value is a reference, use the name of the field.\n\n// If a type is specified (name or value), it takes precedence.\n\n// If we have a type annotation on the parameter, it must match the type\n\n// If we have a name annotation on the parameter, it must match the name\n\n// Use the name of the class as the constructor name\n\n// If we have a type annotation on the parameter, it must match the type\n\n// If we have a name annotation on the parameter, it must match the name\n\n// Don't check for annotation matches in constructors.\n\n// Try to use the field name as a constructor name\n\n// If there is a non-factory method, use it\n\n// Always use the default constructor\n\n// A non-null `type` value, or `value` with a type annotation,\n\n// means that we are constructing a type.\n\n// Use the native constructor if it exists, otherwise fall back to\n\n// Use the name of the class as the default constructor name.\n\n// Check if a constructor exists on the type\n\n// If the value is a reference, use the name of the field.\n\n// Use the default constructor name for a value.\n\n// If the value is a reference, use the name of the field.\n\n// If a type is specified (name or value), it takes precedence.\n\n// If we have a type annotation on the parameter, it must match the type\n\n// If we have a name annotation on the parameter, it must match the name\n\n// Use the name of the class as the constructor name\n\n// If we have a type annotation on the]" time="2.022"><properties><property name="score" value="0.999800980091095" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Why won't anyone on the Senate Banking Committee acknowledge that debt restructuring is needed for the Puerto Rico debt crisis to be solved? I just spent about two hours on CSPAN watching the Committee's hearings about Puerto Rico. Not once was the possibility of debt restructuring, which would mean debt reduction, mentioned. All that was mentioned was higher taxes, pension cuts, and deep government spending cuts, on top of what Puerto Rico is already doing. A complete deadlock. Puerto Rico has been austerity-impaired for a long time now, and has experienced a 12% drop in GDP as a result. No one can grow their way out of debt. Yet, even with this depression-like economy, Puerto Rico is still paying a huge amount of its income to its creditors. A lot of that debt comes from bonds that the government of Puerto Rico sold to investors to get the money to cover the huge losses that its electricity monopoly incurred because of oil price spikes. The government still has no way to fix that, since electricity is now being generated by private companies, due to a privatization law passed in 2013. The government of Puerto Rico could offer a debt exchange to investors, swapping the current bonds for ones with a longer maturity date and a lower interest rate, and then reinvest the proceeds into infrastructure projects, where it has a big deficit. But that possibility is completely ruled out. On the other side, the creditors are as intransigent as ever. They claim they can sue the US government if it tries to change their contracts, which is not true. They also claim that Puerto Rico is defaulting on its debt already, which is also not true. The government of Puerto Rico has been paying its creditors 100 cents on the dollar. There is simply no default risk. This is all a huge waste of time. All that the Puerto Rico hearings did was show that the US Congress is not capable of doing what it should have done long ago: allow Puerto Rico to restructure its debts. That would have given Puerto Rico a fresh start, and the creditors would have ended up with their money back. And we would have avoided all this useless talking. Now, we are just wasting time and making the situation worse. Puerto Rico's economy continues to decline. Puerto Rico's economy is expected to contract again by at least 1% in 2016, and another 2% in 2017. Most of the business owners I interviewed in Puerto Rico last month said they expect business to get worse in 2017. So the longer the creditors keep Puerto Rico in this situation, the worse it will get. In the hearings today, several senators brought up a resolution that the House of Representatives passed yesterday in a partisan vote, saying that Puerto Rico is already in default. That resolution is just to make a political point. The Supreme Court has already ruled that Puerto Rico cannot declare bankruptcy like a US municipality can, and cannot restructure its debt in a court. This was because of a law that was passed in 1984. The US Congress could have changed this law a long time ago, but they did not. The Republicans in the House of Representatives, however, do not want to restructure Puerto Rico's debts. Instead, they are trying to force the government of Puerto Rico to do all the reforms it already is doing, without giving it the opportunity to do the debt restructuring that will put its economy back on track. This was not really mentioned in the Senate hearings, but I found out from reporters that the Democrats on the Senate Banking Committee are not going to move forward with the House resolution. They understand that it is not going to pass. We need Congress to pass the resolution anyway, just to show that it has bipartisan support. In the House hearings, Representative Sean Duffy of Wisconsin said, &quot;This is not an emergency.&quot; This is a statement that no one in Puerto Rico agrees with. By delaying the debt restructuring, the economy of Puerto Rico is getting worse. More people are moving to the US mainland, which will further reduce the income that Puerto Rico is able to collect in taxes. More people are being forced to do cash jobs in the underground economy, because they cannot find other jobs. And those jobs do not get reported to the government. The longer this process is delayed, the worse it will get. But the creditors and some members of Congress do not seem to understand that Puerto Rico needs a fresh start. So they will continue to waste time, talking about things that are already being done, without any hope of any positive outcome. Puerto Rico's creditors just don't want to take a loss. But we are not doing Puerto Rico any favors by letting this situation continue for so long.]" time="2.348"><properties><property name="score" value="0.9987793087959289" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[A 21-year-old man died early Tuesday morning after being struck by a car while riding his bike in Logan Square, police said.\n\nAt 1:26 a.m., authorities responded to the intersection of Fullerton and Armitage avenues where the bicyclist was hit by a car, according to Chicago Police. He was taken to Advocate Illinois Masonic Medical Center, where he was later pronounced dead.\n\nThe man was identified as Fares Khaled, of the 900 block of North Francisco Avenue in Logan Square, according to the Cook County medical examiner's office.\n\nThe driver of the vehicle that struck the cyclist stayed at the scene and cooperated with police, authorities said. The investigation is ongoing.\n\nrsobol@tribune.com\n\nTwitter: @RosemarySobol1]" time="0.390"><properties><property name="score" value="0.9998273849487305" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[I love google chrome's ability to store passwords and usernames.\n\nBut it only works on &quot;http&quot; sites. I often use many other types of internet links, which would be useful to store credentials for.\n\nIs there a way to store and retrieve passwords on all types of web addresses?]" time="0.180"><properties><property name="score" value="0.8682397603988647" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[President Trump got into a heated exchange with CNN White House correspondent Jim Acosta at his press conference on Thursday, calling the reporter \u201cfake news\u201d and challenging his network\u2019s reporting.\n\nBut Trump\u2019s son, Eric Trump, says the president will always be fair and give all reporters the time of day.\n\nIn an appearance on Fox News\u2019 \u201cMedia Buzz,\u201d Eric said Trump may attack CNN but always makes time for the network\u2019s reporters.\n\n\u201c[T]hey can still come in. The reporter that he did call on [at the press conference], we always give them an opportunity at every press conference. He\u2019s always going to be fair,\u201d he said.\n\nADVERTISEMENT\n\n\u201cHe\u2019s always going to give them an opportunity, but we\u2019re not going to let them take potshots at our president and our administration,\u201d Eric Trump added.\n\nEric Trump also predicted that eventually the president will win the fight against \u201cfake news\u201d in the mainstream media.\n\n\u201cMy father is a fighter, and he\u2019s an unbelievable person. I mean, he\u2019s built an unbelievable business, and he\u2019s built an unbelievable life. He\u2019s a person that\u2019s been elected to lead this country, and he\u2019s going to do it with or without the media. He\u2019s going to do it with or without certain networks. And it\u2019s sad because he\u2019s such a good man, and he\u2019s such a good father,\u201d Eric Trump said.\n\n\u201cBut at the same time, if they\u2019re going to put themselves in a position where they\u2019re going to be speaking so much, they\u2019re going to continue to make mistakes,\u201d he added.]" time="1.375"><properties><property name="score" value="0.8800801038742065" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[This is not a bug or something that is wrong. The game is not detecting your Shield as something that it can block. The animation itself looks fine. I don't know how you can replicate it yourself to get a screenshot, but what you are seeing is the default Shieldblock animation without your Shield in your inventory.\n\n\n\nIf you are referring to the broken animation for Normal and Magic Shieldblock, the bug with that is fixed on 0.11.4. I'm not sure when we'll be able to patch that for Ascendancy. Last edited by Mark_GGG on Nov 11, 2015, 8:54:36 AM Posted by Mark_GGG\n\non Grinding Gear Games on Quote this Post\n\nI mean my shield is in my inventory and the character can block while using the skill.\n\n\n\nThat's what I mean about not being a bug.\n\n\n\nIf it's a game mechanic that can be prevented somehow that would be really helpful to know. Posted by Greendogo\n\non on Quote this Post\n\n&quot; Greendogo\n\n\n\nThat's what I mean about not being a bug.\n\n\n\nIf it's a game mechanic that can be prevented somehow that would be really helpful to know. I mean my shield is in my inventory and the character can block while using the skill.That's what I mean about not being a bug.If it's a game mechanic that can be prevented somehow that would be really helpful to know.\n\nThe game can't detect your shield. When you block, the shield is on your back. When you use your shield in the skill, it isn't. The game can't detect your shield. When you block, the shield is on your back. When you use your shield in the skill, it isn't. Posted by Mark_GGG\n\non Grinding Gear Games on Quote this Post\n\n&quot; Mark_GGG &quot; Greendogo\n\n\n\nThat's what I mean about not being a bug.\n\n\n\nIf it's a game mechanic that can be prevented somehow that would be really helpful to know. I mean my shield is in my inventory and the character can block while using the skill.That's what I mean about not being a bug.If it's a game mechanic that can be prevented somehow that would be really helpful to know.\n]" time="1.212"><properties><property name="score" value="0.9997941255569458" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Uninstall all Office add-ins in Office 365\n\nIn this article\n\nWhen you install an Office add-in, you will be able to see the add-in in the Add-ins page in Word, Excel, or PowerPoint. However, it is not the case that you can only install add-ins from this location. In fact, you can add any website to the browser toolbar, and install a specific add-in from there.\n\nThis article describes the steps to completely uninstall all add-ins from all Office applications that you have in your tenant. This will allow you to completely remove all Office add-ins from your Office 365 tenant. This can be useful in a tenant cleanup or a company migration.\n\nImportant Before following the instructions in this article, please verify that your Office 365 subscription has the add-in licensing model and that the Tenant admin has permissions to run PowerShell scripts.\n\nIf you have any questions or feedback, please let us know in the Q&amp;A section on this page.\n\nManual process\n\nUse the following steps to completely uninstall all Office add-ins from all Office applications that you have in your tenant.\n\nImportant Before following the instructions in this article, please verify that your Office 365 subscription has the add-in licensing model and that the Tenant admin has permissions to run PowerShell scripts.\n\nOpen the SharePoint Online Management Shell. Run the following command: Get-PnPUtilization | Remove-PnPUtilization To remove the add-ins from Word, Excel, and PowerPoint, run the following commands: Get-PnPUtilization | ? {$_.Product -like &quot;*Word*&quot;} | Remove-PnPUtilization Get-PnPUtilization | ? {$_.Product -like &quot;*Excel*&quot;} | Remove-PnPUtilization Get-PnPUtilization | ? {$_.Product -like &quot;*PowerPoint*&quot;} | Remove-PnPUtilization To remove the add-ins from Outlook, run the following commands: Get-PnPUtilization | ? {$_.Product -like &quot;*Outlook*&quot;} | Remove-PnPUtilization\n\nPowerShell script\n\nThe following PowerShell script will completely uninstall all Office add-ins from all Office applications that you have in your tenant.\n\nImportant Before following the instructions in this article, please verify that your Office 365 subscription has the add-in licensing model and that the Tenant admin has permissions to run PowerShell scripts.\n\n# run as SharePoint Administrator $spAdmin = [Microsoft.SharePoint.Administration.SPAdminService]::GetAdmin # Connect to the SharePoint Online Management Shell $webApp = Get-SPWebApplication &quot;https://&lt;your_tenant_name&gt;.sharepoint.com&quot; $w = $webApp.LoginPage $w.ReturnUrl = $w.Url.AbsoluteUri + &quot;/_layouts/15/ PowerShell -ApplicationPage.aspx&quot; $cred = Get-Credential $credential = $spAdmin.Credentials.GetNetworkCredential($cred.UserName, $cred.Password) $session = New-PSSession -ConfigurationName Microsoft.Exchange -ConnectionUri https://ps.outlook.com/powershell/ -Credential $cred -Authentication Basic -AllowRedirection Import-PSSession $session -CommandName &quot;Remove-PnPUtilization&quot; -AllowClobber $excel = Get-PnPUtilization | ? {$_.Product -like &quot;*Excel*&quot;} | Remove-PnPUtilization $word = Get-PnPUtilization | ? {$_.Product -like &quot;*Word*&quot;} | Remove-PnPUtilization $powerpoint = Get-PnPUtilization | ? {$_.Product -like &quot;*PowerPoint*&quot;} | Remove-PnPUtilization $outlook = Get-PnPUtilization | ? {$_.Product -like &quot;*Outlook*&quot;} | Remove-PnPUtilization Remove-PSSession $session\n\nNote In the PowerShell script, you can run the commands individually to remove a specific add-in for a specific Office application, or you can remove all Office add-ins by running the Remove-PnPUtilization cmdlet without any parameters.\n\nAdditional information\n\nLearn more about how to remove a specific add-in for a specific Office application\n\nLearn more about how to add or remove an Office add-in in Office 365\n\nLearn more about the SharePoint Online Management Shell]" time="2.332"><properties><property name="score" value="0.11100287735462189" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.11100288&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.11100288
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Shepard and his team have done it again. I've been a fan of &quot;American Chopper&quot; since day one. I used to get angry with Paul Sr. for the way he treated the guys but I guess that was part of the show's draw. I have never seen a father and son fight like that on TV before. I still hope he comes back someday to work with his son, and they can rebuild the relationship they had before. I can't say enough good things about this season of American Chopper, with all the good work the team did for cancer patients, this is one of the most memorable seasons of the series.\n\nThe creativity of the builds is incredible. I'm so glad that they were able to get the Ford back on the road after the fire. They did an amazing job on the bike itself and the amount of work they put into making sure everything was functional on the bike, even the brakes, was inspiring. The customization they did on the ambulance was pretty impressive as well, especially how they made sure all the hydraulic and electrical components worked correctly. I also loved the idea of doing a bike with a hunting theme and the amount of work they put into making the motorcycles look like they belonged to an entire hunting party was great. They did an amazing job on the bike that played homage to the military and even the Great American motorcycle show was awesome! The barber shop bike and &quot;Appalachian Outlaw&quot; motorcycle were just as awesome. The personal touches that were put on these bikes really made them so much better.\n\nI know a lot of people criticize the shows for not showing the whole building process, but in my opinion, I like to see the creativity that the builders put into the bikes. The shows are a lot more enjoyable when I see what goes on in their heads when they come up with their ideas.]" time="0.984"><properties><property name="score" value="0.9997727274894714" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[1060\xb0\n\nFacebook Shows Off Most Personal and In-Depth Profiles\n\nTHR: Facebook has always been about connecting with the people in your life, but the social network is taking things to the next level with new profiles that put the emphasis on the people you care about the most.]" time="0.151"><properties><property name="score" value="0.8068450093269348" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The drama of the week is all about the inability of a local physician to get paid by insurance companies. It is not an uncommon story in that many physicians cannot get paid by insurance companies for a variety of reasons. The difference in this story is that this physician not only tried to get paid, but took the insurance company to court to try to get paid.\n\nIt started when a physician sent an invoice to an insurance company for services rendered. The physician was told that the insurance company did not have the funds to pay the invoice and that they would pay the bill at a later date. The physician sent the insurance company a bill every month for about a year until the insurance company stopped communicating. The physician was then advised by a collection agency that if he wanted to be paid, he would have to take the insurance company to court.\n\nThe physician then did just that. He filed a claim with the District Court seeking the unpaid bill plus interest and costs. The insurance company tried to have the case dismissed by claiming the physician did not have standing to sue. The insurance company argued that the physician only had a contractual right to be paid, but that right had not been breached. They also argued that he could not have been damaged by their failure to pay since it was their money.\n\nThe physician won the case. He was awarded the $10,600 plus interest and costs. The judge agreed that the insurance company breached the contract when it failed to pay the bill. The judge also found that the insurance company's withholding of payment did cause the physician to suffer damages in the form of legal fees to pursue payment.\n\nAs for the insurance company, it is probably a good idea that they got caught. Since their contract was breached and they caused damages, they may be liable for punitive damages which could double the amount of the doctor's claim.]" time="0.826"><properties><property name="score" value="0.06157013773918152" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Forgive me if this has been answered already.\n\n\n\nI bought a standard prepaid plan and activated it last week. I purchased the standard plan with 1 GB of data and 1 GB of minutes and unlimited texts.\n\n\n\nToday, I purchased a pre-paid plan with 1 GB of data and 3 GB of minutes and unlimited texts. When I added the new plan to my phone, it overwrote my previous plan with only 1 GB of data.\n\n\n\nWhen I called Koodo, I was told that I could not keep both plans.\n\n\n\nIs this correct? Can I not use 2 different pre-paid plans? I don't want to lose the 1 GB of data in my phone because it's a limited offer.\n\n\n\nThanks]" time="0.406"><properties><property name="score" value="0.99892657995224" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Land of the Sun\n\n10.03.17 / CD I Love You Like a Brother\n\n08.11.16 / CD I Love You Like A Brother\n\n08.11.16 / LP (analog) Grand Romantic\n\n08.11.16 / CD We Didn't Do It For Your Love\n\n03.08.14 / CD Why Did You Make Me Care?\n\n04.10.13 / CD Some Hearts\n\n02.12.12 / CD]" time="0.255"><properties><property name="score" value="0.8344732522964478" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.83447325&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.83447325
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Google+ on pakotettu entist\xe4 tiukempaan sis\xe4ll\xf6ntuotantoon, kertoo TechCrunch. My\xf6s yhti\xf6n taloudelliset resurssit k\xe4rsiv\xe4t j\xe4\xe4dytyksist\xe4.\n\nLue my\xf6s Google jatkaa keskittymist\xe4 Androidiin\n\nYhti\xf6 aikoo t\xe4sment\xe4\xe4, ett\xe4 j\xe4rjestelm\xe4 ei ole en\xe4\xe4 pelkk\xe4 verkkopalvelu, vaan tiettyyn sis\xe4ll\xf6ntuotantoa vaativaan verkkopakettiin. Verkkopalvelua varten tarvittavat julkaisuprofiilit, lis\xe4\xe4ntyneet palvelut ja Google Play -sovelluskaupan julkaisut ovat poistumassa.\n\nMy\xf6s Googlen lis\xe4\xe4ntyv\xe4t kulut verkkosivujen uudistamisesta kertoivat sis\xe4ll\xf6ntuotannon lis\xe4\xe4ntymisest\xe4. Tuotanto lis\xe4\xe4ntyi Googlen julkistaman ensimm\xe4isen raportin mukaan vuoden 2011 ensimm\xe4isest\xe4 nelj\xe4nneksest\xe4 vuoden 2013 ensimm\xe4iseen nelj\xe4nnekseen verrattuna 10 prosentilla, ja sis\xe4ll\xf6ntuotannon kustannukset kasvoivat samana aikana 48 prosenttia.\n\nGooglen taloudellisen kirjanpidon lukujen perusteella Google+ sai my\xf6s viime vuonna entist\xe4 korkeammat resurssit, mutta kulujen kasvattaminen vaikutti j\xe4\xe4dytykseen.\n\nGoogle on jo aiemmin keskitt\xe4nyt ty\xf6voimansa Android-k\xe4ytt\xf6j\xe4rjestelm\xe4\xe4n, ja samankaltainen kehitys voi my\xf6s Google+:n osalta olla tulossa.]" time="1.282"><properties><property name="score" value="0.9669402837753296" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.96694028&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.96694028
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[the belief or doctrine that the universe is eternal and has no beginning or end.\n\nthe belief or doctrine that the universe is eternal and has no beginning or end.\n\na worldview, esp. in ancient India, that denies all distinctions between spirit and matter.\n\nthe doctrine that the universe is infinite and without beginning or end.\n\nthe doctrine that the universe is infinite and without beginning or end.\n\na form of pantheism that assumes the presence of god as a force or energy pervading the universe.\n\nthe belief that the universe is eternal and has always existed.\n\nthe belief that the universe is eternal and has always existed.\n\nthe doctrine that there is only one god.\n\na philosophy that stresses the unity and continuity of existence.\n\na philosophy that stresses the unity and continuity of existence.\n\nthe philosophical doctrine that there are no immaterial or transcendent entities, such as God or the soul, but that all phenomena of reality may be accounted for by material means.\n\na philosophy holding that truth is apprehended by intuition and not by reason.\n\nthe doctrine that God is not a person but an impersonal principle that accounts for the universe.\n\nthe doctrine that the universe is not a personal, creative agent but a blind, orderly, insensible machine.\n\nthe philosophical doctrine that the universe is not a personal, creative agent but a blind, orderly, insensible machine.\n\nthe doctrine that natural phenomena have no supernatural explanation.\n\na doctrine that is similar to pantheism but is regarded as more fundamental, asserting the essential unity of the universe.\n\nthe doctrine that God, nature, and humanity are not distinct entities but are one.\n\nthe belief that the universe is an organic unity, with every person and event affecting every other.\n\nthe belief that the universe is an organic unity, with every person and event affecting every other.\n\na worldview positing that all life and existence is inseparable from the interdependent parts of the whole.\n\nthe belief that there is an undying and immaterial part of man, esp. the soul.\n\nthe belief that there is an undying and immaterial part of man, esp. the soul.\n\nthe belief that there is a common and universal religious or moral experience in every person.\n\nthe belief that there is a common and universal religious or moral experience in every person.\n\nthe theory that all existing things in the world are formed from and constitute a unity with one fundamental substance, esp. matter or spirit.\n\nthe theory that all existing things in the world are formed from and constitute a unity with one fundamental substance, esp. matter or spirit.\n\nthe theory that only material things and properties exist.\n\nthe theory that only material things and properties exist.\n\nthe doctrine that the world and all its phenomena are merely the product of chance and necessity.\n\nthe doctrine that the world and all its phenomena are merely the product of chance and necessity.\n\nthe doctrine that man is no more than a natural product of evolution, with no spiritual aspect.\n\nthe doctrine that man is no more than a natural product of evolution, with no spiritual aspect.\n\na system of principles and practices based on the teachings of a spiritual leader.\n\nthe worship of idols or icons.\n\nthe worship of idols or icons.\n\na religious practice consisting of the use of some object as a focus of worship.\n\na religious practice consisting of the use of some object as a focus of worship.\n\nthe belief that ultimate reality cannot be expressed in ordinary language.\n\nthe belief that ultimate reality cannot be expressed in ordinary language.\n\nthe doctrine that all human beings possess a dual nature and can achieve a transcendent existence.\n\nthe doctrine that all human beings possess a dual nature and can achieve a transcendent existence.\n\nthe doctrine that only moral and spiritual values are real.\n\nthe doctrine that only moral and spiritual values are real.\n\na belief in the existence of a soul as an immaterial entity distinct from the body.\n\na belief in the existence of a soul as an immaterial entity distinct from the body.\n\nthe doctrine that knowledge is impossible.\n\nthe doctrine that knowledge is impossible.\n\nthe belief that everything is the result of chance.\n\nthe belief that everything is the result of chance.\n\nthe doctrine that all matter is intrinsically evil and that human existence is a burden imposed upon men by evil spirits.\n\nthe doctrine that all matter is intrinsically evil and that human existence is a burden imposed upon men by evil spirits.\n\nthe doctrine that knowledge is merely a by-product of learning and not a result of the processes of the mind.\n\nthe doctrine that knowledge is merely a by-product of learning and not a result of the processes of the mind.\n\nthe doctrine that God is unknowable and that religious truth can be attained only through intuition and inner experience.\n\nthe doctrine that God is unknowable and that religious truth can be attained only through intuition and inner experience.\n\nthe doctrine that life is an illusion and death is the reality.\n\nthe doctrine that life is an illusion and death is the reality.\n\nthe doctrine that the universe is composed of a single substance, such as matter or spirit, and that all things are modifications of this basic substance.\n\nthe doctrine that the universe is composed of a single substance, such as matter or spirit, and that all things are modifications of this basic substance.\n\nthe belief that all phenomena and events in the world are determined by necessary laws and have their causes in previous events.\n\nthe belief that all phenomena and events in the world are determined by necessary laws and have their causes in previous events.\n\nthe doctrine that matter and its motions are eternal and that God, if he exists, is wholly immanent in the world.\n\nthe doctrine that matter and its motions are eternal and that God, if he exists, is wholly immanent in the world.\n\na group of people sharing religious convictions.\n\na group of people sharing religious convictions.\n\nthe doctrine that the world is an aggregate of self-existent individuals.\n\nthe doctrine that the world is an aggregate of self-existent individuals.\n\na philosophy that identifies religion with a specific moral system.\n\nthe belief that knowledge of an eternal reality is gained through intuition.\n\nthe belief that knowledge of an eternal reality is gained through intuition.\n\nthe belief that a divine or other transcendent power has sovereignty over the universe.\n\nthe belief that a divine or other transcendent power has sovereignty over the universe.\n\na comprehensive system of political, economic, and social organization.\n\nthe practice of worshiping with songs, prayers, and other gestures in addition to rites and ceremonies.\n\nthe practice of worshiping with songs, prayers, and other gestures in addition to rites and ceremonies.\n\nthe doctrine that truth is gained through study and contemplation.\n\nthe doctrine that truth is gained through study and contemplation.\n\nthe theory that all qualities in the human mind are relative to the situation in which they are encountered.\n\nthe theory that all qualities in the human mind are relative to the situation in which they are encountered.\n\nthe belief that the divine spark of reason is in all human beings.\n\nthe belief that the divine spark of reason is in all human beings.\n\n]" time="3.684"><properties><property name="score" value="0.9612496569752693" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[The US Forest Service has announced that the Fern Lake Fire is fully contained and the evacuation advisory for North of Tofte has been lifted. This comes after firefighters spent the last week building up the fire lines around the blaze. It also means that Minnesota State Highway 61 is now open to traffic again, as are Highway 1 and Highway 61 (Forest Roads 18 and 55), which run through the fire area.\n\nAs of Monday morning, there are still seven aircraft and 380 firefighters assigned to the fire, which was estimated to be 70 percent contained. It has burned 6,900 acres and its cause remains under investigation.\n\nThe next update is expected by Friday morning.]" time="0.340"><properties><property name="score" value="0.9994612336158752" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[In my world it's important to be organized. Especially when you live with more than two people, which I do. Most of the time, everything is on a schedule. No one is allowed to eat at 11pm unless they're hungry. Everyone is not allowed to leave dirty dishes in the sink unless they're at work. If you're working out at a certain time, you have to shower. No one can go outside unless it's to throw out the garbage. And so on and so forth.\n\nFor some reason, though, the whole &quot;organization&quot; idea flies right out the window when I'm trying to schedule a trip. I tend to plan things to the day and then get annoyed if something goes wrong or someone else wants to do something that may interfere. I hate it when someone wants to do something at a time that I've already booked for another thing. I feel like I should have priority, especially since I've planned this whole thing and if it were cancelled, I would have to pay for everything anyway.\n\nBut, guess what? I've realized I'm wrong. I've realized that while I may have planned the whole thing, other people need to be included too. And that they deserve a say in what goes on too. Maybe that means they have to drive back a different route to get home or they need to leave an hour earlier to get to their house. Maybe they need to stay later than they would have planned. Or maybe they don't want to leave. And that's ok. It doesn't have to ruin anything.\n\nAs long as everyone is included and has a say, it's all good. It doesn't mean that there won't be stress involved. There will be. I'm not going to lie. We may end up fighting about some things. Or they might leave early, which makes us upset. But at the end of the day, we're still family. We still love each other. We still get to have fun.\n\nI guess what I'm saying is that sometimes, I need to let other people in. To let them have a say in what happens. Otherwise, it will just be me against them. And I'd rather be the bigger person. I'd rather be the one who says &quot;Fine, I'll do it this way&quot; instead of the one who gets in an argument. I don't want to be the one who stays up late crying over something that went wrong and couldn't be avoided.\n\nSo, to sum up my whole blog:]" time="1.255"><properties><property name="score" value="0.9855835636456808" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[&quot;Would You Accept A Ridiculously Generous Gift?\n\nThe book, Cults Inside Out, claims to reveal everything you ever wanted to know about cults, from recruiting and controlling techniques to the manipulation of the children of cult members. The book is written by Jeremy Spencer, who claims that his grandfather was a cult leader and that his father was in a cult. It is hoped that by publishing this book that Jeremy Spencer can stop this cycle and prevent any other children from being hurt in this way.\n\nA spokesperson for Jeremy Spencer has confirmed that he is desperately in need of some cash to fund the marketing and promotion of the book. The cost is simply too great to risk doing it alone. He hopes that if he can raise enough money to pay for advertising, then he will be able to share his story and help people.\n\nIf Jeremy Spencer can raise enough money for the book, then he will gift it to the person who donates the most money. The donation can be made via the crowdfunding site Kickstarter.com.\n\nDonating to this crowdfunding campaign is likely to have more benefits than simply receiving a copy of the book. You can be part of the book and even get a speaking part in it. This is a great opportunity to help someone with a great story.&quot;\n\nKickstarter is a crowdfunding website that allows you to donate money to people who are trying to raise money for something. In return for your donations, you can receive gifts from the person you are donating to.\n\nWould You Accept A Ridiculously Generous Gift?\n\nA friend of mine contacted me a few months ago to ask if I would consider reviewing a book for them. The book was called Cults Inside Out and the author is called Jeremy Spencer. He has offered to donate the book to the highest bidder.\n\nCults Inside Out is the first of a trilogy and Jeremy Spencer is trying to raise money to self-publish the second book. As part of his plan, Jeremy is offering to donate the book to the highest bidder. The highest bidder is whoever donates the most money to Jeremy's GoFundMe account.\n\nWould you accept a ridiculously generous gift? I decided that I would.\n\nJeremy Spencer is an interesting character. I had never heard of him before and I was sceptical about some of the things that he had to say. Jeremy is a bit of a controversial character and he is no stranger to the press. He is very controversial and he has a history of being sued for libel, but you can find out more about that on his GoFundMe page.\n\nThe book is about cults and how they use manipulation techniques to recruit people into their organisations. The book also talks about the psychology of the cult members and how they get manipulated. There is some fascinating stuff in there and it is an easy read.\n\nThe book starts with the stories of two children who were born into cults. One child was the son of a cult leader and the other child was the daughter of another cult leader. Both children were exposed to extreme forms of cult abuse and you can really feel the pain that they went through.\n\nI didn't really enjoy reading the stories of the two children because it was so sad. You could really feel their pain. I am glad that Jeremy was able to escape the cult, but I don't think that it is an easy thing to do. There are always going to be repercussions, no matter how many years later.\n\nAfter the stories of the two children, Jeremy goes on to talk about the psychology of the cult members. I didn't realise that there were so many different reasons why people join cults. There are a number of factors that can influence the decision to join a cult.\n\nThere are a number of strategies that cults use to recruit new members and Jeremy goes on to discuss the techniques that they use. He talks about how they use manipulation and persuasion techniques to encourage people to join them. Jeremy is not the first person to talk about these strategies, but he does explain them in a very clear way.\n\nJeremy also talks about the different kinds of cults and what they are like. There are religious cults and political cults and there are even business cults. There are all kinds of cults and it is fascinating to read about how they are structured and what they are all about.\n\nJeremy Spencer has also included information about his family and his time in the cult. I have never been in a cult, so I don't know what it is like, but I was impressed with how well Jeremy handled the subject. He didn't talk about the cult in a negative way and he didn't try to make it sound worse than it was.\n\nThere is a lot of information about cults in this book and Jeremy doesn't try to overstate any of the facts. He doesn't try to talk in hyperbole and he gives very factual descriptions of what it is like to be in a cult.\n\nIf you are a fan of cults, then this book is going to be right up your street. If you are just interested in psychology, then this book is also going to be right up your street. There are a lot of interesting theories and a lot of fascinating facts.\n\nThere are a number of interesting stories and you can really feel the emotion in this book. I am glad that I have read it, but I don't think that I will read the rest of the trilogy.\n\nWhat you will get if you make the highest bid\n\nIf you make the highest bid on Jeremy's GoFundMe page, then he will send you a copy of Cults Inside Out. The money that you donate will be used to help with the marketing and promotion of the book.\n\nI have seen people donate large amounts of money to crowdfunding projects, so it is definitely possible. Jeremy is even offering to sign your copy of the book, if you make the highest bid.\n\nThe book is available in paperback and as an e-book. You will also be able to get it in various different languages. It is likely to be available in a number of different countries.\n\nYou can get a signed copy of the book for just $30 and Jeremy has been very generous with his offers. There are some very generous rewards if you can make the highest bid. Jeremy is determined to get this book out and he will give you the opportunity to help him.\n\nIf you are interested in making the highest bid, then click here. If you are interested in reading more about the book, then click here.\n\nWhat are your thoughts about Jeremy Spencer's generosity? Please leave a comment below.\n\nSave\n\nSave\n\nSave\n\nSave]" time="3.349"><properties><property name="score" value="0.98844484772001" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl['I'd only ever been with two girls'\n\nMany guys worry about their first time having sex \u2013 especially when it comes to having sex for the first time with a girl. How can you be sure she's comfortable? And what should you do if she says no?\n\nHere's how to handle your first time.\n\nGetting ready for the first time\n\nEven though your first time might be really exciting, you need to remember that having sex for the first time can be really awkward.\n\nThis is especially true if you're nervous, so to make it easier for both of you, try to relax, and make sure you're both feeling comfortable and happy to do it. You don't want to be worrying about whether your partner's enjoying it.\n\nThe best way to do this is to talk about it before you get started. Find out what she expects, and what she doesn't. Try to find out how much she knows about sex. It might seem strange to talk about this, but it will make it easier for you both when you get down to it.\n\nYou can talk about contraception too. Both of you should know how you're going to stop pregnancy happening and what to do if it does.\n\nA lot of guys worry about getting their first time over with, but in fact, the longer you take, the more you'll enjoy it. Most girls feel more comfortable the longer it takes. If you try to rush things, you're more likely to make her feel uncomfortable and she might not enjoy it.\n\nWhat if you're with someone new?\n\nIf you're having sex with someone for the first time, you may be nervous. You're not sure what to expect.\n\nYou're likely to be worried that you won't know how to do it properly or that you'll be rubbish in bed. But the fact is, it doesn't matter how good you are. All that matters is that you both enjoy it and that you're safe.\n\nYour worries might make you want to rush things and this might make you less sensitive and not as gentle as you should be. This is especially true if you're with a virgin. She'll be more nervous than you, so make sure you take your time and that you both enjoy it.\n\nYou'll know if she's enjoying it if she shows signs of pleasure, like:\n\nmoaning\n\ngetting very wet (this is the most obvious one!)\n\nshe may make noises or say things to let you know she's enjoying it\n\nremember to look for these signs!\n\nFirst time with a girl: the sex\n\nSex for the first time can be really exciting.\n\nWhat if she says no?\n\nIf your first time involves having sex with a girl, there's a chance she might say no. That's normal. It's important to make sure she's happy to do it. It's your responsibility to make sure she's okay with it and you don't want to do anything that might hurt her or make her feel uncomfortable.\n\nEven if she says yes, she might change her mind when the time comes. This can be a little awkward, especially if you've been going out with her for a while and you've had sex before.\n\nIf this happens, try to deal with it like a man. If you want to take it further, tell her you really want to, but you won't if she's not up for it.\n\nIf she wants to stop, you should respect that. It can happen, even if she's wanted to have sex for a while. If you think she might be pregnant, make sure you take care of that too.\n\nThe actual sex\n\nIt's important that you know how to please a girl, but this is especially important the first time. It will help you build your confidence, and it will also help her enjoy sex more.\n\nHow do you please a girl?\n\nWhat most people think about when it comes to pleasing a girl is what feels good for her. But in fact, what feels good for you is important too. It's the only way you can both enjoy yourselves and have a great time.\n\nBe careful to go slowly. If you go too fast, you might scare her or make her uncomfortable.\n\nWhat if you're not having sex with a virgin?\n\nIf you're not having sex with a virgin, it's unlikely that you'll have to worry about losing your virginity. It's more likely that she will lose hers!\n\nWhen you have sex for the first time, it's still important that you take it slow. It might be even more important because you'll need to show her that you can be gentle and caring. This will make sure she feels comfortable with you. It will also make sure that you have a great time too.\n\nIs she ready for sex?\n\nYou need to know when she's ready to have sex. If you rush it, you could end up doing something that she's not ready for. This could make her feel embarrassed or uncomfortable, which could make you feel embarrassed too.\n\nIf you don't think she's ready, you shouldn't go ahead. That's because it might make her feel like she's not good enough.\n\nJust be patient and make sure you're both happy and comfortable.\n\nYou can help her decide when she's ready for sex by asking her. Ask her what she thinks about having sex. If she says no, respect that, and don't push it.\n\nIf she says yes, talk about what you both think. Ask her if she's ready for it and talk about what you both expect. You'll both feel more relaxed if you know what's going to happen and you're both ready for it.\n\nDon't forget, it's important that you're both ready, not just her!\n\nWhat if you're with a virgin?\n\nSex for the first time can be really scary. But remember, you're both in the same boat and you can help each other through it.\n\nIf you're with a virgin, you need to remember that she's not an expert. It will take her a while to get used to sex and you need to give her time. It will help if you make sure you go slowly.\n\nAs you know, she'll be nervous, so you need to make sure she feels comfortable.\n\nSome girls are really confident and they're great at giving directions. But other girls aren't, so if she needs help, you should be there to help her.\n\nHow do you make sure she's ready?\n\nYou might have to talk about it. You might feel like she's pushing you into having sex, but she might not actually be ready for it.\n\nIf you don't feel comfortable talking to her, you can find out by looking at her. Does she seem relaxed? Is she making the right noises? Is she happy? If she seems happy, you're probably okay.\n\nMake sure you ask her about contraception too. It's important to keep yourself safe, so make sure you both know how you're going to prevent pregnancy.\n\nWhat if you have problems?\n\nSex for the first time can be really scary for some guys. But the good news is that it doesn't have to be.\n\nWhat if you're having trouble getting an erection?\n\nIf you're having trouble getting an erection, it could be that you're nervous. It could also be that you don't feel ready for sex.\n\nIf it's nerves, you should try to relax. Try to take it slowly. You don't want to scare your partner. If you're not ready for sex, then you need to tell her.\n\nIt might help to talk to her about it first. If you think you're ready, but she isn't, you should respect that. You don't want to make her feel bad or make her regret having sex with you.\n\nIt's important that you're both ready for it and that you both feel comfortable. If you're not ready, don't go through with it.\n\nWhat if it doesn't feel right?\n\nSome guys worry about whether they're doing it right. It can be a bit scary if you don't know what to do.\n\nIt's really important that you talk about it. If you think it's not right, you should ask her.\n\nAnd don't worry too much. There's no right way to have sex. The important thing is that both of you are enjoying it and that you're both safe.\n\nIt's really important that you both make sure you're ready for sex. You don't want to go through with it if she's not ready.\n\nIf you're having sex with a virgin, then you need to be more careful. She might not know what she's doing, so it's important that you tell her what you want. You should also tell her if it doesn't feel right.\n\nWhat if it hurts?\n\nIf you're having sex for the first time, there's a chance that it could hurt. But the good news is that this should only happen the first time. If it happens again, you should get help as soon as possible.\n\nYour partner should tell you if it hurts, so you need to listen to her. You should also be able to tell if it hurts. If you can't, then it's not right.\n\nIf it hurts her, then you need to stop and think about what you're doing. It's important that you both know what feels good and what doesn't.\n\nIf you're not sure what to do, then it might be best to practice with a toy. You can use it to find out what she likes and what she doesn't.\n\n]" time="4.860"><properties><property name="score" value="0.9640485379430983" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Public Bookmark *\n\nJames \u201cBucky\u201d Barnes was born in the dead of night, right at the edge of winter and the approaching storm. At least, that\u2019s how his mother likes to tell it. Her mother, a woman who had been born with the gift of sight, had foretold that James would be born during a storm, that his destiny would be caught up in a whirlwind, that he would live a life surrounded by shadows. And in the shadow of the war, there were few people more surrounded by them.\n\nHe lived in the shadow of Steve Rogers and it was only in the wake of his death that James could finally step out from it, could step out of the shadows.\n\nIt is a slow process, this transition. He tries to move forward, but it is difficult.\n\nSome habits are harder to break than others.]" time="0.407"><properties><property name="score" value="0.9855735898017883" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[This paper extends the framework of the intertemporal Tobit model in which explanatory variables are available only at the end of the sample to the multiple dependent variable case. The effects of endogeneity of these explanatory variables on the standard error of the estimated structural coefficients are quantified. A large endogeneity bias, whose size is not affected by the lag length, is found for all sets of variables, except the set with a positive effect on the last observed outcome, which is relatively immune to endogeneity bias. These results are used to analyze the effects of unobserved variables on the estimates of the true effects and on the standard errors of the estimated effects of observed variables. The impact of unobserved variables on the estimates is sometimes substantial.]" time="0.359"><properties><property name="score" value="0.9987894892692566" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Post by Liz \xbb Fri Jan 03, 2018 10:08 am\n\nIt was a tough morning for the county road department this morning with all the road closures due to high water and downed trees. Public Works reports it was a busy day with well over 50 customers reporting issues with their service. The county was very busy all day responding to reported issues, working to clear downed trees, and repairing road damage.\n\nBelow are some pictures from this morning of some of the damage caused by this storm.\n\n\n\nOne of the downed trees on Lakeshore Road.\n\nPictured here is the removal of a large tree on Lakeshore Road.\n\nAnother downed tree on Lakeshore Road\n\nMore trees down on Lakeshore Road\n\nDebris on Lakeshore Road.\n\nTree down on Lewis Street.]" time="0.401"><properties><property name="score" value="0.9998204112052917" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Gary, I don't know who or what the new &quot;Oldsmobile&quot; will be. At this point I have not even seen the car. I think they are planning to unveil it in the spring.\n\n\n\nWhat I would like to see would be an Olds branded CUV with the attitude of the current Enclave. If they can do that, I would buy one. If not, they will be back at square one in the CUV market.\n\n\n\nThe other big issue is the financial status of GM. Is it just teetering on the edge, or has it already fallen into the abyss? If they go bankrupt, will the Olds brand survive? There are a lot of question marks surrounding the new Olds.]" time="0.335"><properties><property name="score" value="0.9998059868812561" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[North Fork is a great neighborhood for enjoying Long Island. It is right next to the Merrick and Seaford-Oyster Bay Expressways. You can head south to get to the beaches of Long Island, east for the Hamptons, or head west to enjoy all the city has to offer. This is a great location if you want to be able to spend time in the city or on the island, or both!\n\n\n\n\n\n\n\n]" time="0.221"><properties><property name="score" value="0.9997227787971497" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[1. Recognition of the conflict by both sides: the de facto government and the organizations that support the peace process\n\nThis stage is important as it identifies the dynamics of the process and the relationship between the parties. If the process is not recognized by one of the parties, the process will be difficult and may not be sustainable.\n\n2. Opening channels: the expression of mutual understanding, for example through exchange of prisoners\n\n3. A period of dialogue: the exchange of prisoners or groups of prisoners in advance of dialogue, by agreement\n\n4. Dialogue: the willingness to talk\n\n5. Agreement: an agreement to implement a cease-fire, a political dialogue or to resolve any other issue\n\n6. Realization of an agreement: actual cessation of hostilities, the release of prisoners or a commitment to implement political reforms\n\n7. Stabilization: a return to normalcy\n\n8. The final stage: the long-term achievement of stability in which peace agreements are sustained.]" time="0.498"><properties><property name="score" value="0.9990487694740295" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[I have been following a special course at Stony Brook University called \u201cLifelong Entrepreneurship\u201d, in which I am studying what it means to run a successful business from an economic perspective. At the same time, I am also writing the script for the book of my life, and taking into account my own perception of what it takes to create a successful business.\n\nBoth have been going great, and this is because I have used all my life experience, as well as learning the skills of how to find a better job, how to start a business, how to build a website, and how to build a strong brand.\n\nThe lessons I have learned through the years, and in the course I am currently taking, is that creating a business is hard work. This applies to any industry, because it\u2019s not just about the idea. It\u2019s about executing the idea, bringing the right people on board, and having a company with the right culture, that truly wants to be a part of something.]" time="0.588"><properties><property name="score" value="0.9997609257698059" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Don't miss the return of the Vancouver Stars to the Coast Soccer League. Vancouver Stars. Score-by-Score. W: L. Goal Differential. Offside. Soccer positions (association football) The outfield players are divided into four main categories: forwards, midfielders, defenders, and goalkeepers. A team is composed of eleven players of which.Vancouver Stars - CSL Senior Division - 20.09.2017 - Whitecaps U-23 vs. StarsAway matches Away Score Home Result Date. The players involved with the Whitecaps' Whitecaps have appeared in an international youth match. Vancouver. In , the Whitecaps qualified for the semifinals of the CONCACAF Champions League for the first time after defeating Puerto Rico Islanders in the quarterfinals with a aggregate score of 4\u20143 and Club Tijuana in the semifinals with a aggregate score of 4\u20142. The following season, the Vancouver Whitecaps won their first CSL Championship. TSR: 0.69. MLS Reserves league records: Attendance.\n\nVancouver Whitecaps FC 2. Regular Season Stats. Soccer - BC Soccer Premier League. Soccer Statistics. Each match will be given a score at the end of the match based on the following scoring system: 3 points for a win, 1 point for a tie, and 0 points for a loss. Statistics - College Soccer - Pro/College. From Wikipedia, the free encyclopedia. The number of teams varies in each season. Check out Vancouver vs Ottawa Fury FC live score, video stream and H2H results. Vancouver Whitecaps F.C. (Vancouver Whitecaps or Whitecaps) is a Canadian professional soccer team based in Vancouver, British Columbia, that competes in Major League Soccer. Soccer Statistics - College Soccer - Pro/College. It was created as a reserve team of the MLS side Vancouver Whitecaps FC, in the United Soccer Leagues Premier Development League , the fourth tier of the American Soccer Pyramid, the Canadian Soccer Pyramid. In , the Vancouver Whitecaps won their first CSL Championship. Vancouver Whitecaps FC 2. The following season, the Vancouver Whitecaps won their first CSL Championship. MLS Reserves league records: Attendance. Vancouver Whitecaps FC. Vancouver Whitecaps FC 2.\n\nCoeur dalene hookup:]" time="1.350"><properties><property name="score" value="0.9549379348754883" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[\xa9 2018 Ontario Real Estate Association. All rights reserved. IDX information is provided exclusively for consumers' personal, non-commercial use and may not be used for any purpose other than to identify prospective properties consumers may be interested in purchasing. Information is deemed reliable but is not guaranteed accurate by the MLS or The Natalie Gabriele Real Estate Team. The data relating to real estate for sale on this website comes in part from the participating Associations/MLS's in the Regional IDX Program of the Toronto Real Estate Board. The properties displayed may not be all the properties available through the MLSR or the TRREB, or may include properties with unavailable data. Notice: The seller/owner of this property expressly disclaims any responsibility for any typographical errors, misinformation, or misprints and will be deemed to have made reasonable efforts to provide accurate information regarding the properties, including costs, terms, dimensions, locations, and amenities. This data is updated daily. Some properties that appear for sale on this website may subsequently have sold and may no longer be available. Information last updated on May 25, 2018 2:48 PM]" time="0.676"><properties><property name="score" value="0.8919932842254639" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.89199328&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.89199328
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Top Review by\n\nExcellent! I made this up yesterday for tonight's dinner and I'm really glad I did. This is a wonderful meal and it's so easy to put together. I used boneless, skinless thighs and was glad I did because the meat stayed moist and tender. This will definitely become a favorite in our house. I will have to make a double batch next time though because we couldn't stop eating it!!]" time="0.290"><properties><property name="score" value="0.5666831135749817" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Managing Leaky Slugs with Slugslinger\n\nYou know how it goes: You bought some organic, free-range eggs. After cracking them open for your morning omelet, you're left with a bunch of shells, or, the collective vernacular would have you believe, &quot;leaky slugs.&quot;\n\n&quot;What am I going to do with these slimy shells?&quot; you ask. &quot;This is just wasteful!&quot; Well, friends, that's where the Sl]" time="0.352"><properties><property name="score" value="0.988314151763916" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.98831415&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.98831415
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Vitaliy Shabaev\n\nVitaliy Shabaev (born June 22, 1984) is a Russian ice hockey goaltender. He is currently playing with Neftekhimik Nizhnekamsk in the Kontinental Hockey League (KHL). He is the brother of ice hockey player Sergei Shabaev.\n]" time="0.275"><properties><property name="score" value="0.7022149562835693" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Orthocisioides\n\nOrthocisioides is a genus of spiders in the Ctenidae family. It was first described in 1929 by Chamberlin &amp; Ivie. , it contains 12 species.\n\n&quot;Orthocisioides&quot; comprises the following species:\n]" time="0.210"><properties><property name="score" value="0.5616979002952576" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Anjaane\n\nAnjaane () is a 2003 Bollywood film, directed by Vikram Bhatt. It stars Ajay Devgn and Ayesha Takia in lead roles and Kunal Khemu, Nafisa Ali and Rajpal Yadav in supporting roles. The film is a remake of the American movie &quot;Unfaithful&quot;.\n\nThis movie was an average grosser at the box office.\n\nA middle class girl, Meghna (Ayesha Takia) lives with her divorced mother, Dr. Kamla Tiwari (Reema Lagoo), a well known psychiatrist, and her younger sister in the suburbs of Mumbai. Her parents had divorced when she was just two years old, and she has hardly seen her father, Colonel Rameshwar Prasad (Deepak Shirke), ever since.\n\nRajeev Mathur (Ajay Devgn), a renowned architect, is married to Sheetal (Ayesha Julka), who is the love of Meghna's life. Meghna falls in love with Rajeev, and they become involved in an illicit relationship. One day, while trying to evade Rajeev, Meghna witnesses an incident where her father, whom she had not met for more than twelve years, is found dead at a friend's farmhouse.\n\nShe goes to his funeral and is completely shocked to see Rajeev there. On confronting him, she is told that he and her father were old friends and colleagues. She accepts this, but not before starting to suspect that Rajeev has been trying to get her out of the way in order to take over her father's position. She confronts Rajeev on this, but he insists that this is not the case.\n\nMeghna, unable to bear the thought that her father was killed, decides to investigate on her own. She tracks down one of the servants at the farmhouse, who reveals that her father was shot, and that Rajeev was there, drunk, when it happened. She confronts Rajeev, who admits that he and her father were involved in a fraud, in which they stole money from one of their clients, and faked his death to escape justice. However, Rajeev swears that he has nothing to do with her father's death.\n\nShe believes him, and accepts that she was wrong in suspecting him. While trying to console her, he kisses her. They decide to meet the next day. However, on the way to her meeting with Rajeev, she finds him being confronted by her mother and sister, who saw him leaving their house. Rajeev confesses to them that he and Meghna have been involved in a relationship, but insists that they didn't do anything wrong.\n\nAfter much debate, her family decides to agree to the relationship. Rajeev insists that Meghna move in with them, but she is shocked to learn that she is pregnant with his child. She insists that Rajeev take a DNA test to prove that the child is his, and he does so.\n\nHe confesses to her that he was never married to his wife, and had no children. He had always been in love with Meghna, but had not told anyone about it, not even her father. Sheetal had left him for another man, who later divorced her, and Rajeev's father was killed by one of his clients, who then blackmailed him. He promises that they will run away together, but Meghna doesn't believe him.\n\nLater that night, Rajeev is killed, and Meghna is accused of the murder. Dr. Tiwari believes her daughter, and begins to suspect that Sheetal might be the murderer. The police refuse to believe Dr. Tiwari's story, and the family falls apart. Meghna is released, but her mother refuses to see her, and she is forced to live with her uncle and his family.\n\nWhen the child is born, the family is forced to admit that the child does bear a close resemblance to Rajeev. They forgive Meghna and re-accept her. She goes to Rajeev's mother (Nafisa Ali), and tells her that the child is Rajeev's. His mother is grief-stricken, but eventually decides to accept Meghna, and asks her to name the child, since Rajeev was also hers. Meghna decides to name the child after her father.\n\n\n\nAll songs were composed by Sandeep Chowta, written by Sameer, and performed by Udit Narayan and Kumar Sanu.\n\n]" time="2.431"><properties><property name="score" value="0.882601335644722" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Shawn Holmes\n\nShawn Holmes (born September 2, 1971) is an American former professional motocross racer.\n\nIn 1991, Holmes made the finals of the 250cc National Championship, finishing in 4th place. He spent most of his professional career racing in the 250cc Motocross Class on the Motocross Grand Prix circuit. He was a member of the 2000 Team Honda 250R National Team, and rode for Red Bull KTM Factory Team in 2002 and 2003. He was sponsored by Xpedition MX Gear from 2000-2009.\n\nHolmes also rode for team AGV, Azonic, and JMC during his career.\n\nHe is married to U.S. Olympic skier Stacey Cook.\n]" time="0.384"><properties><property name="score" value="0.9998226761817932" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Khudadat Rafibeyli\n\nKhudadat Rafibeyli (; 1842\u20131889) was an Azerbaijani novelist, playwright, translator, editor, teacher and publisher.\n\nRafibeyli was born in 1842 in Shamakhy. From 1853 to 1860 he attended the Shamakhy religious school, and then later at the Shamakhi Muslim Boarding School. In 1862, he studied at the Nizamiyya Madrasah in Ganja. At first, he worked as a teacher, but then he moved to Tiflis where he studied at the Caucasian Muslim Gymnasium from 1864 to 1870. Then he returned to Shamakhi and taught at his old school until 1876, when he left for Tiflis again. In 1877, he moved to Baku and worked as a deputy editor and later as an editor at the &quot;Hayat&quot; newspaper. He lived in Tiflis for a third time from 1878 to 1884, and then returned to Baku, where he established a publishing house named &quot;Abdulgaffar&quot;, which published the first Azerbaijani newspaper &quot;Bakinskiy Listok&quot;. He died in 1889.\n\nRafibeyli is considered the founder of the Azerbaijani novel. He is also the author of numerous plays. His best-known plays include &quot;Mumtaz Khan&quot; (1882), &quot;Seyid Huseyn&quot; (1886), &quot;Arshin Mal Alan&quot; (1884), &quot;Khanlar Bahadur&quot; (1886), and &quot;Shirin&quot; (1887). His plays mainly revolve around the heroic deeds of people in the history of Azerbaijan, but his historical plays do not have a profound philosophical content. As well as plays, he also wrote novels and short stories such as &quot;Koroghlu&quot; (1872), &quot;Xirayi&quot; (1872), &quot;Jabal&quot; (1872), &quot;Dashtadem&quot; (1873), &quot;Shah Abbas and Kamaleddin Efendi&quot; (1879), &quot;Bahadur Khan Shaku&quot; (1880), &quot;Shah Mansur&quot; (1882), &quot;Shirin&quot; (1886), &quot;Sulthan Mahmud&quot; (1888), &quot;Dervish Khudadat&quot; (1886), &quot;Agha Mammad Hasan&quot; (1888), &quot;Qataghani&quot; (1889), &quot;Khurshidbanu Natavan&quot; (1889), &quot;Nebiy&quot; (1890), &quot;Unsulli&quot; (1890), and &quot;Gulbakhor&quot; (1890).\n\nHis two most famous works are &quot;Seyid Huseyn&quot; (1886) and &quot;Arshin Mal Alan&quot; (1884). &quot;Seyid Huseyn&quot; was written after the Karabakh liberation movement (1830-1835) and was first staged in 1873. It is a work of realism and it depicts the life of the Azerbaijani feudal landlord Seyid Huseyn. It was written in the tradition of oriental tales and was praised for its artistic value and the depiction of people. &quot;Arshin Mal Alan&quot; is a work of realist-idealistic nature. It tells about the greed of people and how this brings misfortune to them. This work was praised for the beauty of its language and for the pathos of its characters.\n\nIn addition to his plays and novels, Rafibeyli translated several plays by Russian playwrights into Azerbaijani.\n\nRaf]" time="1.855"><properties><property name="score" value="0.8500946760177612" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[We the People of the United States, in Order to form a more perfect Union, establish Justice, insure domestic Tranquility, provide for the common defence, promote the general Welfare, and secure the Blessings of Liberty to ourselves and our Posterity, do ordain and establish this Constitution for the United States of America.\n\nIn looking at the definition of the word constitution, it becomes clear that the purpose of the Constitution was to establish a form of government for the United States of America. This was done to protect the people from the tyranny of the government, and also protect the government from the whims of the people. In this way the people would be ensured freedom to live their lives as they saw fit and the government would be ensured freedom to protect the people from foreign and domestic threats.\n\nIn order to be a free and sovereign people, it was necessary to have a proper government to guard those rights.\n\nHere is the full text of the Constitution:\n\nhttp://www.archives.gov/exhibits/charters/constitution.html\n\nThe text of the Constitution has been added to this blog. It may be found at the following address:\n\nhttp://www.debbieschlussel.com/2012/01/21/constitution-of-the-united-states/]" time="0.698"><properties><property name="score" value="0.9260361790657043" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[You May Also Like\n\nFree quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template Free quote template]" time="4.460"><properties><property name="score" value="0.9994729608297348" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[More videos from Ross Technology\n\nHow to Keep Our Assets Safe in High Risk Situations by Ross Technology\n\nby\n\nDave Devoe, COO at Ross Technology, Inc. explains the importance of keeping our assets safe in high-risk situations.]" time="0.163"><properties><property name="score" value="0.9815222024917603" /></properties></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Schon lange ist Kuba umweltfreundlich und wirtschaftspolitisch auf dem vorderen Rang anzusiedeln. Hier sind dementsprechend einige russische Technologien in den vergangenen Jahren installiert worden. Vor allem Stromerzeugung von Solarzellen hat sich f\xfcr Kuba als erfolgreich erwiesen. In den vergangenen f\xfcnf Jahren wurden 2.000 Solarzellen in Kuba eingesetzt.\n\nDer Nutzen von Solarzellen f\xfcr Kuba\n\nSolarenergie f\xf6rdert auf Kuba gesundheitlich viele Vorteile. Erstens erzeugt sie Strom, was f\xfcr einen Staat, der wie Kuba so viele wichtige Schritte in Sachen Energie f\xfcr eine modernisierte Gesellschaft geht, eine gro\xdfe Rolle spielt. Die Energiegewinnung wird damit auch in Kuba demographisch ausgeglichen.\n\nZweitens l\xf6st die Sonnenenergie den direkten Nutzen f\xfcr die medizinische Versorgung aus. Das bedeutet, dass Solarzellen auch f\xfcr die Versorgung von medizinischem Ger\xe4t eingesetzt werden k\xf6nnen, wodurch wiederum die Gesundheit der Menschen verbessert wird. So gesehen ist die Installation von Solarzellen eine entscheidende Sache, die f\xfcr eine langfristige Gesundheit von Menschen gesorgt hat.\n\nWie werden Solarzellen produziert?\n\nAuch wenn das Wetter auf Kuba eine gro\xdfe Rolle spielt, kann man aufgrund der Klima\xe4nderungen auf der Insel mit einem bedeutenden Einfluss auf die Anwendung von Solarenergie rechnen. Daher werden die Solarzellen auf Kuba in gro\xdfem Umfang eingesetzt. Wie genau das im Einzelnen geschieht, zeigt das Beispiel des Leuchtturms \u201eLa Revoluci\xf3n\u201c in der Provinz Las Tunas. Dieser wurde mit Solarzellen gebaut, weil die vorhandene Infrastruktur einige Herausforderungen mit sich bringt. Die Installation der Solarzellen hat dementsprechend unmittelbare Erfolge mit sich gebracht.\n\nDer Leuchtturm befindet sich auf einer Klippe, die einen Abstand von 1.250 Metern zum Festland hat. Es gibt keine M\xf6glichkeit, in den Turm ein Stromkabel einzuziehen. Das ist deshalb ein Problem, weil dieser Turm eine gewisse Gr\xf6\xdfe hat. Die M\xf6glichkeit, dass der Strom aus der Natur umweltfreundlich gewonnen werden kann, wird durch die Installation der Solarzellen genutzt. Wie es ebenfalls erw\xe4hnt wurde, befindet sich der Leuchtturm auf einer Klippe und somit auch einige Kilometer von der Siedlung entfernt. Deshalb ist es hier sehr lohnenswert, Solarzellen zu installieren.\n\nDie gro\xdfe Stromproduktion des Leuchtturms ist das Resultat des Einsatzes von Solarenergie. Der gesamte Energiebedarf wird damit gedeckt. Solarzellen haben sich dementsprechend als eine \xe4u\xdferst wertvolle Sache f\xfcr Kuba erwiesen. Dadurch wird eine stromproduzierende Stelle sichergestellt, die unabh\xe4ngig vom Regenwetter ist. Somit haben Solarzellen einen unglaublich hohen Nutzen f\xfcr Kuba.\n\nFazit\n\nDie Installation von Solarzellen auf Kuba stellt somit einen sehr wichtigen Baustein f\xfcr das Funktionieren der Gesellschaft dar. Die Produktion von Strom aus Sonnenenergie f\xf6rdert die Gesundheit, da sie dem Bev\xf6lkerungswachstum Rechnung tr\xe4gt. Solarzellen sind daher auch in Kuba eine wichtige Sache, um die Energieversorgung zu sichern.]" time="2.830"><properties><property name="score" value="0.9998112320899963" /></properties><failure message="AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99981123&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/GPT-3-175b_samples.jsonl is an LLM-generated sample, misclassified as human-generated with confidence 0.99981123
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_gpt3_jsonl[Cardiac intervention in men is a critical procedure performed with the best equipment available. Sometimes, however, there are complications that are difficult to solve. For example, the images provided by a coronary angiography may not be clear enough for the doctor to determine the problem. In this situation, it is necessary to perform an intervention for coronary angioplasty. This is a technique that was introduced in the 1970s and has been improved many times over the years.\n\nWhat is coronary angioplasty?\n\nCoronary angioplasty is a technique that allows the narrowing of a coronary artery to be dilated. This is done with the use of a catheter that has a flexible balloon at its end. This is inflated and deflated so that the artery is dilated. Once the dilation is achieved, a stent is implanted to keep the artery dilated and allow blood to flow properly. The stent is a small metal ring that is placed inside the lumen of the artery. It acts as a scaffold and prevents the vessel from closing. This is important because, if this happened, the part of the heart that receives blood from this artery would not receive any more.\n\nCoronary angioplasty is performed under general anesthesia. The patient is given nitroglycerin so that the pain caused by the dilation of the artery is less intense. Sometimes, if the patient has chest pain, an injection of painkillers may be needed.\n\nAfter the procedure is completed, the patient is observed to make sure he has not had any complications. If the patient develops a hematoma, the artery is usually clamped again and the hematoma is removed by inserting a small wire.\n\nThe risk of complications is low. They can develop in patients who have had angioplasty many times or who have had previous heart attacks.\n\nThe main advantage of angioplasty is that it provides immediate relief for the patient. If the stent works well, the patient may need to take medication for the rest of his life. This is an important advantage for those who have suffered a heart attack and are at risk of a second one.\n\nIn certain cases, it is necessary to repeat the intervention several times until the lumen of the artery becomes stable. Usually, the procedure is done after 4-6 weeks, depending on the patient\u2019s progress.]" time="1.196"><properties><property name="score" value="0.3618270953496297" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i0]" time="1.105"><properties><property name="score" value="0.33803772926330566" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i1]" time="1.288"><properties><property name="score" value="0.9909436305363973" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i2]" time="0.723"><properties><property name="score" value="0.21650105714797974" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i3]" time="0.714"><properties><property name="score" value="0.9852544367313385" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i4]" time="0.792"><properties><property name="score" value="0.9996868371963501" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i5]" time="0.842"><properties><property name="score" value="0.15916651487350464" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i6]" time="1.292"><properties><property name="score" value="0.8950265645980835" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i7]" time="0.812"><properties><property name="score" value="0.9805915355682373" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i8]" time="1.376"><properties><property name="score" value="0.4221643606821696" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i9]" time="1.543"><properties><property name="score" value="0.9949692885080973" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i10]" time="0.644"><properties><property name="score" value="0.9889686703681946" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i11]" time="1.401"><properties><property name="score" value="0.9995559453964233" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i12]" time="0.709"><properties><property name="score" value="0.9032338857650757" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i13]" time="1.253"><properties><property name="score" value="0.4955202341079712" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i14]" time="0.703"><properties><property name="score" value="0.004489243030548096" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i15]" time="1.274"><properties><property name="score" value="0.9933535854021708" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i16]" time="1.315"><properties><property name="score" value="0.9751461346944174" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i17]" time="0.983"><properties><property name="score" value="0.9997092485427856" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i18]" time="1.297"><properties><property name="score" value="0.9994335969289144" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i19]" time="1.549"><properties><property name="score" value="0.9996739625930786" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i20]" time="0.457"><properties><property name="score" value="0.9998303651809692" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i21]" time="1.318"><properties><property name="score" value="0.9996842940648397" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i22]" time="0.511"><properties><property name="score" value="0.9996777772903442" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i23]" time="1.474"><properties><property name="score" value="0.930377463499705" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i24]" time="0.588"><properties><property name="score" value="0.24683469533920288" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i25]" time="0.975"><properties><property name="score" value="0.9998238682746887" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i26]" time="1.004"><properties><property name="score" value="0.9997778534889221" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i27]" time="1.480"><properties><property name="score" value="0.999743382136027" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i28]" time="0.861"><properties><property name="score" value="0.9996185302734375" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i29]" time="0.614"><properties><property name="score" value="0.16794809699058533" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i30]" time="0.999"><properties><property name="score" value="0.9776257276535034" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i31]" time="0.476"><properties><property name="score" value="0.9997774958610535" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i32]" time="1.172"><properties><property name="score" value="0.999674121538798" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i33]" time="1.450"><properties><property name="score" value="0.5434567332267761" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i34]" time="1.560"><properties><property name="score" value="0.996519406636556" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i35]" time="0.758"><properties><property name="score" value="0.9756788015365601" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i36]" time="1.204"><properties><property name="score" value="0.9997512896855673" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i37]" time="1.395"><properties><property name="score" value="0.9998273650805155" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i38]" time="1.006"><properties><property name="score" value="0.8932798902193705" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i39]" time="1.292"><properties><property name="score" value="0.9887590805689493" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i40]" time="1.319"><properties><property name="score" value="0.9987603425979614" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i41]" time="0.651"><properties><property name="score" value="0.10070693492889404" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i42]" time="1.174"><properties><property name="score" value="0.9277148842811584" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i43]" time="0.802"><properties><property name="score" value="0.9995073080062866" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i44]" time="1.331"><properties><property name="score" value="0.37132346630096436" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i45]" time="1.276"><properties><property name="score" value="0.8899807532628378" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i46]" time="1.359"><properties><property name="score" value="0.3044572671254476" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i47]" time="1.243"><properties><property name="score" value="0.9976347287495931" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i48]" time="1.207"><properties><property name="score" value="0.8850578864415487" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i49]" time="0.938"><properties><property name="score" value="0.9250558316707611" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i50]" time="1.027"><properties><property name="score" value="0.9997213780879974" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i51]" time="1.380"><properties><property name="score" value="0.899264375368754" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i52]" time="1.253"><properties><property name="score" value="0.9976159930229187" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i53]" time="1.178"><properties><property name="score" value="0.8728211124738058" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i54]" time="1.290"><properties><property name="score" value="0.41954439878463745" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i55]" time="1.483"><properties><property name="score" value="0.8843557387590408" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i56]" time="1.160"><properties><property name="score" value="0.9995518326759338" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i57]" time="0.702"><properties><property name="score" value="0.9406405389308929" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i58]" time="1.128"><properties><property name="score" value="0.9853964447975159" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i59]" time="0.948"><properties><property name="score" value="0.9991322755813599" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i60]" time="0.726"><properties><property name="score" value="0.9986104071140289" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i61]" time="0.858"><properties><property name="score" value="0.9996419250965118" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i62]" time="1.488"><properties><property name="score" value="0.9693665901819865" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i63]" time="1.665"><properties><property name="score" value="0.9996444384256998" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i64]" time="1.377"><properties><property name="score" value="0.7763914664586385" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i65]" time="1.406"><properties><property name="score" value="0.9926962455113729" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i66]" time="1.397"><properties><property name="score" value="0.4900870124499003" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i67]" time="1.315"><properties><property name="score" value="0.9998161196708679" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i68]" time="1.319"><properties><property name="score" value="0.9929679234822592" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i69]" time="1.442"><properties><property name="score" value="0.9975364406903585" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i70]" time="1.198"><properties><property name="score" value="0.9997075001398722" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i71]" time="1.072"><properties><property name="score" value="0.8668341437975565" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i72]" time="1.391"><properties><property name="score" value="0.9930587808291117" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i73]" time="1.416"><properties><property name="score" value="0.9991673628489176" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i74]" time="1.314"><properties><property name="score" value="0.9984186689058939" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i75]" time="1.056"><properties><property name="score" value="0.9998294711112976" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i76]" time="1.383"><properties><property name="score" value="0.9823128382364908" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i77]" time="0.959"><properties><property name="score" value="0.9998108446598053" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i78]" time="1.111"><properties><property name="score" value="0.999560683965683" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i79]" time="1.397"><properties><property name="score" value="0.9949497977892557" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i80]" time="1.449"><properties><property name="score" value="0.9275935292243958" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i81]" time="0.752"><properties><property name="score" value="0.9986278712749481" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i82]" time="0.447"><properties><property name="score" value="0.9998196959495544" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i83]" time="0.482"><properties><property name="score" value="0.9386201500892639" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i84]" time="1.345"><properties><property name="score" value="0.9997007846832275" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i85]" time="1.458"><properties><property name="score" value="0.9997418522834778" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i86]" time="0.983"><properties><property name="score" value="0.9931584000587463" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i87]" time="0.587"><properties><property name="score" value="0.02845624089241028" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i88]" time="0.883"><properties><property name="score" value="0.9997766017913818" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i89]" time="0.941"><properties><property name="score" value="0.1178252100944519" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i90]" time="1.330"><properties><property name="score" value="0.9342873692512512" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i91]" time="0.880"><properties><property name="score" value="0.9926381707191467" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i92]" time="1.297"><properties><property name="score" value="0.9735442201296488" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i93]" time="1.507"><properties><property name="score" value="0.9992989301681519" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i94]" time="1.535"><properties><property name="score" value="0.9997666080792745" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i95]" time="1.329"><properties><property name="score" value="0.9639825026194254" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i96]" time="0.633"><properties><property name="score" value="0.8643794655799866" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i97]" time="0.851"><properties><property name="score" value="0.9998237788677216" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i98]" time="1.420"><properties><property name="score" value="0.9997178713480631" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i99]" time="0.923"><properties><property name="score" value="0.9997498691082001" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i100]" time="0.590"><properties><property name="score" value="0.989646703004837" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i101]" time="1.315"><properties><property name="score" value="0.9990643262863159" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i102]" time="0.833"><properties><property name="score" value="0.9998265206813812" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i103]" time="0.952"><properties><property name="score" value="0.999778687953949" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i104]" time="1.455"><properties><property name="score" value="0.999812384446462" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i105]" time="1.556"><properties><property name="score" value="0.9997713168462118" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i106]" time="0.765"><properties><property name="score" value="0.9991114437580109" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i107]" time="1.258"><properties><property name="score" value="0.9775026241938273" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i108]" time="0.902"><properties><property name="score" value="0.9985421597957611" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i109]" time="0.616"><properties><property name="score" value="0.9938603639602661" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i110]" time="0.978"><properties><property name="score" value="0.9998281896114349" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i111]" time="1.550"><properties><property name="score" value="0.9978554646174113" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i112]" time="1.120"><properties><property name="score" value="0.9998286962509155" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i113]" time="0.752"><properties><property name="score" value="0.9990198612213135" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i114]" time="0.660"><properties><property name="score" value="0.7760953307151794" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i115]" time="1.411"><properties><property name="score" value="0.994027316570282" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i116]" time="0.783"><properties><property name="score" value="0.13339316844940186" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i117]" time="1.279"><properties><property name="score" value="0.4955790837605794" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i118]" time="1.435"><properties><property name="score" value="0.9992024302482605" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i119]" time="0.930"><properties><property name="score" value="0.9987260103225708" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i120]" time="0.855"><properties><property name="score" value="0.7633346021175385" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i121]" time="1.207"><properties><property name="score" value="0.9949203729629517" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i122]" time="0.372"><properties><property name="score" value="0.9997890591621399" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i123]" time="1.386"><properties><property name="score" value="0.9949796597162882" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i124]" time="1.250"><properties><property name="score" value="0.8591053684552511" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i125]" time="1.225"><properties><property name="score" value="0.999702125787735" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i126]" time="1.384"><properties><property name="score" value="0.9997357527414957" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i127]" time="1.217"><properties><property name="score" value="0.9997792641321818" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i128]" time="1.111"><properties><property name="score" value="0.3779126803080241" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i129]" time="1.053"><properties><property name="score" value="0.9947935044765472" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i130]" time="0.716"><properties><property name="score" value="0.9870868921279907" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i131]" time="1.335"><properties><property name="score" value="0.9919312795003256" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i132]" time="1.169"><properties><property name="score" value="0.8799259463946024" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i133]" time="1.749"><properties><property name="score" value="0.9997896154721578" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i134]" time="0.924"><properties><property name="score" value="0.9996479749679565" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i135]" time="0.671"><properties><property name="score" value="0.9865493774414062" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i136]" time="0.776"><properties><property name="score" value="0.012389898300170898" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i137]" time="0.388"><properties><property name="score" value="0.999779999256134" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i138]" time="0.643"><properties><property name="score" value="0.8451639413833618" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i139]" time="0.894"><properties><property name="score" value="0.9996845126152039" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i140]" time="1.200"><properties><property name="score" value="0.9638854662577311" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i141]" time="1.121"><properties><property name="score" value="0.43379439910252887" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i142]" time="1.271"><properties><property name="score" value="0.9996374845504761" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i143]" time="1.056"><properties><property name="score" value="0.30324288209279376" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i144]" time="1.313"><properties><property name="score" value="0.9997352759043375" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i145]" time="0.852"><properties><property name="score" value="0.9454884827136993" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i146]" time="0.821"><properties><property name="score" value="0.9992068707942963" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i147]" time="0.842"><properties><property name="score" value="0.9998197853565216" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i148]" time="1.233"><properties><property name="score" value="0.9664163788159689" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i149]" time="0.951"><properties><property name="score" value="0.9997240900993347" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i150]" time="1.180"><properties><property name="score" value="0.9995720187822977" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i151]" time="0.854"><properties><property name="score" value="0.9997027814388275" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i152]" time="1.070"><properties><property name="score" value="0.8554263512293497" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i153]" time="1.404"><properties><property name="score" value="0.9816227356592814" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i154]" time="1.276"><properties><property name="score" value="0.8549584547678629" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i155]" time="1.139"><properties><property name="score" value="0.9791819850603739" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i156]" time="0.905"><properties><property name="score" value="0.9997849464416504" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i157]" time="0.992"><properties><property name="score" value="0.9997730851173401" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i158]" time="0.828"><properties><property name="score" value="0.9157783389091492" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i159]" time="1.175"><properties><property name="score" value="0.9679648280143738" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i160]" time="0.906"><properties><property name="score" value="0.9998067319393158" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i161]" time="0.765"><properties><property name="score" value="0.9859792590141296" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i162]" time="1.310"><properties><property name="score" value="0.34770673513412476" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i163]" time="0.792"><properties><property name="score" value="0.9607382118701935" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i164]" time="0.788"><properties><property name="score" value="0.9773186147212982" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i165]" time="1.246"><properties><property name="score" value="0.9997359216213226" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i166]" time="1.548"><properties><property name="score" value="0.9957007169723511" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i167]" time="1.146"><properties><property name="score" value="0.9838612973690033" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i168]" time="0.603"><properties><property name="score" value="0.9997808337211609" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i169]" time="0.916"><properties><property name="score" value="0.9815739989280701" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i170]" time="1.210"><properties><property name="score" value="0.3290391167004903" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i171]" time="1.485"><properties><property name="score" value="0.9951123595237732" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i172]" time="1.108"><properties><property name="score" value="0.9995128214359283" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i173]" time="0.882"><properties><property name="score" value="0.9973663091659546" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i174]" time="0.899"><properties><property name="score" value="0.9998238980770111" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i175]" time="1.545"><properties><property name="score" value="0.9717009663581848" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i176]" time="1.529"><properties><property name="score" value="0.9949198365211487" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i177]" time="0.700"><properties><property name="score" value="0.10073179006576538" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i178]" time="1.357"><properties><property name="score" value="0.9185638825098673" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i179]" time="1.281"><properties><property name="score" value="0.9881505370140076" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i180]" time="1.320"><properties><property name="score" value="0.9922490914662679" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i181]" time="1.245"><properties><property name="score" value="0.9570252696673075" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i182]" time="1.177"><properties><property name="score" value="0.955115795135498" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i183]" time="1.742"><properties><property name="score" value="0.4999205470085144" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i184]" time="1.549"><properties><property name="score" value="0.9997824629147848" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i185]" time="0.433"><properties><property name="score" value="0.9982485771179199" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.99824858&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.99824858
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i186]" time="1.581"><properties><property name="score" value="0.9997588793436686" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i187]" time="1.530"><properties><property name="score" value="0.5978996008634567" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i188]" time="0.980"><properties><property name="score" value="0.9998172223567963" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i189]" time="0.834"><properties><property name="score" value="0.9994898736476898" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i190]" time="0.814"><properties><property name="score" value="0.11377820372581482" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i191]" time="1.143"><properties><property name="score" value="0.4779413342475891" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i192]" time="1.403"><properties><property name="score" value="0.9997712969779968" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i193]" time="1.438"><properties><property name="score" value="0.9998119076093038" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i194]" time="1.223"><properties><property name="score" value="0.9569030602773031" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i195]" time="1.142"><properties><property name="score" value="0.936046838760376" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i196]" time="1.442"><properties><property name="score" value="0.9998053510983785" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i197]" time="0.873"><properties><property name="score" value="0.9435257315635681" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i198]" time="1.361"><properties><property name="score" value="0.9997823039690653" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i199]" time="1.284"><properties><property name="score" value="0.9997632304827372" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i200]" time="0.532"><properties><property name="score" value="0.7708204090595245" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i201]" time="0.650"><properties><property name="score" value="0.7888340651988983" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i202]" time="1.128"><properties><property name="score" value="0.45915019512176514" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i203]" time="1.099"><properties><property name="score" value="0.9997828602790833" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i204]" time="1.296"><properties><property name="score" value="0.9980584383010864" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i205]" time="1.013"><properties><property name="score" value="0.005759567022323608" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.00575957&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.00575957
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i206]" time="1.714"><properties><property name="score" value="0.9997466206550598" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i207]" time="1.429"><properties><property name="score" value="0.998549242814382" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i208]" time="0.858"><properties><property name="score" value="0.9998161494731903" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i209]" time="0.521"><properties><property name="score" value="0.9996397495269775" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.99963975&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.99963975
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i210]" time="1.660"><properties><property name="score" value="0.9996961951255798" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i211]" time="1.489"><properties><property name="score" value="0.9906701644261678" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i212]" time="0.699"><properties><property name="score" value="0.9994919002056122" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i213]" time="1.496"><properties><property name="score" value="0.9997666080792745" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i214]" time="1.474"><properties><property name="score" value="0.9997553030649821" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i215]" time="1.537"><properties><property name="score" value="0.9997705022493998" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i216]" time="0.592"><properties><property name="score" value="0.9998192191123962" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i217]" time="1.336"><properties><property name="score" value="0.9998032848040262" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i218]" time="1.385"><properties><property name="score" value="0.9979944626490275" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i219]" time="1.219"><properties><property name="score" value="0.4017282525698344" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i220]" time="0.882"><properties><property name="score" value="0.9939359724521637" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i221]" time="1.308"><properties><property name="score" value="0.40284043550491333" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i222]" time="0.687"><properties><property name="score" value="0.9257247149944305" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i223]" time="1.532"><properties><property name="score" value="0.36213167508443195" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i224]" time="1.171"><properties><property name="score" value="0.933285137017568" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i225]" time="1.338"><properties><property name="score" value="0.9913025895754496" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i226]" time="1.122"><properties><property name="score" value="0.9998289346694946" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i227]" time="0.934"><properties><property name="score" value="0.9258537590503693" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i228]" time="1.332"><properties><property name="score" value="0.9958118200302124" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i229]" time="1.456"><properties><property name="score" value="0.32380592823028564" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i230]" time="1.458"><properties><property name="score" value="0.9978257020314535" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i231]" time="1.347"><properties><property name="score" value="0.9196378986040751" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i232]" time="0.769"><properties><property name="score" value="0.9920766651630402" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i233]" time="1.140"><properties><property name="score" value="0.44510964552561444" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i234]" time="0.993"><properties><property name="score" value="0.0019770562648773193" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i235]" time="1.466"><properties><property name="score" value="0.9997947017351786" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i236]" time="0.829"><properties><property name="score" value="0.9997296035289764" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i237]" time="1.264"><properties><property name="score" value="0.9870392481486002" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i238]" time="0.564"><properties><property name="score" value="0.9998205304145813" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i239]" time="0.841"><properties><property name="score" value="0.9994953274726868" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i240]" time="1.745"><properties><property name="score" value="0.5818083882331848" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i241]" time="1.509"><properties><property name="score" value="0.9982151389122009" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i242]" time="0.592"><properties><property name="score" value="0.9998216032981873" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i243]" time="0.957"><properties><property name="score" value="0.46856488784154254" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i244]" time="1.484"><properties><property name="score" value="0.9951716661453247" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i245]" time="0.931"><properties><property name="score" value="0.9887405335903168" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i246]" time="1.315"><properties><property name="score" value="0.9292695919672648" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i247]" time="1.643"><properties><property name="score" value="0.9997970461845398" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i248]" time="1.599"><properties><property name="score" value="0.9998156428337097" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i249]" time="0.444"><properties><property name="score" value="0.8398001790046692" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.83980018&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.83980018
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i250]" time="0.379"><properties><property name="score" value="0.9997627139091492" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i251]" time="1.057"><properties><property name="score" value="0.9998066127300262" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i252]" time="0.918"><properties><property name="score" value="0.9997933208942413" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i253]" time="1.221"><properties><property name="score" value="0.999396562576294" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i254]" time="0.729"><properties><property name="score" value="0.9990728795528412" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i255]" time="0.804"><properties><property name="score" value="0.9991051554679871" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i256]" time="0.955"><properties><property name="score" value="0.9997564852237701" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i257]" time="0.372"><properties><property name="score" value="0.999830961227417" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i258]" time="0.748"><properties><property name="score" value="0.9907004535198212" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i259]" time="1.408"><properties><property name="score" value="0.9730121890703837" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i260]" time="1.272"><properties><property name="score" value="0.8662718137105306" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i261]" time="0.994"><properties><property name="score" value="0.9998304843902588" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i262]" time="1.214"><properties><property name="score" value="0.9997014800707499" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i263]" time="1.105"><properties><property name="score" value="0.359250287214915" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i264]" time="1.275"><properties><property name="score" value="0.9998231331507365" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i265]" time="1.057"><properties><property name="score" value="0.9998226463794708" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i266]" time="1.469"><properties><property name="score" value="0.9993946353594462" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i267]" time="0.591"><properties><property name="score" value="0.16414836049079895" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i268]" time="1.323"><properties><property name="score" value="0.9989624619483948" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i269]" time="1.593"><properties><property name="score" value="0.9973666866620382" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i270]" time="1.591"><properties><property name="score" value="0.9996719559033712" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i271]" time="1.392"><properties><property name="score" value="0.470283846060435" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i272]" time="1.228"><properties><property name="score" value="0.4949455261230469" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i273]" time="0.722"><properties><property name="score" value="0.9983004629611969" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i274]" time="1.165"><properties><property name="score" value="0.9963137904802958" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i275]" time="1.196"><properties><property name="score" value="0.9995774626731873" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i276]" time="0.851"><properties><property name="score" value="0.9700413346290588" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i277]" time="1.721"><properties><property name="score" value="0.9988305370012919" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i278]" time="0.453"><properties><property name="score" value="0.999764621257782" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i279]" time="0.693"><properties><property name="score" value="0.9688458740711212" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i280]" time="0.855"><properties><property name="score" value="0.9997222423553467" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i281]" time="1.475"><properties><property name="score" value="0.9998233119646708" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i282]" time="1.069"><properties><property name="score" value="0.3680997093518575" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i283]" time="0.860"><properties><property name="score" value="0.14331617951393127" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i284]" time="1.171"><properties><property name="score" value="0.9069510300954183" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i285]" time="1.391"><properties><property name="score" value="0.9992924133936564" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i286]" time="0.910"><properties><property name="score" value="0.997644305229187" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i287]" time="0.844"><properties><property name="score" value="0.22944581508636475" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i288]" time="1.396"><properties><property name="score" value="0.977345327536265" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i289]" time="1.492"><properties><property name="score" value="0.999823530515035" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i290]" time="0.810"><properties><property name="score" value="0.9996005892753601" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i291]" time="1.208"><properties><property name="score" value="0.46785082419713336" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i292]" time="1.131"><properties><property name="score" value="0.9835694034894308" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i293]" time="1.210"><properties><property name="score" value="0.9998272955417633" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i294]" time="0.818"><properties><property name="score" value="0.8657499849796295" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.86574998&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.86574998
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i295]" time="1.215"><properties><property name="score" value="0.99820609887441" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i296]" time="1.099"><properties><property name="score" value="0.9998141527175903" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i297]" time="0.488"><properties><property name="score" value="0.9998027682304382" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i298]" time="1.148"><properties><property name="score" value="0.9303361773490906" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i299]" time="1.423"><properties><property name="score" value="0.33334418137868244" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i300]" time="1.315"><properties><property name="score" value="0.8967978755633036" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i301]" time="0.937"><properties><property name="score" value="0.23168018460273743" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i302]" time="1.464"><properties><property name="score" value="0.9967179695765177" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i303]" time="1.380"><properties><property name="score" value="0.8995934724807739" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i304]" time="1.152"><properties><property name="score" value="0.9944936831792196" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i305]" time="0.882"><properties><property name="score" value="0.999821275472641" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i306]" time="1.438"><properties><property name="score" value="0.9997816880544027" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i307]" time="1.511"><properties><property name="score" value="0.9996403654416403" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i308]" time="0.946"><properties><property name="score" value="0.9997463822364807" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i309]" time="1.659"><properties><property name="score" value="0.99978635708491" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i310]" time="1.275"><properties><property name="score" value="0.9722052415211996" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i311]" time="1.140"><properties><property name="score" value="0.3776501814524333" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i312]" time="1.553"><properties><property name="score" value="0.9827204346656799" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i313]" time="0.807"><properties><property name="score" value="0.9962862133979797" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i314]" time="0.952"><properties><property name="score" value="0.9998031854629517" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i315]" time="1.375"><properties><property name="score" value="0.9997701446215311" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i316]" time="1.370"><properties><property name="score" value="0.9752153952916464" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i317]" time="1.080"><properties><property name="score" value="0.4668264190355937" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i318]" time="0.681"><properties><property name="score" value="0.7951816916465759" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i319]" time="1.565"><properties><property name="score" value="0.9995260238647461" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i320]" time="1.028"><properties><property name="score" value="0.9971193671226501" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i321]" time="1.182"><properties><property name="score" value="0.3165487249692281" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i322]" time="0.700"><properties><property name="score" value="0.8113732039928436" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i323]" time="1.230"><properties><property name="score" value="0.999673088391622" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i324]" time="1.433"><properties><property name="score" value="0.3314584294954936" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i325]" time="1.362"><properties><property name="score" value="0.9990200400352478" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i326]" time="1.157"><properties><property name="score" value="0.9318121472994486" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i327]" time="1.389"><properties><property name="score" value="0.9997791647911072" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i328]" time="1.535"><properties><property name="score" value="0.9993801712989807" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i329]" time="1.516"><properties><property name="score" value="0.8596369425455729" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i330]" time="1.286"><properties><property name="score" value="0.9839429060618082" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i331]" time="1.433"><properties><property name="score" value="0.34853412707646686" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i332]" time="1.143"><properties><property name="score" value="0.9998250007629395" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i333]" time="1.248"><properties><property name="score" value="0.999597430229187" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i334]" time="1.416"><properties><property name="score" value="0.9997936089833578" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i335]" time="1.398"><properties><property name="score" value="0.9820278088251749" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i336]" time="0.998"><properties><property name="score" value="0.9997231662273407" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i337]" time="0.729"><properties><property name="score" value="0.9998116493225098" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i338]" time="1.493"><properties><property name="score" value="0.9993781646092733" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i339]" time="1.089"><properties><property name="score" value="0.9941166440645853" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i340]" time="1.623"><properties><property name="score" value="0.9997347593307495" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i341]" time="0.701"><properties><property name="score" value="0.9992089569568634" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i342]" time="0.857"><properties><property name="score" value="0.9970196485519409" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i343]" time="1.233"><properties><property name="score" value="0.9997150103251139" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i344]" time="1.086"><properties><property name="score" value="0.8686124881108602" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i345]" time="1.123"><properties><property name="score" value="0.9996887842814127" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i346]" time="0.836"><properties><property name="score" value="0.9911136031150818" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i347]" time="1.503"><properties><property name="score" value="0.9997722506523132" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i348]" time="1.279"><properties><property name="score" value="0.9726395805676779" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i349]" time="1.336"><properties><property name="score" value="0.9668539961179098" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i350]" time="1.470"><properties><property name="score" value="0.9996213515599569" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i351]" time="1.490"><properties><property name="score" value="0.943882405757904" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i352]" time="0.689"><properties><property name="score" value="0.9942803978919983" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i353]" time="1.261"><properties><property name="score" value="0.9473390976587931" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i354]" time="1.395"><properties><property name="score" value="0.9998175303141276" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i355]" time="1.032"><properties><property name="score" value="0.22016912698745728" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i356]" time="1.223"><properties><property name="score" value="0.4081790049870809" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i357]" time="1.343"><properties><property name="score" value="0.9744828343391418" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i358]" time="1.339"><properties><property name="score" value="0.9997121691703796" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i359]" time="1.148"><properties><property name="score" value="0.4651947021484375" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i360]" time="0.939"><properties><property name="score" value="0.9997012913227081" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i361]" time="0.931"><properties><property name="score" value="0.9998263120651245" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i362]" time="0.872"><properties><property name="score" value="0.998675137758255" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i363]" time="1.131"><properties><property name="score" value="0.910467286904653" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i364]" time="1.118"><properties><property name="score" value="0.3645209272702535" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i365]" time="1.038"><properties><property name="score" value="0.9997944831848145" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i366]" time="1.236"><properties><property name="score" value="0.9730593164761862" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i367]" time="1.433"><properties><property name="score" value="0.9997708598772684" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i368]" time="1.383"><properties><property name="score" value="0.9356527328491211" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i369]" time="0.721"><properties><property name="score" value="0.7681886553764343" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i370]" time="1.385"><properties><property name="score" value="0.9997631708780924" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i371]" time="1.575"><properties><property name="score" value="0.9997720122337341" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i372]" time="1.310"><properties><property name="score" value="0.9998096426328024" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i373]" time="0.862"><properties><property name="score" value="0.9917227029800415" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i374]" time="0.459"><properties><property name="score" value="0.9998243451118469" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i375]" time="1.083"><properties><property name="score" value="0.9998009204864502" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i376]" time="0.931"><properties><property name="score" value="0.9998080730438232" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i377]" time="0.890"><properties><property name="score" value="0.9997630715370178" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i378]" time="0.734"><properties><property name="score" value="0.996850460767746" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i379]" time="1.037"><properties><property name="score" value="0.9978292286396027" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i380]" time="1.018"><properties><property name="score" value="0.9990755617618561" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i381]" time="0.603"><properties><property name="score" value="0.9067650139331818" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i382]" time="1.424"><properties><property name="score" value="0.9982543587684631" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i383]" time="0.757"><properties><property name="score" value="0.9996610879898071" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i384]" time="1.413"><properties><property name="score" value="0.9996083378791809" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i385]" time="0.688"><properties><property name="score" value="0.9877739548683167" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i386]" time="1.399"><properties><property name="score" value="0.9848360419273376" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i387]" time="1.190"><properties><property name="score" value="0.9994925061861674" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i388]" time="1.419"><properties><property name="score" value="0.999783972899119" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i389]" time="0.664"><properties><property name="score" value="0.9331748187541962" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.93317482&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.93317482
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i390]" time="1.102"><properties><property name="score" value="0.9997382462024689" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i391]" time="1.502"><properties><property name="score" value="0.6146008372306824" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i392]" time="0.848"><properties><property name="score" value="0.9326446056365967" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i393]" time="1.320"><properties><property name="score" value="0.992564857006073" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i394]" time="1.161"><properties><property name="score" value="0.9997221827507019" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i395]" time="1.338"><properties><property name="score" value="0.9891208608945211" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i396]" time="1.127"><properties><property name="score" value="0.9557789862155914" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i397]" time="1.302"><properties><property name="score" value="0.9607579310735067" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i398]" time="1.684"><properties><property name="score" value="0.3332634170850118" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i399]" time="0.589"><properties><property name="score" value="0.9996505975723267" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i400]" time="0.974"><properties><property name="score" value="0.8328061103820801" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i401]" time="1.302"><properties><property name="score" value="0.9988013307253519" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i402]" time="1.412"><properties><property name="score" value="0.9997929533322653" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i403]" time="0.919"><properties><property name="score" value="0.999637633562088" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i404]" time="1.159"><properties><property name="score" value="0.8559324741363525" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i405]" time="1.032"><properties><property name="score" value="0.9997300207614899" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i406]" time="1.203"><properties><property name="score" value="0.9978288014729818" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i407]" time="0.737"><properties><property name="score" value="0.9964352548122406" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i408]" time="1.096"><properties><property name="score" value="0.9997871816158295" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i409]" time="1.327"><properties><property name="score" value="0.9788492520650228" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i410]" time="1.483"><properties><property name="score" value="0.9997337460517883" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i411]" time="1.522"><properties><property name="score" value="0.390387495358785" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i412]" time="1.406"><properties><property name="score" value="0.9190970063209534" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i413]" time="1.172"><properties><property name="score" value="0.9675678809483846" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i414]" time="0.900"><properties><property name="score" value="0.6857260763645172" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.68572608&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.68572608
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i415]" time="1.576"><properties><property name="score" value="0.9997232357660929" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i416]" time="1.967"><properties><property name="score" value="0.9600036442279816" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i417]" time="1.363"><properties><property name="score" value="0.3333057165145874" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i418]" time="1.137"><properties><property name="score" value="0.9997202157974243" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i419]" time="0.905"><properties><property name="score" value="0.9997901916503906" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i420]" time="1.512"><properties><property name="score" value="0.33728774388631183" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i421]" time="0.961"><properties><property name="score" value="0.9930786192417145" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i422]" time="0.854"><properties><property name="score" value="0.9827590882778168" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i423]" time="0.839"><properties><property name="score" value="0.9997731149196625" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i424]" time="1.213"><properties><property name="score" value="0.9896275401115417" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i425]" time="1.545"><properties><property name="score" value="0.9988274574279785" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i426]" time="0.497"><properties><property name="score" value="0.9998225569725037" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i427]" time="0.846"><properties><property name="score" value="0.9891308844089508" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i428]" time="1.364"><properties><property name="score" value="0.999810258547465" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i429]" time="0.671"><properties><property name="score" value="0.12494552135467529" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i430]" time="1.377"><properties><property name="score" value="0.3334817091623942" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i431]" time="0.928"><properties><property name="score" value="0.9962072670459747" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i432]" time="1.215"><properties><property name="score" value="0.4870337247848511" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i433]" time="1.593"><properties><property name="score" value="0.9997305274009705" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i434]" time="1.470"><properties><property name="score" value="0.322096327940623" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i435]" time="0.602"><properties><property name="score" value="0.04333442449569702" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.04333442&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.04333442
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i436]" time="1.064"><properties><property name="score" value="0.9998150169849396" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i437]" time="1.104"><properties><property name="score" value="0.9177845120429993" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i438]" time="0.528"><properties><property name="score" value="0.999747097492218" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i439]" time="0.702"><properties><property name="score" value="0.9608858227729797" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i440]" time="1.058"><properties><property name="score" value="0.9935017625490824" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i441]" time="1.354"><properties><property name="score" value="0.9987963438034058" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i442]" time="1.467"><properties><property name="score" value="0.3763480583826701" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i443]" time="0.985"><properties><property name="score" value="0.9998284578323364" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i444]" time="1.026"><properties><property name="score" value="0.9988632202148438" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i445]" time="1.395"><properties><property name="score" value="0.9951344132423401" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i446]" time="0.906"><properties><property name="score" value="0.9997049868106842" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i447]" time="0.661"><properties><property name="score" value="0.9954937100410461" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i448]" time="0.842"><properties><property name="score" value="0.9907903969287872" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i449]" time="0.830"><properties><property name="score" value="0.936634749174118" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.93663475&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.93663475
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i450]" time="1.444"><properties><property name="score" value="0.9995803833007812" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i451]" time="0.695"><properties><property name="score" value="0.9304200708866119" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i452]" time="0.792"><properties><property name="score" value="0.9953885972499847" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i453]" time="1.085"><properties><property name="score" value="0.8908483187357584" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i454]" time="0.819"><properties><property name="score" value="0.9201360642910004" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i455]" time="0.968"><properties><property name="score" value="0.9988473951816559" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i456]" time="1.201"><properties><property name="score" value="0.9936461051305135" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i457]" time="0.767"><properties><property name="score" value="0.9996996819972992" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i458]" time="0.648"><properties><property name="score" value="0.8752228319644928" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i459]" time="1.548"><properties><property name="score" value="0.9995403091112772" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i460]" time="1.063"><properties><property name="score" value="0.9997361898422241" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i461]" time="0.705"><properties><property name="score" value="0.8234121799468994" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i462]" time="1.608"><properties><property name="score" value="0.999176025390625" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i463]" time="1.074"><properties><property name="score" value="0.9907151858011881" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i464]" time="1.374"><properties><property name="score" value="0.9989715218544006" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i465]" time="0.961"><properties><property name="score" value="0.997618705034256" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i466]" time="1.350"><properties><property name="score" value="0.9997843702634176" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i467]" time="1.453"><properties><property name="score" value="0.9997875690460205" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i468]" time="1.269"><properties><property name="score" value="0.36824264129002887" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i469]" time="1.313"><properties><property name="score" value="0.9985217253367106" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i470]" time="1.005"><properties><property name="score" value="0.9998142421245575" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i471]" time="1.006"><properties><property name="score" value="0.22085893154144287" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.22085893&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.22085893
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i472]" time="0.620"><properties><property name="score" value="0.18621662259101868" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i473]" time="0.888"><properties><property name="score" value="0.9027727246284485" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i474]" time="0.744"><properties><property name="score" value="0.9873996078968048" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i475]" time="1.144"><properties><property name="score" value="0.9308174252510071" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i476]" time="1.198"><properties><property name="score" value="0.7309227585792542" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i477]" time="1.094"><properties><property name="score" value="0.9997558295726776" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i478]" time="0.994"><properties><property name="score" value="0.9255541563034058" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i479]" time="1.226"><properties><property name="score" value="0.4798274040222168" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i480]" time="1.440"><properties><property name="score" value="0.9957079092661539" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i481]" time="0.979"><properties><property name="score" value="0.9777027070522308" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i482]" time="1.461"><properties><property name="score" value="0.5250254422426224" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i483]" time="0.794"><properties><property name="score" value="0.03184893727302551" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i484]" time="0.814"><properties><property name="score" value="0.9997502863407135" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i485]" time="1.254"><properties><property name="score" value="0.4917664925257365" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i486]" time="1.231"><properties><property name="score" value="0.9717037677764893" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i487]" time="0.739"><properties><property name="score" value="0.9287342727184296" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i488]" time="0.950"><properties><property name="score" value="0.9997928142547607" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i489]" time="0.895"><properties><property name="score" value="0.999799907207489" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i490]" time="1.429"><properties><property name="score" value="0.9970369537671407" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i491]" time="1.209"><properties><property name="score" value="0.481997549533844" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i492]" time="1.215"><properties><property name="score" value="0.3737957278887431" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i493]" time="0.911"><properties><property name="score" value="0.031074106693267822" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i494]" time="1.225"><properties><property name="score" value="0.9569424192110697" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i495]" time="0.369"><properties><property name="score" value="0.9812410473823547" /></properties><failure message="AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.98124105&#10;assert 'AI' == 'Human'&#10;  - Human&#10;  + AI">E   AssertionError: samples/news.jsonl is a human-generated sample, misclassified as AI-generated with confidence 0.98124105
+    assert 'AI' == 'Human'
+      - Human
+      + AI</failure></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i496]" time="1.007"><properties><property name="score" value="0.9997844696044922" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i497]" time="0.986"><properties><property name="score" value="0.9973243474960327" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i498]" time="1.682"><properties><property name="score" value="0.363461176554362" /></properties></testcase><testcase classname="test_roberta_detect" name="test_humannews_jsonl[i499]" time="1.991"><properties><property name="score" value="0.9997951785723368" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i0]" time="1.442"><properties><property name="score" value="0.9984716375668844" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i1]" time="1.288"><properties><property name="score" value="0.3971804579099019" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.39718046&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.39718046
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i2]" time="0.766"><properties><property name="score" value="0.21446457505226135" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.21446458&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.21446458
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i3]" time="0.951"><properties><property name="score" value="0.01565384864807129" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01565385&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01565385
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i4]" time="0.843"><properties><property name="score" value="0.008036643266677856" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i5]" time="0.991"><properties><property name="score" value="0.005597591400146484" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00559759&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00559759
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i6]" time="1.055"><properties><property name="score" value="0.9818304777145386" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i7]" time="1.024"><properties><property name="score" value="0.9998103976249695" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i8]" time="1.439"><properties><property name="score" value="0.9982912540435791" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i9]" time="1.014"><properties><property name="score" value="0.11578366160392761" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i10]" time="0.662"><properties><property name="score" value="0.9109793305397034" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i11]" time="1.227"><properties><property name="score" value="0.9996596574783325" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i12]" time="0.643"><properties><property name="score" value="0.986138105392456" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i13]" time="1.004"><properties><property name="score" value="0.16550901532173157" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i14]" time="0.816"><properties><property name="score" value="0.9984333515167236" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i15]" time="1.120"><properties><property name="score" value="0.9917864998181661" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i16]" time="1.046"><properties><property name="score" value="0.32962626218795776" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i17]" time="0.842"><properties><property name="score" value="0.9974329173564911" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i18]" time="1.113"><properties><property name="score" value="0.9989084998766581" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.9989085&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.9989085
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i19]" time="1.549"><properties><property name="score" value="0.5452145487070084" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i20]" time="0.537"><properties><property name="score" value="0.9980182647705078" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i21]" time="1.251"><properties><property name="score" value="0.33240509033203125" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i22]" time="0.613"><properties><property name="score" value="0.8025331497192383" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i23]" time="1.361"><properties><property name="score" value="0.4001869360605876" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i24]" time="0.550"><properties><property name="score" value="0.9768226444721222" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i25]" time="0.957"><properties><property name="score" value="0.9832465946674347" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i26]" time="1.002"><properties><property name="score" value="0.004813969135284424" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00481397&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00481397
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i27]" time="1.321"><properties><property name="score" value="0.9993326663970947" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i28]" time="0.699"><properties><property name="score" value="0.009102582931518555" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i29]" time="0.718"><properties><property name="score" value="0.9952973425388336" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i30]" time="0.778"><properties><property name="score" value="0.9985516369342804" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.99855164&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.99855164
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i31]" time="0.429"><properties><property name="score" value="0.8445727229118347" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.84457272&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.84457272
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i32]" time="1.218"><properties><property name="score" value="0.9981215000152588" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i33]" time="1.181"><properties><property name="score" value="0.9975699782371521" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i34]" time="1.273"><properties><property name="score" value="0.3351157307624817" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i35]" time="0.691"><properties><property name="score" value="0.9845669567584991" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.98456696&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.98456696
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i36]" time="1.434"><properties><property name="score" value="0.38238511482874554" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i37]" time="1.635"><properties><property name="score" value="0.3328786293665568" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i38]" time="1.222"><properties><property name="score" value="0.4947515328725179" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.49475153&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.49475153
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i39]" time="1.297"><properties><property name="score" value="0.2151894768079122" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i40]" time="1.860"><properties><property name="score" value="0.45227333903312683" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i41]" time="0.847"><properties><property name="score" value="0.05955237150192261" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i42]" time="1.231"><properties><property name="score" value="0.9998224973678589" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i43]" time="0.968"><properties><property name="score" value="0.9994767606258392" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i44]" time="1.213"><properties><property name="score" value="0.9659350713094076" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i45]" time="1.318"><properties><property name="score" value="0.2908060948053996" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i46]" time="1.342"><properties><property name="score" value="0.9991037249565125" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i47]" time="1.059"><properties><property name="score" value="0.9741453925768534" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i48]" time="0.889"><properties><property name="score" value="0.07056388258934021" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07056388&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.07056388
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i49]" time="0.744"><properties><property name="score" value="0.9994921684265137" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i50]" time="1.046"><properties><property name="score" value="0.9351189335187277" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i51]" time="1.250"><properties><property name="score" value="0.9615907470385233" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i52]" time="1.156"><properties><property name="score" value="0.18217414617538452" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i53]" time="1.054"><properties><property name="score" value="0.33729161818822223" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i54]" time="1.950"><properties><property name="score" value="0.560616135597229" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i55]" time="1.204"><properties><property name="score" value="0.9771826267242432" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i56]" time="1.007"><properties><property name="score" value="0.9984424710273743" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i57]" time="0.660"><properties><property name="score" value="0.04373675584793091" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04373676&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.04373676
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i58]" time="1.011"><properties><property name="score" value="0.345586617787679" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i59]" time="0.877"><properties><property name="score" value="3.2007694244384766e-05" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i60]" time="0.826"><properties><property name="score" value="0.02622130513191223" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i61]" time="0.884"><properties><property name="score" value="0.09790685772895813" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.09790686&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.09790686
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i62]" time="1.169"><properties><property name="score" value="0.9777837594350179" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i63]" time="1.295"><properties><property name="score" value="0.3607455293337504" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i64]" time="1.094"><properties><property name="score" value="0.20523607730865479" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i65]" time="1.089"><properties><property name="score" value="0.4964865446090698" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i66]" time="1.174"><properties><property name="score" value="0.37407634655634564" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i67]" time="1.148"><properties><property name="score" value="0.9852803746859232" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i68]" time="1.236"><properties><property name="score" value="0.40663886070251465" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i69]" time="1.426"><properties><property name="score" value="0.2960206667582194" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.29602067&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.29602067
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i70]" time="1.162"><properties><property name="score" value="0.33144797881444293" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i71]" time="1.215"><properties><property name="score" value="0.9922697345415751" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i72]" time="1.440"><properties><property name="score" value="0.9998179078102112" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i73]" time="1.510"><properties><property name="score" value="0.8928097635507584" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i74]" time="1.253"><properties><property name="score" value="0.32942086458206177" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i75]" time="0.958"><properties><property name="score" value="0.992289662361145" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i76]" time="1.211"><properties><property name="score" value="0.9902872840563456" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i77]" time="0.925"><properties><property name="score" value="0.9998202323913574" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i78]" time="1.066"><properties><property name="score" value="0.941917339960734" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i79]" time="1.276"><properties><property name="score" value="0.9998111526171366" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i80]" time="1.531"><properties><property name="score" value="0.9998150666554769" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i81]" time="0.810"><properties><property name="score" value="0.7725012302398682" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.77250123&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.77250123
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i82]" time="0.535"><properties><property name="score" value="0.8218924105167389" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i83]" time="0.426"><properties><property name="score" value="0.9994544386863708" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i84]" time="1.166"><properties><property name="score" value="0.9978791276613871" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i85]" time="1.314"><properties><property name="score" value="0.32142478227615356" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i86]" time="0.999"><properties><property name="score" value="0.9997696876525879" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i87]" time="0.651"><properties><property name="score" value="0.9960881769657135" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i88]" time="0.900"><properties><property name="score" value="0.9068448841571808" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i89]" time="0.827"><properties><property name="score" value="0.08880490064620972" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0888049&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0888049
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i90]" time="1.226"><properties><property name="score" value="0.41937051216761273" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i91]" time="1.050"><properties><property name="score" value="0.9986235797405243" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i92]" time="1.125"><properties><property name="score" value="0.8924668828646342" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i93]" time="1.498"><properties><property name="score" value="0.3655366500218709" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.36553665&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.36553665
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i94]" time="1.154"><properties><property name="score" value="0.9820608496665955" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i95]" time="1.418"><properties><property name="score" value="0.32370181878407794" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i96]" time="0.620"><properties><property name="score" value="0.7021366655826569" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i97]" time="0.833"><properties><property name="score" value="0.7859208881855011" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i98]" time="1.342"><properties><property name="score" value="0.9997673034667969" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i99]" time="0.964"><properties><property name="score" value="0.00537535548210144" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i100]" time="0.826"><properties><property name="score" value="0.16417846083641052" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.16417846&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.16417846
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i101]" time="1.276"><properties><property name="score" value="0.19342943032582602" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i102]" time="0.896"><properties><property name="score" value="0.9997978210449219" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i103]" time="1.046"><properties><property name="score" value="0.45603781938552856" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i104]" time="1.519"><properties><property name="score" value="0.9937559167544047" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i105]" time="1.637"><properties><property name="score" value="0.9997353553771973" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i106]" time="0.894"><properties><property name="score" value="0.9958354532718658" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i107]" time="1.459"><properties><property name="score" value="0.9980374574661255" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i108]" time="1.295"><properties><property name="score" value="0.9992041289806366" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i109]" time="0.754"><properties><property name="score" value="0.8020311295986176" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i110]" time="1.056"><properties><property name="score" value="0.9998000264167786" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i111]" time="1.508"><properties><property name="score" value="0.33311182260513306" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i112]" time="1.392"><properties><property name="score" value="0.9183404048283895" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i113]" time="0.931"><properties><property name="score" value="0.9109807908535004" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.91098079&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.91098079
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i114]" time="0.809"><properties><property name="score" value="0.9993641376495361" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i115]" time="1.184"><properties><property name="score" value="0.9963726798693339" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i116]" time="0.734"><properties><property name="score" value="0.9823826253414154" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i117]" time="1.144"><properties><property name="score" value="0.9674155712127686" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i118]" time="1.232"><properties><property name="score" value="0.9988779028256735" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i119]" time="0.929"><properties><property name="score" value="0.9653203189373016" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i120]" time="0.860"><properties><property name="score" value="3.7670135498046875e-05" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i121]" time="1.152"><properties><property name="score" value="0.22724034388860068" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i122]" time="0.338"><properties><property name="score" value="0.9660729169845581" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.96607292&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.96607292
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i123]" time="1.013"><properties><property name="score" value="0.34186528126398724" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i124]" time="0.860"><properties><property name="score" value="0.00121307373046875" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i125]" time="0.771"><properties><property name="score" value="0.08050402998924255" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.08050403&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.08050403
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i126]" time="1.343"><properties><property name="score" value="0.341347614924113" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i127]" time="1.094"><properties><property name="score" value="0.1968269944190979" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.19682699&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.19682699
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i128]" time="0.950"><properties><property name="score" value="0.9998016953468323" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i129]" time="0.952"><properties><property name="score" value="0.9998169541358948" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i130]" time="0.585"><properties><property name="score" value="0.2229829728603363" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.22298297&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.22298297
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i131]" time="0.996"><properties><property name="score" value="0.8400131265322367" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i132]" time="1.070"><properties><property name="score" value="0.9949625333150228" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i133]" time="1.235"><properties><property name="score" value="0.2446980873743693" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.24469809&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.24469809
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i134]" time="1.039"><properties><property name="score" value="0.7975748777389526" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i135]" time="0.671"><properties><property name="score" value="0.011171519756317139" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i136]" time="0.701"><properties><property name="score" value="0.9893593192100525" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i137]" time="0.336"><properties><property name="score" value="0.9987586736679077" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.99875867&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.99875867
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i138]" time="0.618"><properties><property name="score" value="0.9428516328334808" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i139]" time="0.819"><properties><property name="score" value="0.9544914066791534" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.95449141&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.95449141
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i140]" time="0.971"><properties><property name="score" value="0.9963037967681885" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i141]" time="0.937"><properties><property name="score" value="0.9993414282798767" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i142]" time="1.193"><properties><property name="score" value="0.373771866162618" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i143]" time="1.111"><properties><property name="score" value="0.8645265301068624" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i144]" time="1.119"><properties><property name="score" value="0.19636871417363486" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.19636871&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.19636871
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i145]" time="0.962"><properties><property name="score" value="0.8778989712397257" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i146]" time="0.905"><properties><property name="score" value="0.997542679309845" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.99754268&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.99754268
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i147]" time="0.783"><properties><property name="score" value="0.9527691304683685" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i148]" time="1.150"><properties><property name="score" value="0.30617515246073407" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i149]" time="1.035"><properties><property name="score" value="0.0026652514934539795" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00266525&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00266525
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i150]" time="1.206"><properties><property name="score" value="0.35932717720667523" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i151]" time="0.901"><properties><property name="score" value="0.9987233579158783" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i152]" time="1.106"><properties><property name="score" value="0.3838367462158203" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i153]" time="1.203"><properties><property name="score" value="0.26602703332901" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.26602703&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.26602703
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i154]" time="1.083"><properties><property name="score" value="0.9411871433258057" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i155]" time="1.126"><properties><property name="score" value="0.9833460052808126" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i156]" time="0.801"><properties><property name="score" value="0.10408371686935425" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.10408372&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.10408372
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i157]" time="0.872"><properties><property name="score" value="0.9998106956481934" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i158]" time="0.971"><properties><property name="score" value="0.9998233914375305" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i159]" time="0.942"><properties><property name="score" value="0.9626654982566833" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i160]" time="0.843"><properties><property name="score" value="0.997225284576416" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i161]" time="0.861"><properties><property name="score" value="0.7527625262737274" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.75276253&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.75276253
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i162]" time="1.180"><properties><property name="score" value="0.9726232091585795" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i163]" time="0.676"><properties><property name="score" value="0.8754819929599762" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i164]" time="0.655"><properties><property name="score" value="0.9980902075767517" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i165]" time="0.882"><properties><property name="score" value="0.9986684024333954" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i166]" time="1.140"><properties><property name="score" value="0.38990447918574017" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i167]" time="1.040"><properties><property name="score" value="0.22493608792622885" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.22493609&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.22493609
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i168]" time="0.532"><properties><property name="score" value="0.8811719417572021" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i169]" time="0.769"><properties><property name="score" value="0.9625748097896576" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i170]" time="0.878"><properties><property name="score" value="0.9997120499610901" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i171]" time="1.450"><properties><property name="score" value="0.994014839331309" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i172]" time="1.178"><properties><property name="score" value="0.9937088886896769" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i173]" time="0.905"><properties><property name="score" value="0.9997485280036926" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i174]" time="0.835"><properties><property name="score" value="0.9975769221782684" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.99757692&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.99757692
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i175]" time="1.136"><properties><property name="score" value="0.16934722661972046" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.16934723&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.16934723
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i176]" time="1.438"><properties><property name="score" value="0.33420048157374066" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i177]" time="0.461"><properties><property name="score" value="0.9997127652168274" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i178]" time="1.266"><properties><property name="score" value="0.4528254469235738" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i179]" time="1.185"><properties><property name="score" value="0.9658588767051697" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i180]" time="1.367"><properties><property name="score" value="0.3674633304278056" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i181]" time="0.896"><properties><property name="score" value="0.999805748462677" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i182]" time="1.091"><properties><property name="score" value="0.48963656028111774" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i183]" time="1.238"><properties><property name="score" value="0.9127458333969116" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i184]" time="1.193"><properties><property name="score" value="0.9434432586034139" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i185]" time="0.369"><properties><property name="score" value="0.9998165965080261" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i186]" time="1.323"><properties><property name="score" value="0.24821442365646362" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i187]" time="1.101"><properties><property name="score" value="0.3190438151359558" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.31904382&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.31904382
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i188]" time="0.765"><properties><property name="score" value="0.9996525943279266" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i189]" time="0.974"><properties><property name="score" value="0.9997535645961761" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i190]" time="1.130"><properties><property name="score" value="0.0003777742385864258" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00037777&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00037777
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i191]" time="1.037"><properties><property name="score" value="0.2969159483909607" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i192]" time="1.341"><properties><property name="score" value="0.31423717737197876" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.31423718&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.31423718
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i193]" time="1.028"><properties><property name="score" value="0.3207826415697734" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i194]" time="1.114"><properties><property name="score" value="0.20694303512573242" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.20694304&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.20694304
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i195]" time="0.952"><properties><property name="score" value="0.1638094186782837" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.16380942&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.16380942
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i196]" time="1.524"><properties><property name="score" value="0.8958070278167725" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.89580703&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.89580703
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i197]" time="0.821"><properties><property name="score" value="0.7919905483722687" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i198]" time="1.368"><properties><property name="score" value="0.2273817459742228" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.22738175&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.22738175
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i199]" time="1.349"><properties><property name="score" value="0.999819815158844" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i200]" time="0.654"><properties><property name="score" value="0.9358334839344025" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i201]" time="0.675"><properties><property name="score" value="0.9681703150272369" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i202]" time="1.024"><properties><property name="score" value="0.8811724583307902" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i203]" time="1.256"><properties><property name="score" value="0.9996959567070007" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i204]" time="1.267"><properties><property name="score" value="0.9988239407539368" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i205]" time="1.102"><properties><property name="score" value="0.9143550992012024" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i206]" time="1.484"><properties><property name="score" value="0.9998050332069397" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i207]" time="1.458"><properties><property name="score" value="0.9985008041063944" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i208]" time="0.876"><properties><property name="score" value="0.9997817873954773" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i209]" time="0.516"><properties><property name="score" value="0.9997778534889221" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i210]" time="1.744"><properties><property name="score" value="0.45390379428863525" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.45390379&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.45390379
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i211]" time="1.180"><properties><property name="score" value="0.9645174344380697" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.96451743&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.96451743
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i212]" time="0.697"><properties><property name="score" value="0.014139801263809204" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0141398&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.0141398
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i213]" time="1.074"><properties><property name="score" value="0.30586401621500653" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.30586402&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.30586402
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i214]" time="0.966"><properties><property name="score" value="0.0012382864952087402" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00123829&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00123829
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i215]" time="1.558"><properties><property name="score" value="0.7649497985839844" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i216]" time="0.551"><properties><property name="score" value="0.22665414214134216" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.22665414&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.22665414
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i217]" time="1.350"><properties><property name="score" value="0.2324406901995341" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.23244069&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.23244069
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i218]" time="0.949"><properties><property name="score" value="0.10933494567871094" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.10933495&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.10933495
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i219]" time="1.191"><properties><property name="score" value="0.9992026686668396" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i220]" time="0.730"><properties><property name="score" value="0.0010360479354858398" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i221]" time="1.223"><properties><property name="score" value="0.9997883439064026" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i222]" time="0.708"><properties><property name="score" value="0.8325699865818024" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.83256999&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.83256999
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i223]" time="1.074"><properties><property name="score" value="0.9552381634712219" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i224]" time="0.790"><properties><property name="score" value="0.8990232348442078" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i225]" time="1.173"><properties><property name="score" value="0.8596663673718771" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i226]" time="0.904"><properties><property name="score" value="0.9997937679290771" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i227]" time="0.858"><properties><property name="score" value="0.999798059463501" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i228]" time="1.294"><properties><property name="score" value="0.313859224319458" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i229]" time="1.396"><properties><property name="score" value="0.9998095234235128" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i230]" time="0.976"><properties><property name="score" value="0.8750312527020773" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i231]" time="1.349"><properties><property name="score" value="0.9990344444910685" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i232]" time="0.647"><properties><property name="score" value="0.9843467175960541" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i233]" time="0.992"><properties><property name="score" value="0.3243661125500997" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.32436611&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.32436611
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i234]" time="0.794"><properties><property name="score" value="0.9972749650478363" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i235]" time="1.452"><properties><property name="score" value="0.9897317886352539" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i236]" time="0.982"><properties><property name="score" value="0.9998087882995605" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i237]" time="1.258"><properties><property name="score" value="0.3772616386413574" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i238]" time="0.472"><properties><property name="score" value="0.9998264908790588" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i239]" time="0.826"><properties><property name="score" value="0.9434821009635925" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.9434821&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.9434821
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i240]" time="1.231"><properties><property name="score" value="0.2360661824544271" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.23606618&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.23606618
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i241]" time="1.009"><properties><property name="score" value="0.8543733557065328" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i242]" time="0.720"><properties><property name="score" value="0.12212443351745605" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i243]" time="0.860"><properties><property name="score" value="0.9998204112052917" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i244]" time="1.274"><properties><property name="score" value="0.278206209341685" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.27820621&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.27820621
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i245]" time="0.790"><properties><property name="score" value="0.24779346585273743" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i246]" time="1.269"><properties><property name="score" value="0.99347323179245" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i247]" time="1.430"><properties><property name="score" value="0.31648315985997516" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i248]" time="1.297"><properties><property name="score" value="0.9801744023958842" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i249]" time="0.434"><properties><property name="score" value="0.9997214674949646" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i250]" time="0.444"><properties><property name="score" value="0.9997923970222473" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i251]" time="0.822"><properties><property name="score" value="0.0014596283435821533" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i252]" time="0.916"><properties><property name="score" value="0.06244000792503357" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i253]" time="1.024"><properties><property name="score" value="0.21221359570821127" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i254]" time="0.709"><properties><property name="score" value="0.9779192507266998" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i255]" time="0.685"><properties><property name="score" value="0.021025359630584717" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i256]" time="0.912"><properties><property name="score" value="0.03339836001396179" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03339836&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03339836
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i257]" time="0.356"><properties><property name="score" value="0.997398853302002" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i258]" time="0.796"><properties><property name="score" value="0.033320844173431396" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i259]" time="1.216"><properties><property name="score" value="0.9986899892489115" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i260]" time="1.173"><properties><property name="score" value="0.9932873845100403" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i261]" time="0.814"><properties><property name="score" value="0.9997689127922058" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i262]" time="1.081"><properties><property name="score" value="0.4418433705965678" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i263]" time="1.184"><properties><property name="score" value="0.3033297657966614" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.30332977&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.30332977
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i264]" time="1.092"><properties><property name="score" value="0.2632368803024292" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.26323688&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.26323688
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i265]" time="0.952"><properties><property name="score" value="0.8458989858627319" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.84589899&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.84589899
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i266]" time="1.294"><properties><property name="score" value="0.9454140067100525" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.94541401&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.94541401
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i267]" time="0.635"><properties><property name="score" value="0.547631174325943" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i268]" time="1.257"><properties><property name="score" value="0.22925875584284464" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i269]" time="1.208"><properties><property name="score" value="0.9875039060910543" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i270]" time="1.288"><properties><property name="score" value="0.9907547434171041" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i271]" time="1.255"><properties><property name="score" value="0.39421987533569336" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i272]" time="0.930"><properties><property name="score" value="0.9903567731380463" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i273]" time="0.809"><properties><property name="score" value="0.10118627548217773" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i274]" time="0.869"><properties><property name="score" value="0.005864381790161133" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00586438&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00586438
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i275]" time="0.852"><properties><property name="score" value="0.9888796806335449" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i276]" time="0.806"><properties><property name="score" value="0.9273589551448822" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i277]" time="1.096"><properties><property name="score" value="0.9950051108996073" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i278]" time="0.484"><properties><property name="score" value="0.9995142221450806" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i279]" time="0.905"><properties><property name="score" value="0.9938617944717407" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i280]" time="0.844"><properties><property name="score" value="0.9998198747634888" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i281]" time="1.195"><properties><property name="score" value="0.9998160401980082" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i282]" time="0.837"><properties><property name="score" value="0.9998201727867126" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i283]" time="0.869"><properties><property name="score" value="3.841519355773926e-05" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i284]" time="1.067"><properties><property name="score" value="0.8664819598197937" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i285]" time="1.801"><properties><property name="score" value="0.9333625286817551" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.93336253&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.93336253
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i286]" time="0.888"><properties><property name="score" value="0.0700884759426117" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i287]" time="0.666"><properties><property name="score" value="0.9880577325820923" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.98805773&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.98805773
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i288]" time="1.404"><properties><property name="score" value="0.9894526998202006" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i289]" time="1.343"><properties><property name="score" value="0.938876211643219" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i290]" time="0.768"><properties><property name="score" value="0.9996740818023682" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.99967408&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.99967408
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i291]" time="1.050"><properties><property name="score" value="0.8390604257583618" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i292]" time="1.198"><properties><property name="score" value="0.801337738831838" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i293]" time="1.148"><properties><property name="score" value="0.9285288254419962" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i294]" time="0.916"><properties><property name="score" value="0.22359808286031088" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i295]" time="1.363"><properties><property name="score" value="0.9988283514976501" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i296]" time="0.924"><properties><property name="score" value="0.9161336421966553" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i297]" time="0.709"><properties><property name="score" value="0.06770780682563782" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i298]" time="1.089"><properties><property name="score" value="0.42239407698313397" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.42239408&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.42239408
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i299]" time="1.143"><properties><property name="score" value="0.9992963671684265" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i300]" time="1.103"><properties><property name="score" value="0.24533597628275552" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i301]" time="0.988"><properties><property name="score" value="0.07422128319740295" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i302]" time="1.516"><properties><property name="score" value="0.16596071422100067" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i303]" time="1.367"><properties><property name="score" value="0.9486927588780721" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i304]" time="1.182"><properties><property name="score" value="0.32159632444381714" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i305]" time="1.110"><properties><property name="score" value="0.32962314287821454" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.32962314&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.32962314
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i306]" time="1.244"><properties><property name="score" value="0.300148069858551" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.30014807&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.30014807
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i307]" time="1.594"><properties><property name="score" value="0.9909399002790451" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i308]" time="0.861"><properties><property name="score" value="0.998137891292572" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i309]" time="1.203"><properties><property name="score" value="0.33911128838857013" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i310]" time="1.149"><properties><property name="score" value="0.9570591052373251" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i311]" time="1.148"><properties><property name="score" value="0.12570178508758545" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.12570179&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.12570179
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i312]" time="1.108"><properties><property name="score" value="0.24458370606104532" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.24458371&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.24458371
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i313]" time="0.651"><properties><property name="score" value="0.9987660646438599" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i314]" time="0.934"><properties><property name="score" value="0.9040626883506775" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i315]" time="1.139"><properties><property name="score" value="0.9905946056048075" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.99059461&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.99059461
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i316]" time="1.065"><properties><property name="score" value="0.4891369740168254" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i317]" time="1.184"><properties><property name="score" value="0.9726180036862692" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i318]" time="0.740"><properties><property name="score" value="0.19890066981315613" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i319]" time="1.373"><properties><property name="score" value="0.24630429347356161" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.24630429&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.24630429
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i320]" time="1.007"><properties><property name="score" value="0.7290011246999105" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i321]" time="1.216"><properties><property name="score" value="0.9826884865760803" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i322]" time="0.728"><properties><property name="score" value="0.018715322017669678" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01871532&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.01871532
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i323]" time="0.913"><properties><property name="score" value="0.998868465423584" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i324]" time="1.126"><properties><property name="score" value="0.9950304627418518" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i325]" time="1.181"><properties><property name="score" value="0.9987582166989645" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i326]" time="1.176"><properties><property name="score" value="0.9997972051302592" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i327]" time="1.231"><properties><property name="score" value="0.9106310606002808" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i328]" time="1.379"><properties><property name="score" value="0.9997920393943787" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i329]" time="1.149"><properties><property name="score" value="0.9982194900512695" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i330]" time="1.223"><properties><property name="score" value="0.9980531136194865" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i331]" time="1.302"><properties><property name="score" value="0.9647947351137797" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i332]" time="0.973"><properties><property name="score" value="0.999536395072937" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i333]" time="1.323"><properties><property name="score" value="0.9812896251678467" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.98128963&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.98128963
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i334]" time="1.275"><properties><property name="score" value="0.29868970314661664" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i335]" time="1.271"><properties><property name="score" value="0.8961419065793356" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i336]" time="0.990"><properties><property name="score" value="0.15897536277770996" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i337]" time="0.771"><properties><property name="score" value="0.9980240166187286" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i338]" time="1.241"><properties><property name="score" value="0.8539116581281027" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i339]" time="1.093"><properties><property name="score" value="0.9994984070460001" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i340]" time="1.435"><properties><property name="score" value="0.333291490872701" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i341]" time="0.717"><properties><property name="score" value="0.8785853087902069" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i342]" time="0.961"><properties><property name="score" value="0.012039899826049805" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i343]" time="1.189"><properties><property name="score" value="0.33158332109451294" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.33158332&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.33158332
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i344]" time="0.963"><properties><property name="score" value="0.4536684552828471" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i345]" time="1.223"><properties><property name="score" value="0.2897172768910726" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.28971728&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.28971728
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i346]" time="0.805"><properties><property name="score" value="0.23302221298217773" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.23302221&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.23302221
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i347]" time="1.410"><properties><property name="score" value="0.9997149705886841" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i348]" time="1.302"><properties><property name="score" value="0.9994203646977743" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i349]" time="1.423"><properties><property name="score" value="0.9997955362002054" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i350]" time="1.555"><properties><property name="score" value="0.8914795716603597" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i351]" time="1.519"><properties><property name="score" value="0.17602312564849854" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.17602313&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.17602313
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i352]" time="0.687"><properties><property name="score" value="0.999663382768631" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i353]" time="0.812"><properties><property name="score" value="0.04285329580307007" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i354]" time="1.165"><properties><property name="score" value="0.9997973243395487" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i355]" time="0.959"><properties><property name="score" value="0.9998230934143066" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i356]" time="1.159"><properties><property name="score" value="0.17779409885406494" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.1777941&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.1777941
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i357]" time="1.161"><properties><property name="score" value="0.16271225611368814" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i358]" time="1.190"><properties><property name="score" value="0.9997569521268209" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i359]" time="1.195"><properties><property name="score" value="0.32566650708516437" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i360]" time="0.865"><properties><property name="score" value="0.9915338158607483" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i361]" time="0.771"><properties><property name="score" value="0.7997847497463226" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i362]" time="0.797"><properties><property name="score" value="0.9558180570602417" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.95581806&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.95581806
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i363]" time="0.906"><properties><property name="score" value="0.00013533234596252441" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00013533&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00013533
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i364]" time="0.988"><properties><property name="score" value="0.998431921005249" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i365]" time="0.839"><properties><property name="score" value="0.9998015761375427" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i366]" time="0.903"><properties><property name="score" value="0.9958670139312744" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i367]" time="1.101"><properties><property name="score" value="0.9837559660275778" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i368]" time="1.057"><properties><property name="score" value="0.8798910776774088" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i369]" time="0.904"><properties><property name="score" value="0.9740511476993561" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.97405115&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.97405115
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i370]" time="1.234"><properties><property name="score" value="0.9974707365036011" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i371]" time="1.230"><properties><property name="score" value="0.9998213450113932" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i372]" time="1.049"><properties><property name="score" value="0.9993818203608195" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i373]" time="0.817"><properties><property name="score" value="0.9997968077659607" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i374]" time="0.494"><properties><property name="score" value="0.9998303651809692" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.99983037&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.99983037
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i375]" time="1.363"><properties><property name="score" value="0.9994657238324484" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i376]" time="0.843"><properties><property name="score" value="0.9998241662979126" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i377]" time="0.771"><properties><property name="score" value="0.9994522929191589" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i378]" time="0.742"><properties><property name="score" value="0.9997837543487549" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i379]" time="0.812"><properties><property name="score" value="0.9998185038566589" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i380]" time="1.094"><properties><property name="score" value="0.9306139548619589" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i381]" time="0.649"><properties><property name="score" value="0.9875688552856445" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.98756886&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.98756886
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i382]" time="1.627"><properties><property name="score" value="0.46601293981075287" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i383]" time="0.825"><properties><property name="score" value="0.9981416165828705" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i384]" time="1.356"><properties><property name="score" value="0.860397736231486" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.86039774&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.86039774
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i385]" time="0.738"><properties><property name="score" value="0.032534003257751465" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i386]" time="0.903"><properties><property name="score" value="0.9726037979125977" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i387]" time="1.163"><properties><property name="score" value="0.8430514931678772" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.84305149&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.84305149
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i388]" time="1.253"><properties><property name="score" value="0.9415754874547323" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i389]" time="0.914"><properties><property name="score" value="0.8991599977016449" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i390]" time="1.350"><properties><property name="score" value="0.8820819854736328" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i391]" time="1.496"><properties><property name="score" value="0.8800054341554642" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i392]" time="0.879"><properties><property name="score" value="0.9783064126968384" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.97830641&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.97830641
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i393]" time="1.326"><properties><property name="score" value="0.3266701896985372" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i394]" time="0.979"><properties><property name="score" value="0.9998260140419006" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i395]" time="0.867"><properties><property name="score" value="0.04345899820327759" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i396]" time="1.068"><properties><property name="score" value="0.9852522810300192" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i397]" time="1.280"><properties><property name="score" value="0.9984065492947897" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i398]" time="1.422"><properties><property name="score" value="0.9350697795550028" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.93506978&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.93506978
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i399]" time="0.569"><properties><property name="score" value="0.9998249411582947" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i400]" time="1.033"><properties><property name="score" value="0.9902715881665548" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i401]" time="1.246"><properties><property name="score" value="0.9837998946507772" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.98379989&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.98379989
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i402]" time="1.097"><properties><property name="score" value="0.37364596128463745" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i403]" time="0.766"><properties><property name="score" value="0.005705893039703369" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00570589&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00570589
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i404]" time="1.282"><properties><property name="score" value="0.33193325996398926" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i405]" time="0.854"><properties><property name="score" value="0.9980986714363098" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i406]" time="0.889"><properties><property name="score" value="0.9998132586479187" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i407]" time="0.706"><properties><property name="score" value="0.8979074954986572" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i408]" time="1.046"><properties><property name="score" value="0.17821133136749268" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.17821133&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.17821133
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i409]" time="1.093"><properties><property name="score" value="0.9962007602055868" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i410]" time="1.376"><properties><property name="score" value="0.9998043179512024" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i411]" time="1.573"><properties><property name="score" value="0.5427448600530624" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i412]" time="1.128"><properties><property name="score" value="0.9716955423355103" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i413]" time="0.941"><properties><property name="score" value="0.9998222589492798" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i414]" time="1.058"><properties><property name="score" value="0.30805687109629315" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.30805687&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.30805687
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i415]" time="1.203"><properties><property name="score" value="0.4280824462572734" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i416]" time="1.378"><properties><property name="score" value="0.2786294221878052" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.27862942&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.27862942
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i417]" time="1.239"><properties><property name="score" value="0.46603743235270184" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i418]" time="1.177"><properties><property name="score" value="0.34561028083165485" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.34561028&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.34561028
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i419]" time="0.841"><properties><property name="score" value="0.9833200573921204" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i420]" time="1.176"><properties><property name="score" value="0.37242265542348224" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i421]" time="0.889"><properties><property name="score" value="0.9522815644741058" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i422]" time="0.733"><properties><property name="score" value="0.9670200049877167" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i423]" time="0.820"><properties><property name="score" value="0.9998239278793335" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i424]" time="1.213"><properties><property name="score" value="0.9108660022417704" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i425]" time="1.189"><properties><property name="score" value="0.2795597513516744" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i426]" time="0.574"><properties><property name="score" value="0.7719220519065857" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i427]" time="0.790"><properties><property name="score" value="0.00021317601203918457" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00021318&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00021318
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i428]" time="1.201"><properties><property name="score" value="0.9735872745513916" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i429]" time="0.661"><properties><property name="score" value="0.8226242065429688" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.82262421&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.82262421
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i430]" time="1.230"><properties><property name="score" value="0.9586420655250549" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i431]" time="0.779"><properties><property name="score" value="0.9998043775558472" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i432]" time="1.187"><properties><property name="score" value="0.9161777297655741" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i433]" time="1.288"><properties><property name="score" value="0.45823009808858234" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i434]" time="1.246"><properties><property name="score" value="0.9995513955752054" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i435]" time="0.563"><properties><property name="score" value="0.8459224104881287" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i436]" time="0.758"><properties><property name="score" value="0.9824471771717072" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.98244718&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.98244718
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i437]" time="1.436"><properties><property name="score" value="0.3096470634142558" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.30964706&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.30964706
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i438]" time="0.691"><properties><property name="score" value="0.9994584321975708" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i439]" time="0.669"><properties><property name="score" value="0.000558018684387207" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00055802&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00055802
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i440]" time="0.880"><properties><property name="score" value="0.9197102387746176" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.91971024&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.91971024
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i441]" time="0.931"><properties><property name="score" value="0.02130967378616333" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i442]" time="1.327"><properties><property name="score" value="0.2196908394495646" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i443]" time="0.820"><properties><property name="score" value="0.000525057315826416" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00052506&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00052506
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i444]" time="1.053"><properties><property name="score" value="0.9888309836387634" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i445]" time="1.587"><properties><property name="score" value="0.9610344171524048" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i446]" time="0.986"><properties><property name="score" value="0.8469152748584747" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i447]" time="0.739"><properties><property name="score" value="0.22933197021484375" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i448]" time="0.932"><properties><property name="score" value="0.6083444952964783" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i449]" time="0.728"><properties><property name="score" value="0.0035271644592285156" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i450]" time="1.253"><properties><property name="score" value="0.9995936155319214" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i451]" time="0.711"><properties><property name="score" value="0.053137898445129395" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i452]" time="0.847"><properties><property name="score" value="0.9998250305652618" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i453]" time="1.023"><properties><property name="score" value="0.2133470376332601" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i454]" time="0.767"><properties><property name="score" value="0.1352967917919159" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i455]" time="0.843"><properties><property name="score" value="0.9998075366020203" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i456]" time="1.169"><properties><property name="score" value="0.3325723210970561" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i457]" time="0.728"><properties><property name="score" value="0.058573514223098755" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05857351&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.05857351
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i458]" time="1.081"><properties><property name="score" value="0.00021454691886901855" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00021455&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00021455
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i459]" time="1.390"><properties><property name="score" value="0.8666579723358154" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.86665797&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.86665797
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i460]" time="1.164"><properties><property name="score" value="0.3328990936279297" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i461]" time="0.865"><properties><property name="score" value="0.007166147232055664" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i462]" time="1.532"><properties><property name="score" value="0.12201745808124542" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.12201746&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.12201746
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i463]" time="1.102"><properties><property name="score" value="0.1448967456817627" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.14489675&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.14489675
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i464]" time="1.079"><properties><property name="score" value="0.9470290740331014" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i465]" time="1.206"><properties><property name="score" value="0.2659171422322591" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i466]" time="1.059"><properties><property name="score" value="0.2916155258814494" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i467]" time="1.509"><properties><property name="score" value="0.2068381905555725" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.20683819&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.20683819
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i468]" time="1.079"><properties><property name="score" value="0.9201197822888693" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i469]" time="0.935"><properties><property name="score" value="0.004412949085235596" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00441295&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.00441295
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i470]" time="1.060"><properties><property name="score" value="0.9909900724887848" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i471]" time="0.898"><properties><property name="score" value="0.9865541756153107" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i472]" time="0.714"><properties><property name="score" value="0.9925233721733093" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i473]" time="0.666"><properties><property name="score" value="0.9957118034362793" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i474]" time="0.701"><properties><property name="score" value="0.9521922171115875" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i475]" time="0.895"><properties><property name="score" value="0.9283293783664703" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i476]" time="1.361"><properties><property name="score" value="0.9194203615188599" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i477]" time="0.866"><properties><property name="score" value="0.021271467208862305" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i478]" time="1.082"><properties><property name="score" value="0.8782995541890463" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i479]" time="1.177"><properties><property name="score" value="0.9982547163963318" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i480]" time="1.334"><properties><property name="score" value="0.9997851252555847" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i481]" time="0.769"><properties><property name="score" value="0.9970488250255585" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i482]" time="1.358"><properties><property name="score" value="0.9998069008191427" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i483]" time="0.857"><properties><property name="score" value="0.9997877478599548" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i484]" time="0.762"><properties><property name="score" value="0.959233820438385" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i485]" time="1.303"><properties><property name="score" value="0.9998240272204081" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i486]" time="1.213"><properties><property name="score" value="0.9997357328732809" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i487]" time="0.733"><properties><property name="score" value="0.005379319190979004" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i488]" time="0.826"><properties><property name="score" value="0.16408127546310425" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i489]" time="0.876"><properties><property name="score" value="0.018272846937179565" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i490]" time="1.064"><properties><property name="score" value="0.9122047225634257" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.91220472&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.91220472
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i491]" time="1.134"><properties><property name="score" value="0.8099290529886881" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.80992905&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.80992905
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i492]" time="1.030"><properties><property name="score" value="0.3112519184748332" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.31125192&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.31125192
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i493]" time="0.819"><properties><property name="score" value="0.9997879862785339" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i494]" time="1.100"><properties><property name="score" value="0.999708871046702" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i495]" time="0.382"><properties><property name="score" value="0.9998096823692322" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i496]" time="0.827"><properties><property name="score" value="0.0315193235874176" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03151932&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.03151932
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i497]" time="0.932"><properties><property name="score" value="0.05561530590057373" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i498]" time="1.409"><properties><property name="score" value="0.9944070130586624" /></properties></testcase><testcase classname="test_roberta_detect" name="test_chatgptnews_jsonl[i499]" time="1.361"><properties><property name="score" value="0.3267669479052226" /></properties><failure message="AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.32676695&#10;assert 'Human' == 'AI'&#10;  - AI&#10;  + Human">E   AssertionError: samples/news.jsonl is a AI-generated sample, misclassified as human-generated with confidence 0.32676695
+    assert 'Human' == 'AI'
+      - AI
+      + Human</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i0]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i1]" time="0.003"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i2]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i3]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i4]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i5]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i6]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i7]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i8]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i9]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i10]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i11]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i12]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i13]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i14]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i15]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i16]" time="0.003"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i17]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i18]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i19]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i20]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i21]" time="0.003"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i22]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i23]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i24]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i25]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i26]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i27]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i28]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i29]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i30]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i31]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i32]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i33]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i34]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i35]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i36]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i37]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i38]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i39]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i40]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i41]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i42]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i43]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i44]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i45]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i46]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i47]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i48]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i49]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i50]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i51]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i52]" time="0.003"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i53]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i54]" time="0.003"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i55]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i56]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i57]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i58]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i59]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i60]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i61]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i62]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i63]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i64]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i65]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i66]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i67]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i68]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i69]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i70]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i71]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i72]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i73]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i74]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i75]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i76]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i77]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i78]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i79]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i80]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i81]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i82]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i83]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i84]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i85]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i86]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i87]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i88]" time="0.003"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i89]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i90]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i91]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i92]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i93]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i94]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i95]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i96]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i97]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i98]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i99]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i100]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i101]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i102]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i103]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i104]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i105]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i106]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i107]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i108]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i109]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i110]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i111]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i112]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i113]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i114]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i115]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i116]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i117]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i118]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i119]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i120]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i121]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i122]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i123]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i124]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i125]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i126]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i127]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i128]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i129]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i130]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i131]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i132]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i133]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i134]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i135]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i136]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i137]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i138]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i139]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i140]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i141]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i142]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i143]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i144]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i145]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i146]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i147]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i148]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i149]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i150]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i151]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i152]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i153]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i154]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i155]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i156]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i157]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i158]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i159]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i160]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i161]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i162]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i163]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i164]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i165]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i166]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i167]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i168]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i169]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i170]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i171]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i172]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i173]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i174]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i175]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i176]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i177]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i178]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i179]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i180]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i181]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i182]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i183]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i184]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i185]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i186]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i187]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i188]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i189]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i190]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i191]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i192]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i193]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i194]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i195]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i196]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i197]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i198]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i199]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i200]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i201]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i202]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i203]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i204]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i205]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i206]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i207]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i208]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i209]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i210]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i211]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i212]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i213]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i214]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i215]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i216]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i217]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i218]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i219]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i220]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i221]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i222]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i223]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i224]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i225]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i226]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i227]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i228]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i229]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i230]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i231]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i232]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i233]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i234]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i235]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i236]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i237]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i238]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i239]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i240]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i241]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i242]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i243]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i244]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i245]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i246]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i247]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i248]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i249]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i250]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i251]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i252]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i253]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i254]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i255]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i256]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i257]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i258]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i259]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i260]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i261]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i262]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i263]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i264]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i265]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i266]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i267]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i268]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i269]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i270]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i271]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i272]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i273]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i274]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i275]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i276]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i277]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i278]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i279]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i280]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i281]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i282]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i283]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i284]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i285]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i286]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i287]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i288]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i289]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i290]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i291]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i292]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i293]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i294]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i295]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i296]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i297]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i298]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i299]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i300]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i301]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i302]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i303]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i304]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i305]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i306]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i307]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i308]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i309]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i310]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i311]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i312]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i313]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i314]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i315]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i316]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i317]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i318]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i319]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i320]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i321]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i322]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i323]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i324]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i325]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i326]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i327]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i328]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i329]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i330]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i331]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i332]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i333]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i334]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i335]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i336]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i337]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i338]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i339]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i340]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i341]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i342]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i343]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i344]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i345]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i346]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i347]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i348]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i349]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i350]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i351]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i352]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i353]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i354]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i355]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i356]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i357]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i358]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i359]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i360]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i361]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i362]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i363]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i364]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i365]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i366]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i367]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i368]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i369]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i370]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i371]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i372]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i373]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i374]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i375]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i376]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i377]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i378]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i379]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i380]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i381]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i382]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i383]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i384]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i385]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i386]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i387]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i388]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i389]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i390]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i391]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i392]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i393]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i394]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i395]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i396]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i397]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i398]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i399]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i400]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i401]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i402]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i403]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i404]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i405]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i406]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i407]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i408]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i409]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i410]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i411]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i412]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i413]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i414]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i415]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i416]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i417]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i418]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i419]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i420]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i421]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i422]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i423]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i424]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i425]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i426]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i427]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i428]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i429]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i430]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i431]" time="0.003"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i432]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i433]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i434]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i435]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i436]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i437]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i438]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i439]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i440]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i441]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i442]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i443]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i444]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i445]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i446]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i447]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i448]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i449]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i450]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i451]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i452]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i453]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i454]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i455]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i456]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i457]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i458]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i459]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i460]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i461]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i462]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i463]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i464]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i465]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i466]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i467]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i468]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i469]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i470]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i471]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i472]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i473]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i474]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i475]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i476]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i477]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i478]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i479]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i480]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i481]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i482]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i483]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i484]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i485]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i486]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i487]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i488]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i489]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i490]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i491]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i492]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i493]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i494]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i495]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i496]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i497]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i498]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_human_jsonl[i499]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i0]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i1]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i2]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i3]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i4]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i5]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i6]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i7]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i8]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i9]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i10]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i11]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i12]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i13]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i14]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i15]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i16]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i17]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i18]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i19]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i20]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i21]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i22]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i23]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i24]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i25]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i26]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i27]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i28]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i29]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i30]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i31]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i32]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i33]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i34]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i35]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i36]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i37]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i38]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i39]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i40]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i41]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i42]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i43]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i44]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i45]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i46]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i47]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i48]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i49]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i50]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i51]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i52]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i53]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i54]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i55]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i56]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i57]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i58]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i59]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i60]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i61]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i62]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i63]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i64]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i65]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i66]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i67]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i68]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i69]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i70]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i71]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i72]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i73]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i74]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i75]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i76]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i77]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i78]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i79]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i80]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i81]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i82]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i83]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i84]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i85]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i86]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i87]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i88]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i89]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i90]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i91]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i92]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i93]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i94]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i95]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i96]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i97]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i98]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i99]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i100]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i101]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i102]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i103]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i104]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i105]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i106]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i107]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i108]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i109]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i110]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i111]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i112]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i113]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i114]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i115]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i116]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i117]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i118]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i119]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i120]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i121]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i122]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i123]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i124]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i125]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i126]" time="0.003"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i127]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i128]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i129]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i130]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i131]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i132]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i133]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i134]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i135]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i136]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i137]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i138]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i139]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i140]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i141]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i142]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i143]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i144]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i145]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i146]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i147]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i148]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i149]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i150]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i151]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i152]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i153]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i154]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i155]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i156]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i157]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i158]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i159]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i160]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i161]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i162]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i163]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i164]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i165]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i166]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i167]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i168]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i169]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i170]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i171]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i172]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i173]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i174]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i175]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i176]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i177]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i178]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i179]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i180]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i181]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i182]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i183]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i184]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i185]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i186]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i187]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i188]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i189]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i190]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i191]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i192]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i193]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i194]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i195]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i196]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i197]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i198]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i199]" time="0.003"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i200]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i201]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i202]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i203]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i204]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i205]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i206]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i207]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i208]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i209]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i210]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i211]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i212]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i213]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i214]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i215]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i216]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i217]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i218]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i219]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i220]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i221]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i222]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i223]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i224]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i225]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i226]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i227]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i228]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i229]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i230]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i231]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i232]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i233]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i234]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i235]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i236]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i237]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i238]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i239]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i240]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i241]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i242]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i243]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i244]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i245]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i246]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i247]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i248]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i249]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i250]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i251]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i252]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i253]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i254]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i255]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i256]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i257]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i258]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i259]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i260]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i261]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i262]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i263]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i264]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i265]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i266]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i267]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i268]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i269]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i270]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i271]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i272]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i273]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i274]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i275]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i276]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i277]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i278]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i279]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i280]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i281]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i282]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i283]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i284]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i285]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i286]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i287]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i288]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i289]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i290]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i291]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i292]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i293]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i294]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i295]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i296]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i297]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i298]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i299]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i300]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i301]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i302]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i303]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i304]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i305]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i306]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i307]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i308]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i309]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i310]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i311]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i312]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i313]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i314]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i315]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i316]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i317]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i318]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i319]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i320]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i321]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i322]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i323]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i324]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i325]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i326]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i327]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i328]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i329]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i330]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i331]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i332]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i333]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i334]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i335]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i336]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i337]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i338]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i339]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i340]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i341]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i342]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i343]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i344]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i345]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i346]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i347]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i348]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i349]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i350]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i351]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i352]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i353]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i354]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i355]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i356]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i357]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i358]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i359]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i360]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i361]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i362]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i363]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i364]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i365]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i366]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i367]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i368]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i369]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i370]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i371]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i372]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i373]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i374]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i375]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i376]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i377]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i378]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i379]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i380]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i381]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i382]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i383]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i384]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i385]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i386]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i387]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i388]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i389]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i390]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i391]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i392]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i393]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i394]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i395]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i396]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i397]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i398]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i399]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i400]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i401]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i402]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i403]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i404]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i405]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i406]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i407]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i408]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i409]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i410]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i411]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i412]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i413]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i414]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i415]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i416]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i417]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i418]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i419]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i420]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i421]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i422]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i423]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i424]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i425]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i426]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i427]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i428]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i429]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i430]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i431]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i432]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i433]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i434]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i435]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i436]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i437]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i438]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i439]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i440]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i441]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i442]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i443]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i444]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i445]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i446]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i447]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i448]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i449]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i450]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i451]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i452]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i453]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i454]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i455]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i456]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i457]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i458]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i459]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i460]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i461]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i462]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i463]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i464]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i465]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i466]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i467]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i468]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i469]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i470]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i471]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i472]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i473]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i474]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i475]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i476]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i477]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i478]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i479]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i480]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i481]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i482]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i483]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i484]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i485]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i486]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i487]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i488]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i489]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i490]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i491]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i492]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i493]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i494]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i495]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i496]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i497]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i498]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_generation_jsonl[i499]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i0]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i1]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i2]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i3]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i4]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i5]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i6]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i7]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i8]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i9]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i10]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i11]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i12]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i13]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i14]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i15]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i16]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i17]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i18]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i19]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i20]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i21]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i22]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i23]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i24]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i25]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i26]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i27]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i28]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i29]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i30]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i31]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i32]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i33]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i34]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i35]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i36]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i37]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i38]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i39]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i40]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i41]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i42]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i43]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i44]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i45]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i46]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i47]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i48]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i49]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i50]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i51]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i52]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i53]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i54]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i55]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i56]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i57]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i58]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i59]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i60]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i61]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i62]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i63]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i64]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i65]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i66]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i67]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i68]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i69]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i70]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i71]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i72]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i73]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i74]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i75]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i76]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i77]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i78]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i79]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i80]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i81]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i82]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i83]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i84]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i85]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i86]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i87]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i88]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i89]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i90]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i91]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i92]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i93]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i94]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i95]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i96]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i97]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i98]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i99]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i100]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i101]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i102]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i103]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i104]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i105]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i106]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i107]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i108]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i109]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i110]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i111]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i112]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i113]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i114]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i115]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i116]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i117]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i118]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i119]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i120]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i121]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i122]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i123]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i124]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i125]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i126]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i127]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i128]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i129]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i130]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i131]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i132]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i133]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i134]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i135]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i136]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i137]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i138]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i139]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i140]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i141]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i142]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i143]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i144]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i145]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i146]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i147]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i148]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i149]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i150]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i151]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i152]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i153]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i154]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i155]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i156]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i157]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i158]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i159]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i160]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i161]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i162]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i163]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i164]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i165]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i166]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i167]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i168]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i169]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i170]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i171]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i172]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i173]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i174]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i175]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i176]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i177]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i178]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i179]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i180]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i181]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i182]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i183]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i184]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i185]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i186]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i187]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i188]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i189]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i190]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i191]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i192]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i193]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i194]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i195]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i196]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i197]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i198]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i199]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i200]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i201]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i202]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i203]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i204]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i205]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i206]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i207]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i208]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i209]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i210]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i211]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i212]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i213]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i214]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i215]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i216]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i217]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i218]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i219]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i220]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i221]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i222]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i223]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i224]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i225]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i226]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i227]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i228]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i229]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i230]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i231]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i232]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i233]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i234]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i235]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i236]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i237]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i238]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i239]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i240]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i241]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i242]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i243]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i244]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i245]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i246]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i247]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i248]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i249]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i250]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i251]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i252]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i253]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i254]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i255]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i256]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i257]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i258]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i259]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i260]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i261]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i262]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i263]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i264]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i265]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i266]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i267]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i268]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i269]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i270]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i271]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i272]" time="0.003"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i273]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i274]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i275]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i276]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i277]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i278]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i279]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i280]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i281]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i282]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i283]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i284]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i285]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i286]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i287]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i288]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i289]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i290]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i291]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i292]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i293]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i294]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i295]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i296]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i297]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i298]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i299]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i300]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i301]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i302]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i303]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i304]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i305]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i306]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i307]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i308]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i309]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i310]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i311]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i312]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i313]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i314]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i315]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i316]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i317]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i318]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i319]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i320]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i321]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i322]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i323]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i324]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i325]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i326]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i327]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i328]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i329]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i330]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i331]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i332]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i333]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i334]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i335]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i336]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i337]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i338]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i339]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i340]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i341]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i342]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i343]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i344]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i345]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i346]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i347]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i348]" time="0.003"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i349]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i350]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i351]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i352]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i353]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i354]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i355]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i356]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i357]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i358]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i359]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i360]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i361]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i362]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i363]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i364]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i365]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i366]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i367]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i368]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i369]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i370]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i371]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i372]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i373]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i374]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i375]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i376]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i377]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i378]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i379]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i380]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i381]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i382]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i383]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i384]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i385]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i386]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i387]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i388]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i389]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i390]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i391]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i392]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i393]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i394]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i395]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i396]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i397]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i398]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i399]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i400]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i401]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i402]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i403]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i404]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i405]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i406]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i407]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i408]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i409]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i410]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i411]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i412]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i413]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i414]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i415]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i416]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i417]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i418]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i419]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i420]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i421]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i422]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i423]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i424]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i425]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i426]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i427]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i428]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i429]" time="0.003"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i430]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i431]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i432]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i433]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i434]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i435]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i436]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i437]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i438]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i439]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i440]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i441]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i442]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i443]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i444]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i445]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i446]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i447]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i448]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i449]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i450]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i451]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i452]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i453]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i454]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i455]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i456]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i457]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i458]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i459]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i460]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i461]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i462]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i463]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i464]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i465]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i466]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i467]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i468]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i469]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i470]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i471]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i472]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i473]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i474]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i475]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i476]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i477]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i478]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i479]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i480]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i481]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i482]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i483]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i484]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i485]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i486]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i487]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i488]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i489]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i490]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i491]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i492]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i493]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i494]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i495]" time="0.002"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i496]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i497]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i498]" time="0.001"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase><testcase classname="test_roberta_detect" name="test_cheat_polish_jsonl[i499]" time="0.064"><failure message="NameError: name 'FUZZINESS' is not defined">E   NameError: name 'FUZZINESS' is not defined</failure></testcase></testsuite></testsuites>
\ No newline at end of file