{"meta":{"page":1,"per_page":50,"max_per_page":100,"total":570,"total_is_capped":false,"direct_labels_cover":570,"predictions_cover":570,"direct_label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline (scores rank; they never assert a category)","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12"},"query_hash":"3cc3e5abc8f5","filters":{"category":"research_integrity"}},"results":[{"id":"W2587319655","doi":"10.1186/s12961-016-0164-6","title":"The concept of ‘vulnerability’ in research ethics: an in-depth analysis of policies and guidelines","year":2017,"lang":"en","type":"article","venue":"Health Research Policy and Systems","topic":"Ethics in Clinical Research","field":"Medicine","cited_by":276,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"McGill University; Université de Montréal; McGill University Health Centre; Montreal Clinical Research Institute","funders":"Fonds de Recherche du Québec - Santé; Canadian Institutes of Health Research; Faculty of Medicine, McGill University; McGill University","keywords":"Vulnerability (computing); Research ethics; Stakeholder; Normative; Vulnerability assessment; Harm; Sociology; Engineering ethics; Political science; Public relations; Psychology; Social psychology; Law; Psychological resilience; Computer science; Computer security","retraction":null,"screen_n_in":null,"score":{"opus":0.9560629307251761,"gpt":0.7983703746638118,"spread":0.1576925560613643,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","sts","research_integrity"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.2130552,0.0001265838,0.001026681,0.002194667,0.001201585,0.000185394,0.0007283159,0.0005216478,0.000003761215],"category_scores_gemma":[0.3169491,0.00008418197,0.00007306251,0.001707158,0.006684565,0.0001243095,0.0006618639,0.005855765,9.547331e-7],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003091175,"about_ca_system_score_gemma":0.005884327,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.345382,"about_ca_topic_score_gemma":0.1358867,"domain_scores_codex":[0.9805462,0.01093852,0.002153739,0.0006461868,0.004251175,0.001464193],"domain_scores_gemma":[0.8874477,0.1000691,0.0005042203,0.003157314,0.007748589,0.001073046],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"observational","study_design_scores_codex":[0.0008109713,0.0003803282,0.4641863,0.005849359,0.0002369327,0.00001894074,0.03087243,0.00006603987,0.0004899189,0.4794055,0.0006091522,0.01707421],"study_design_scores_gemma":[0.001991933,0.002466258,0.9166916,0.003584989,0.0000160044,0.000008214905,0.03214162,0.009360497,0.0001645103,0.03005264,0.003385122,0.0001366344],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8505239,0.009925827,0.000003930452,0.1354179,0.00004542227,0.001435732,0.00004590037,0.000006274169,0.002595135],"genre_scores_gemma":[0.9676583,0.03084426,0.00004761546,0.0001872034,0.0002283815,0.00008234623,0.000004283962,0.00001563226,0.0009319663],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4525054,"threshold_uncertainty_score":0.9997514,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["research_integrity"],"domain":null,"study_design":"qualitative","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch"],"domain":"methods","study_design":"qualitative","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W2023640666","doi":"10.1177/1476750307083720","title":"Reciprocity","year":2008,"lang":"en","type":"article","venue":"Action Research","topic":"Community Health and Development","field":"Health Professions","cited_by":274,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":true},"ca_institutions":"University of Toronto; York University","funders":"Health Canada","keywords":"Reciprocity (cultural anthropology); Sociology; Participatory action research; Reciprocal; Action research; Citizen journalism; Public relations; Engineering ethics; Action (physics); Political science; Social science; Pedagogy; Law","retraction":null,"screen_n_in":null,"score":{"opus":0.908520586474894,"gpt":0.730495074282825,"spread":0.178025512192069,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["sts","insufficient_payload"],"consensus_categories":["insufficient_payload"],"category_scores_codex":[0.00270143,0.00004659672,0.00008757771,0.0001882548,0.004276739,0.000002609545,0.0001510636,0.0001170903,0.002558912],"category_scores_gemma":[0.0008892317,0.00004188497,0.00001763206,0.0004724428,0.00008229344,0.00008393534,0.000169429,0.001864092,0.005192222],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003819389,"about_ca_system_score_gemma":0.001205579,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002135283,"about_ca_topic_score_gemma":0.0006325569,"domain_scores_codex":[0.9971705,0.001400181,0.000222887,0.0001265641,0.000506538,0.0005732956],"domain_scores_gemma":[0.997729,0.001226829,0.00003337441,0.0003134912,0.0004786615,0.0002186312],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0002374109,0.0001377511,0.1725577,0.0002325487,0.00000986416,0.0000173793,0.01117717,5.48254e-7,0.0003870522,0.002080782,0.7776242,0.03553754],"study_design_scores_gemma":[0.0002836552,0.00003890921,0.2507353,0.00002362169,2.8961e-7,0.000003766741,0.003041156,0.00001312504,0.0000758715,0.0008611671,0.7448793,0.00004376666],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.835637,0.00005605267,0.00007452022,0.006985358,0.000426912,0.000629092,0.000001302086,0.0001037744,0.156086],"genre_scores_gemma":[0.9536966,0.0008744555,0.0005384698,0.001441742,0.0002811041,0.0002744031,0.00000986085,0.00001101327,0.04287236],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.1180596,"threshold_uncertainty_score":0.9983529,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"methods","study_design":"qualitative","genre":"empirical","about_ca_system":false,"about_ca_topic":true,"confidence":"high"},{"model":"gpt","categories":["metaresearch"],"domain":"methods","study_design":"qualitative","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W2082951106","doi":"10.4103/0019-5545.82529","title":"The MMR vaccine and autism: Sensation, refutation, retraction, and fraud","year":2011,"lang":"en","type":"article","venue":"Indian Journal of Psychiatry","topic":"Vaccine Coverage and Hesitancy","field":"Social Sciences","cited_by":258,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"","keywords":"MMR vaccine; Autism; Measles-Mumps-Rubella Vaccine; Publicity; Measles; Medicine; Misconduct; Vaccination; Rubella; Rubella vaccine; Psychiatry; Pediatrics; Psychology; Family medicine; Virology; Political science; Law","retraction":null,"screen_n_in":null,"score":{"opus":0.0139472442936847,"gpt":0.2683230300903904,"spread":0.2543757857967057,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008828377,0.00006499394,0.00009939287,0.00008475652,0.000756285,0.00008235766,0.0001044219,0.00008738601,0.000102789],"category_scores_gemma":[0.0001358401,0.00004777081,0.0000348173,0.0001542185,0.00002655878,0.0003887423,0.000007909307,0.0002314938,0.000005272272],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002064441,"about_ca_system_score_gemma":0.0002365076,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00009314701,"about_ca_topic_score_gemma":0.001575481,"domain_scores_codex":[0.9991543,0.0001162709,0.0003167477,0.00008005271,0.0001898079,0.0001428621],"domain_scores_gemma":[0.9992158,0.00007151265,0.000344033,0.00007971541,0.0001569731,0.0001319685],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0001024862,0.00005063811,0.8470428,0.00001589104,0.00008446111,0.0000368146,0.04136388,4.017197e-7,0.00001393165,0.06054829,0.03329627,0.01744416],"study_design_scores_gemma":[0.0004461016,0.0001210994,0.8600401,0.00004117297,0.00003481353,0.0001794881,0.008314361,8.114359e-7,0.000009298678,0.1074425,0.02328477,0.00008549771],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9566888,0.004937846,0.0001706278,0.03181776,0.002172454,0.00008969598,0.000001328629,0.00001170403,0.004109787],"genre_scores_gemma":[0.9949922,0.002110131,0.001489458,0.0003137605,0.0007653332,5.965699e-7,2.311516e-7,0.000006994235,0.000321322],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.04689417,"threshold_uncertainty_score":0.5816809,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["research_integrity"],"domain":null,"study_design":"theoretical_or_conceptual","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["research_integrity"],"domain":null,"study_design":"design_other","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W1987707043","doi":"10.1136/bmj.h210","title":"Firm action needed on predatory journals","year":2015,"lang":"en","type":"editorial","venue":"BMJ","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":243,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"","keywords":"Action (physics); Business; Public relations; Law and economics; Economics; Political science","retraction":null,"screen_n_in":null,"score":{"opus":0.7969926508729327,"gpt":0.6778639477958661,"spread":0.1191287030770666,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","scholarly_communication","research_integrity","insufficient_payload"],"consensus_categories":["metaresearch","bibliometrics","insufficient_payload"],"category_scores_codex":[0.07453775,0.0003023434,0.0006567832,0.08824164,0.0002552495,0.005935813,0.00401319,0.001299479,0.00191309],"category_scores_gemma":[0.3744678,0.0002046667,0.0003103461,0.1151029,0.0001050054,0.000630898,0.0007479962,0.001859084,0.005137558],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005909696,"about_ca_system_score_gemma":0.002213753,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001251702,"about_ca_topic_score_gemma":0.00002060145,"domain_scores_codex":[0.9520504,0.0006182799,0.0012438,0.001019096,0.04437969,0.0006887168],"domain_scores_gemma":[0.9646193,0.01673616,0.001270443,0.001828136,0.01453772,0.001008263],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00007796607,0.00007901764,0.0000950033,0.000006105177,0.00002103017,0.00002432573,0.00002812518,0.00001827629,0.00001642444,0.000005224414,0.9534311,0.04619735],"study_design_scores_gemma":[0.0004521565,0.0002922913,0.0004840363,0.00004061568,0.000007792441,0.000002787902,0.0001961423,0.000192195,0.00005272311,0.001847212,0.9961783,0.0002537403],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"editorial","genre_gemma":"editorial","genre_scores_codex":[0.001093037,0.001137071,0.0001123373,0.001476189,0.9631559,0.0004378086,0.0003500536,0.00004000721,0.03219762],"genre_scores_gemma":[0.001018158,0.0005184346,0.0001356163,0.0001064794,0.9289952,0.00003192541,0.00005132973,0.00003637673,0.0691065],"genre_candidate":"editorial","genre_consensus":"editorial","teacher_disagreement_score":0.2999301,"threshold_uncertainty_score":0.999997,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","scholarly_communication"],"domain":"evaluation","study_design":"not_applicable","genre":"editorial","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","research_integrity","scholarly_communication"],"domain":"methods","study_design":"not_applicable","genre":"editorial","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W2120841701","doi":"10.1371/journal.pbio.0060073","title":"Research Ethics Recommendations for Whole-Genome Research: Consensus Statement","year":2008,"lang":"en","type":"article","venue":"PLoS Biology","topic":"Ethics in Clinical Research","field":"Medicine","cited_by":241,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Memorial University of Newfoundland; Ontario Genomics; Canadian Institutes of Health Research; University of British Columbia; University of Toronto; Université de Montréal; University of Alberta","funders":"Institute of Genetics; Ontario Genomics; Canadian Institutes of Health Research; Ontario Genomics Institute; National Institutes of Health; Genome Alberta","keywords":"Biology; Statement (logic); Research ethics; Genome; Engineering ethics; Computational biology; MEDLINE; Genetics; Biotechnology; Gene; Political science; Law","retraction":null,"screen_n_in":null,"score":{"opus":0.9590869121845906,"gpt":0.7257237591152245,"spread":0.2333631530693662,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","sts","research_integrity","insufficient_payload"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.0383897,0.0001391018,0.0004150032,0.0006325558,0.001240395,0.00001961539,0.0004730258,0.001055081,0.0004927726],"category_scores_gemma":[0.1138889,0.0001179223,0.0001102492,0.0007503588,0.003609186,0.00002355006,0.0005150395,0.00965144,0.0007802548],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003839539,"about_ca_system_score_gemma":0.003062065,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001860952,"about_ca_topic_score_gemma":0.0002853023,"domain_scores_codex":[0.9925839,0.002983437,0.0008338369,0.0008049169,0.001431679,0.001362255],"domain_scores_gemma":[0.8922917,0.09969997,0.00007783875,0.0009647698,0.006504261,0.0004614786],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.004039606,0.004903358,0.01251352,0.002023085,0.001105505,0.0003563423,0.01024649,0.000005323096,0.3065028,0.4048087,0.2502012,0.003294102],"study_design_scores_gemma":[0.002400459,0.004789475,0.00135455,0.0002085566,0.00002160366,0.00004770132,0.00200961,0.0001320967,0.003631822,0.1720754,0.8131369,0.0001917628],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"commentary","genre_gemma":"empirical","genre_scores_codex":[0.2227045,0.0007242655,0.0004017035,0.7549924,0.0003597194,0.00355068,0.0007011633,0.0001322578,0.01643336],"genre_scores_gemma":[0.9288422,0.007686749,0.02347495,0.003649188,0.001028508,0.001157405,0.001100517,0.00009793099,0.03296247],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.7513431,"threshold_uncertainty_score":0.9999977,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["research_integrity"],"domain":null,"study_design":"not_applicable","genre":"other","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch"],"domain":"methods","study_design":"not_applicable","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"split"},{"id":"W2765089630","doi":"10.1089/trgh.2017.0012","title":"Guidance and Ethical Considerations for Undertaking Transgender Health Research and Institutional Review Boards Adjudicating this Research","year":2017,"lang":"en","type":"review","venue":"Transgender Health","topic":"LGBTQ Health, Identity, and Policy","field":"Psychology","cited_by":237,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Youth Services Bureau of Ottawa","funders":"","keywords":"Transgender; Confidentiality; Temptation; Informed consent; Guard (computer science); Framing (construction); Public relations; Institutional review board; Research ethics; Psychology; Political science; Engineering ethics; Medical education; Medicine; Social psychology; Law; Alternative medicine; Psychiatry; Computer science; Engineering","retraction":null,"screen_n_in":null,"score":{"opus":0.9016664340896657,"gpt":0.705548719146325,"spread":0.1961177149433407,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","metaepi_narrow","sts","research_integrity"],"consensus_categories":["research_integrity"],"category_scores_codex":[0.03228002,0.0007035956,0.003210115,0.0009490529,0.008292201,0.0003149565,0.0007066667,0.0013632,0.000414871],"category_scores_gemma":[0.001508835,0.0006626823,0.000393509,0.0006471505,0.00206902,0.000160113,0.0001071248,0.006908168,0.00009470047],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009179602,"about_ca_system_score_gemma":0.01713663,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01002835,"about_ca_topic_score_gemma":0.005162391,"domain_scores_codex":[0.9829872,0.008063841,0.002707469,0.002027533,0.001522517,0.002691458],"domain_scores_gemma":[0.9902228,0.005161493,0.000802003,0.001609107,0.0008314854,0.001373177],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00001748544,0.0001183147,0.00001541792,0.2976727,0.0001697504,0.000009200358,0.002714725,5.552148e-8,4.651073e-8,0.1293429,0.09367245,0.4762669],"study_design_scores_gemma":[0.0008861159,0.0005423461,0.0002327235,0.02606298,0.000159124,0.0003640517,0.0001163025,0.000003855967,1.768081e-8,0.01122516,0.9599301,0.000477231],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.000001646685,0.8923207,0.0008071803,0.08888856,0.0005950396,0.006833727,0.0006998556,0.00008680849,0.009766462],"genre_scores_gemma":[0.0001894146,0.9858176,0.002040268,0.006929072,0.001005374,0.001682674,0.0002070898,0.0001511797,0.001977358],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.8662577,"threshold_uncertainty_score":0.9999332,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"methods","study_design":"not_applicable","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch"],"domain":"methods","study_design":"design_other","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"split"},{"id":"W2112084168","doi":"10.1111/j.1365-2958.2005.04703.x","title":"Weapons of mass retraction","year":2005,"lang":"en","type":"review","venue":"Molecular Microbiology","topic":"Bacterial Genetics and Biotechnology","field":"Biochemistry, Genetics and Molecular Biology","cited_by":203,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"SickKids Foundation; Hospital for Sick Children; University of Toronto","funders":"Natural Sciences and Engineering Research Council of Canada; Cystic Fibrosis Foundation","keywords":"Pilus; Periplasmic space; Biology; Flagellum; Bacterial conjugation; Fimbriae Proteins; Secretion; Cell envelope; Function (biology); Virulence; Bacteria; Cell biology; Computational biology; Microbiology; Escherichia coli; Genetics; Gene; Biochemistry","retraction":null,"screen_n_in":null,"score":{"opus":0.01276240747529528,"gpt":0.276006894292041,"spread":0.2632444868167457,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","research_integrity"],"consensus_categories":[],"category_scores_codex":[0.0001738301,0.0004653053,0.001146449,0.0001956455,0.0000348697,0.000009956259,0.0005338066,0.00261975,0.00006274263],"category_scores_gemma":[0.00005380343,0.0004202442,0.0006137945,0.000181961,0.0002484591,0.000001087599,0.0002356322,0.0004096088,0.00005769203],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000309577,"about_ca_system_score_gemma":0.0001938115,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000004015714,"about_ca_topic_score_gemma":0.000004624183,"domain_scores_codex":[0.9979463,0.0002375021,0.000676965,0.0006951343,0.00004430651,0.0003997478],"domain_scores_gemma":[0.9985268,0.00001226384,0.0005063119,0.0008110366,0.00008554736,0.0000580583],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00001042373,0.00003462236,3.258141e-7,0.0004491713,0.0001886146,0.000005689156,8.806743e-7,6.427669e-7,0.6833119,0.00005747508,0.0005505143,0.3153897],"study_design_scores_gemma":[0.0001496655,0.0002820218,3.104896e-7,0.0002877309,0.0002671203,0.0001987175,0.000001561551,1.430498e-7,0.09964076,0.00001001882,0.8988104,0.0003515158],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.0005608145,0.9971569,0.0004830715,0.00003669879,0.0006000762,0.0004334854,0.0002678121,0.00001948669,0.0004416635],"genre_scores_gemma":[0.0003900949,0.9950572,0.001768835,0.00006940095,0.0003071318,0.0000359661,0.001893266,0.0000801655,0.0003980005],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.8982599,"threshold_uncertainty_score":0.9998249,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"evaluation","study_design":"not_applicable","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","scholarly_communication","research_integrity"],"domain":"evaluation","study_design":"design_other","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"},{"id":"W2920377922","doi":"10.1007/s11948-019-00094-3","title":"Scientific Integrity Principles and Best Practices: Recommendations from a Scientific Integrity Consortium","year":2019,"lang":"en","type":"article","venue":"Science and Engineering Ethics","topic":"Academic integrity and plagiarism","field":"Social Sciences","cited_by":170,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of British Columbia; Canadian Nutrition Society","funders":"Canadian Institutes of Health Research; National Academies of Sciences, Engineering, and Medicine","keywords":"Scientific integrity; Transparency (behavior); Engineering ethics; Best practice; Publication; Harmonization; Scientific misconduct; Research integrity; Process (computing); Philosophy of science; Academic integrity; Scientific evidence; Data integrity; Research ethics; Quality (philosophy); Computer science; Political science; Medicine; Engineering; Law; Computer security","retraction":null,"screen_n_in":null,"score":{"opus":0.1559152298375802,"gpt":0.3841029606226381,"spread":0.2281877307850579,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","sts","scholarly_communication","research_integrity"],"consensus_categories":["sts"],"category_scores_codex":[0.01873629,0.0001533028,0.0001807345,0.0002873384,0.002647763,0.001970425,0.0005943908,0.0006372292,0.0001070427],"category_scores_gemma":[0.0206525,0.0001436727,0.0000288933,0.001208807,0.00364398,0.001797068,0.0002418663,0.004470821,0.00007452969],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001520027,"about_ca_system_score_gemma":0.001995496,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006098246,"about_ca_topic_score_gemma":0.002735912,"domain_scores_codex":[0.9971929,0.0001725971,0.0002653745,0.0007154839,0.001132056,0.000521597],"domain_scores_gemma":[0.9965155,0.001783548,0.0001942748,0.0003233328,0.000876499,0.0003068312],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00001804736,0.0001594435,0.02592501,0.0001849126,0.00003438337,0.000004920286,0.1584096,0.0001014895,0.02633044,0.7785549,0.001965666,0.008311264],"study_design_scores_gemma":[0.0002144895,0.00005091704,0.002869517,0.0004159062,0.00003523625,0.000006819417,0.03012756,0.01519452,0.00121075,0.00570522,0.9436748,0.0004942587],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9709828,0.0003514658,0.00067432,0.01751079,0.003793447,0.0003033499,0.00006362767,0.0001210421,0.00619914],"genre_scores_gemma":[0.9900523,0.0003720737,0.006282815,0.0002389479,0.0001452413,0.000008995726,0.00001734199,0.000008881978,0.00287343],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9417092,"threshold_uncertainty_score":0.9990675,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"methods","study_design":"theoretical_or_conceptual","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","research_integrity"],"domain":"methods","study_design":"theoretical_or_conceptual","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"agree"},{"id":"W1603273306","doi":"10.1177/160940691201100208","title":"Ethics in Qualitative Research: A View of the Participants' and Researchers' World from a Critical Standpoint","year":2012,"lang":"en","type":"article","venue":"International Journal of Qualitative Methods","topic":"Qualitative Research Methods and Ethics","field":"Social Sciences","cited_by":167,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"","keywords":"Beneficence; Transparency (behavior); Research ethics; Engineering ethics; Respect for persons; Economic Justice; Qualitative research; Vocabulary; Sociology; Psychology; Autonomy; Social science; Political science; Law","retraction":null,"screen_n_in":null,"score":{"opus":0.9688948298709127,"gpt":0.8332285134866256,"spread":0.1356663163842871,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","sts","research_integrity"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.4183765,0.0001474343,0.0005266941,0.0006807555,0.0003468565,0.0001182443,0.0009626934,0.0001872563,0.0002516778],"category_scores_gemma":[0.3282923,0.0001029294,0.0001955975,0.001221005,0.007228603,0.0007376932,0.0003508626,0.003400315,0.000003655985],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005138294,"about_ca_system_score_gemma":0.001394215,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003997945,"about_ca_topic_score_gemma":0.001991097,"domain_scores_codex":[0.6875256,0.3051711,0.001452868,0.0002499321,0.004686192,0.0009142988],"domain_scores_gemma":[0.7362962,0.2562111,0.0005718652,0.0002262481,0.00615579,0.0005388077],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"qualitative","study_design_scores_codex":[0.0002191084,0.0001518169,0.0005247926,0.00002735848,0.0001113704,0.000004154846,0.6021863,6.938253e-7,0.0007698619,0.3884732,0.0001186771,0.007412635],"study_design_scores_gemma":[0.0003716286,0.00009556516,0.002032316,0.0005127305,0.00001176911,0.000001023141,0.5217708,0.00001247774,0.001120362,0.4596048,0.01438012,0.00008645483],"study_design_candidate":"qualitative","study_design_consensus":"qualitative","genre_codex":"commentary","genre_gemma":"empirical","genre_scores_codex":[0.3177725,0.0246542,0.2610137,0.3777421,0.003367331,0.001096385,0.0003186128,0.00001946648,0.01401571],"genre_scores_gemma":[0.6959472,0.001160413,0.3017209,0.0003620993,0.0004427646,0.00002010156,8.433867e-7,0.00001804041,0.0003276197],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.3781748,"threshold_uncertainty_score":0.9988989,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["research_integrity"],"domain":null,"study_design":"qualitative","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch"],"domain":"methods","study_design":"theoretical_or_conceptual","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"split"},{"id":"W2926567176","doi":"10.1136/bmjopen-2018-026516","title":"Knowledge and motivations of researchers publishing in presumed predatory journals: a survey","year":2019,"lang":"en","type":"article","venue":"BMJ Open","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":160,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Waterloo; University of Ottawa; Jewish General Hospital; Institut du Savoir Montfort; McGill University; Ottawa Hospital","funders":"Canadian Institutes of Health Research; Ottawa Hospital Anesthesia Alternate Funds Association","keywords":"Publishing; Medicine; Psychological intervention; Law; Psychiatry; Political science","retraction":null,"screen_n_in":null,"score":{"opus":0.9177965981725121,"gpt":0.7046737378033378,"spread":0.2131228603691744,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","scholarly_communication","insufficient_payload"],"consensus_categories":["metaresearch","bibliometrics"],"category_scores_codex":[0.1739013,0.00007749222,0.0003312395,0.03416999,0.00008212571,0.008267575,0.004144824,0.0001052637,0.0009398088],"category_scores_gemma":[0.2431802,0.00005554944,0.00003906314,0.1161056,0.0001075579,0.004209267,0.003837105,0.0003892387,0.0001154252],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00007752397,"about_ca_system_score_gemma":0.0008272337,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002134968,"about_ca_topic_score_gemma":0.001305276,"domain_scores_codex":[0.9906144,0.00190393,0.0009143365,0.0005604162,0.005557302,0.0004496672],"domain_scores_gemma":[0.9776171,0.01597779,0.0003915449,0.001064246,0.004543257,0.0004060529],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00006925716,0.0000938103,0.9502512,0.00001184831,0.000007349843,0.000001551836,0.0003827097,0.00001574663,0.0001796146,0.0002351482,0.02334965,0.02540216],"study_design_scores_gemma":[0.0006909508,0.00005686073,0.9891419,0.00005437819,4.717402e-7,0.000001350397,0.0005031046,0.002807582,0.0000988395,0.001610964,0.004959528,0.00007401575],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9533338,0.0006387595,0.00007307276,0.0006674282,0.0001485,0.002743282,0.0000650462,0.000003218559,0.04232689],"genre_scores_gemma":[0.9908869,0.00006337643,0.0003441161,0.00002451333,0.00001569336,0.00005940539,0.000004024901,0.000007059421,0.008594934],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.08193564,"threshold_uncertainty_score":0.9999735,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"incentives","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["scholarly_communication"],"domain":null,"study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W2113866902","doi":"10.1177/1049732303259842","title":"Ethical Dilemmas in Research on Internet Communities","year":2004,"lang":"en","type":"article","venue":"Qualitative Health Research","topic":"Focus Groups and Qualitative Methods","field":"Social Sciences","cited_by":158,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto; Public Health Ontario","funders":"","keywords":"Beneficence; The Internet; Autonomy; Harm; Economic Justice; Research ethics; Ethical issues; Engineering ethics; Internet research; Public relations; Psychology; Internet privacy; Sociology; Political science; Social psychology; Law; Computer science","retraction":null,"screen_n_in":null,"score":{"opus":0.937739584532181,"gpt":0.7846211578711447,"spread":0.1531184266610363,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","sts","research_integrity"],"consensus_categories":["metaresearch","sts"],"category_scores_codex":[0.2686574,0.0001723515,0.0004674215,0.001441296,0.002149611,0.0001897933,0.000914363,0.0003966079,0.0003746543],"category_scores_gemma":[0.01456298,0.0001573273,0.0000785404,0.003221313,0.005848081,0.0001836553,0.0002718344,0.007988679,0.0005839396],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002733174,"about_ca_system_score_gemma":0.002696031,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.2573316,"about_ca_topic_score_gemma":0.1009283,"domain_scores_codex":[0.81703,0.1753336,0.0007801901,0.0004952658,0.004111363,0.002249557],"domain_scores_gemma":[0.9540235,0.04336854,0.00009267224,0.0004610645,0.001458822,0.0005953537],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"qualitative","study_design_scores_codex":[0.0001071304,0.0001639306,0.00005380827,0.00005419323,0.000004656192,0.000006027364,0.4748898,0.000005140143,0.000002837115,0.51892,0.002352488,0.003439924],"study_design_scores_gemma":[0.0005544907,0.001339784,0.0007841635,0.0003625378,2.304453e-7,2.04144e-7,0.6337432,0.000006991852,0.00003818686,0.2904013,0.07265185,0.0001170641],"study_design_candidate":"qualitative","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5256313,0.001004087,0.002095188,0.3299152,0.0003370107,0.002219442,0.0000472178,0.0001152459,0.1386353],"genre_scores_gemma":[0.9888693,0.0004540108,0.005284961,0.0019621,0.0002067919,0.0002124582,0.00001429105,0.00003265781,0.002963436],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.463238,"threshold_uncertainty_score":0.9991494,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["research_integrity"],"domain":null,"study_design":"theoretical_or_conceptual","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","research_integrity"],"domain":"methods","study_design":"qualitative","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"split"},{"id":"W4249889786","doi":"10.12688/f1000research.15256.1","title":"What is a predatory journal? A scoping review","year":2018,"lang":"en","type":"review","venue":"F1000Research","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":154,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Ottawa Hospital; University of Ottawa","funders":"Ottawa Hospital Anesthesia Alternate Funds Association; Ottawa Hospital Research Institute; University of Ottawa","keywords":"PsycINFO; Grey literature; MEDLINE; CINAHL; Web of science; Medicine; Information retrieval; Computer science; Biology","retraction":null,"screen_n_in":null,"score":{"opus":0.9152973410968032,"gpt":0.7496890964298902,"spread":0.1656082446669129,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","metaepi_narrow","bibliometrics","scholarly_communication","open_science","research_integrity","insufficient_payload"],"consensus_categories":["metaresearch","bibliometrics","insufficient_payload"],"category_scores_codex":[0.1526493,0.0006879686,0.003806303,0.1078312,0.0009572492,0.02663212,0.0164496,0.000695651,0.02420215],"category_scores_gemma":[0.1051354,0.0004125765,0.002047736,0.2761776,0.000774161,0.003172324,0.005682404,0.003014539,0.02809673],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006484403,"about_ca_system_score_gemma":0.008336073,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000009779443,"about_ca_topic_score_gemma":0.000005209153,"domain_scores_codex":[0.9498433,0.004448206,0.003775579,0.00223753,0.0374523,0.002243045],"domain_scores_gemma":[0.9553101,0.01898348,0.001865007,0.004794673,0.0165202,0.002526517],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00000159701,0.0000402071,0.000004013262,0.02608021,0.00004341976,0.00006166766,0.00002013719,6.281368e-9,2.590152e-8,0.000008139489,0.3715579,0.6021827],"study_design_scores_gemma":[0.00008043241,0.0001371202,0.000001225176,0.3154487,0.00005466002,0.0001965904,0.00003515115,0.000009269235,6.34924e-7,0.0004700719,0.6832628,0.0003033261],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[5.255653e-7,0.9891187,0.0001062777,0.0008828244,0.002619705,0.003716286,0.00009015624,0.00003979166,0.003425746],"genre_scores_gemma":[5.298901e-8,0.9639038,0.0002296602,0.0005779005,0.0011173,0.0002231885,0.00001602876,0.0001024554,0.03382955],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.6018794,"threshold_uncertainty_score":0.9998326,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","scholarly_communication"],"domain":"evaluation","study_design":"systematic_review","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["research_integrity","scholarly_communication"],"domain":null,"study_design":"systematic_review","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W2889499003","doi":"10.1503/cmaj.180154","title":"How predatory journals leak into PubMed","year":2018,"lang":"en","type":"article","venue":"Canadian Medical Association Journal","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":146,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":true,"about_ca":false},"ca_institutions":"Ottawa Hospital","funders":"University of Ottawa","keywords":"Key (lock); Data science; Computer science; MEDLINE; World Wide Web; Computer security; Political science","retraction":null,"screen_n_in":null,"score":{"opus":0.3334433946008996,"gpt":0.4961553382486549,"spread":0.1627119436477553,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","scholarly_communication","insufficient_payload"],"consensus_categories":["metaresearch","bibliometrics","insufficient_payload"],"category_scores_codex":[0.08107378,0.0001390894,0.0003110713,0.03402918,0.001081469,0.01064864,0.002797762,0.0005501052,0.01038845],"category_scores_gemma":[0.3063317,0.00009917778,0.0002036654,0.06545956,0.0002076342,0.001307828,0.0001640679,0.001497262,0.001171165],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001971778,"about_ca_system_score_gemma":0.005237541,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0007693563,"about_ca_topic_score_gemma":0.03113579,"domain_scores_codex":[0.9711182,0.0007254727,0.0008256675,0.0003913632,0.02564573,0.001293546],"domain_scores_gemma":[0.9762717,0.002183804,0.0006431378,0.0003524112,0.01287234,0.007676636],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.000002213984,0.00001248318,0.1015512,5.924e-7,0.00002489375,0.0001193744,0.0001234219,2.134901e-7,0.00000590086,0.0001070481,0.4496251,0.4484275],"study_design_scores_gemma":[0.0003753556,0.00006154756,0.1490853,0.00001317011,0.000003368051,0.0001801671,0.0005085059,0.0004152021,0.00003005905,0.005552176,0.8436528,0.0001223041],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"commentary","genre_gemma":"empirical","genre_scores_codex":[0.41811,0.004612911,0.009230076,0.4997303,0.0150188,0.0004934419,0.00009336133,0.00005080419,0.0526603],"genre_scores_gemma":[0.9742511,0.000494603,0.0002422246,0.006159651,0.00583498,0.000007096136,0.000001337181,0.00001413167,0.01299491],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.556141,"threshold_uncertainty_score":0.9996065,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","bibliometrics","research_integrity"],"domain":"evaluation","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","bibliometrics","scholarly_communication","research_integrity"],"domain":"evaluation","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"split"},{"id":"W2167083241","doi":"10.1525/jer.2010.5.4.23","title":"Researchers' Perspectives on Collective/Community Co-Authorship in Community-Based Participatory Indigenous Research","year":2010,"lang":"en","type":"article","venue":"Journal of Empirical Research on Human Research Ethics","topic":"Mental Health and Patient Involvement","field":"Health Professions","cited_by":143,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Dalhousie University","funders":"Canadian Institutes of Health Research","keywords":"Indigenous; Community-based participatory research; Participatory action research; Citizen journalism; Thematic analysis; Sociology; Public relations; Value (mathematics); Qualitative research; Political science; Social science; Law","retraction":null,"screen_n_in":null,"score":{"opus":0.9721388023156835,"gpt":0.7766190470881619,"spread":0.1955197552275216,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","metaepi_narrow","sts","research_integrity","insufficient_payload"],"consensus_categories":["metaresearch","sts","research_integrity","insufficient_payload"],"category_scores_codex":[0.5064174,0.0004420521,0.00100109,0.008870157,0.02797791,0.000245241,0.004061491,0.002178831,0.001091743],"category_scores_gemma":[0.09583548,0.0003656672,0.0002754447,0.00601737,0.006127353,0.0003091239,0.001235926,0.2490152,0.0008444211],"about_ca_system_candidate":true,"about_ca_system_consensus":true,"about_ca_system_score_codex":0.009017,"about_ca_system_score_gemma":0.02715219,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007181549,"about_ca_topic_score_gemma":0.01635125,"domain_scores_codex":[0.7033116,0.2745714,0.002100617,0.000597438,0.01459729,0.004821596],"domain_scores_gemma":[0.7895505,0.1909528,0.0004292162,0.002170802,0.01395684,0.00293978],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"qualitative","study_design_scores_codex":[0.01455392,0.02158808,0.1580751,0.00459101,0.0001206924,0.0005297887,0.7379432,0.00005178028,0.004892328,0.01697019,0.03978676,0.0008971339],"study_design_scores_gemma":[0.009262708,0.06372354,0.1392913,0.008578548,0.00001039576,0.000006718061,0.607281,0.0001304724,0.003034831,0.1121344,0.05580376,0.0007423294],"study_design_candidate":"qualitative","study_design_consensus":"qualitative","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9536884,0.0001763211,0.000004443382,0.01033011,0.0003301762,0.00365988,0.00006137898,0.00003547268,0.03171386],"genre_scores_gemma":[0.993802,0.0002749264,0.0001426507,0.001682727,0.0007813583,0.000637506,0.00004885977,0.0001179952,0.002511961],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4105819,"threshold_uncertainty_score":0.9999335,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"methods","study_design":"qualitative","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"high"},{"model":"gpt","categories":["research_integrity","scholarly_communication"],"domain":null,"study_design":"qualitative","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W2811055236","doi":"10.12688/f1000research.15256.2","title":"What is a predatory journal? A scoping review","year":2018,"lang":"en","type":"review","venue":"F1000Research","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":141,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Ottawa Hospital; University of Ottawa","funders":"Ottawa Hospital Anesthesia Alternate Funds Association; Ottawa Hospital Research Institute; University of Ottawa","keywords":"PsycINFO; MEDLINE; Grey literature; Web of science; CINAHL; Medicine; Psychology; Meta-analysis; Political science; Pathology","retraction":null,"screen_n_in":null,"score":{"opus":0.9152973410968032,"gpt":0.7496890964298902,"spread":0.1656082446669129,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","metaepi_narrow","bibliometrics","scholarly_communication","open_science","research_integrity","insufficient_payload"],"consensus_categories":["metaresearch","bibliometrics","insufficient_payload"],"category_scores_codex":[0.1526493,0.0006879686,0.003806303,0.1078312,0.0009572492,0.02663212,0.0164496,0.000695651,0.02420215],"category_scores_gemma":[0.1051354,0.0004125765,0.002047736,0.2761776,0.000774161,0.003172324,0.005682404,0.003014539,0.02809673],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006484403,"about_ca_system_score_gemma":0.008336073,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000009779443,"about_ca_topic_score_gemma":0.000005209153,"domain_scores_codex":[0.9498433,0.004448206,0.003775579,0.00223753,0.0374523,0.002243045],"domain_scores_gemma":[0.9553101,0.01898348,0.001865007,0.004794673,0.0165202,0.002526517],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00000159701,0.0000402071,0.000004013262,0.02608021,0.00004341976,0.00006166766,0.00002013719,6.281368e-9,2.590152e-8,0.000008139489,0.3715579,0.6021827],"study_design_scores_gemma":[0.00008043241,0.0001371202,0.000001225176,0.3154487,0.00005466002,0.0001965904,0.00003515115,0.000009269235,6.34924e-7,0.0004700719,0.6832628,0.0003033261],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[5.255653e-7,0.9891187,0.0001062777,0.0008828244,0.002619705,0.003716286,0.00009015624,0.00003979166,0.003425746],"genre_scores_gemma":[5.298901e-8,0.9639038,0.0002296602,0.0005779005,0.0011173,0.0002231885,0.00001602876,0.0001024554,0.03382955],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.6018794,"threshold_uncertainty_score":0.9998326,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity","scholarly_communication"],"domain":"evaluation","study_design":"systematic_review","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","research_integrity"],"domain":"methods","study_design":"design_other","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W2117161745","doi":"10.1177/104973202129120313","title":"Leveling the Playing Field? Exploring the Merits of the Ethics-as-Process Approach for Judging Qualitative Research Proposals","year":2002,"lang":"en","type":"article","venue":"Qualitative Health Research","topic":"Qualitative Research Methods and Ethics","field":"Social Sciences","cited_by":136,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"","keywords":"Qualitative research; Engineering ethics; Research ethics; Field (mathematics); Process (computing); Informed consent; Management science; Psychology; Sociology; Medicine; Social science; Computer science; Alternative medicine; Engineering","retraction":null,"screen_n_in":null,"score":{"opus":0.9614351793248613,"gpt":0.7788944891919308,"spread":0.1825406901329305,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","sts","research_integrity"],"consensus_categories":["metaresearch","sts"],"category_scores_codex":[0.4866171,0.0002927598,0.000632529,0.0004751255,0.02056025,0.000407509,0.002909255,0.0003456052,0.0001132285],"category_scores_gemma":[0.2455912,0.0001589287,0.000285747,0.004909304,0.01083141,0.0005625841,0.0005835901,0.009187994,0.0000288247],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006521303,"about_ca_system_score_gemma":0.004151767,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01667967,"about_ca_topic_score_gemma":0.001525489,"domain_scores_codex":[0.7182604,0.2661942,0.001465182,0.001030865,0.009762552,0.003286815],"domain_scores_gemma":[0.6899015,0.2989656,0.0004780786,0.001065857,0.009094584,0.0004944389],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"qualitative","study_design_scores_codex":[0.0000811397,0.00008665583,0.00001309365,0.0007507377,0.00004672326,3.349381e-7,0.6318705,0.000009576025,0.00007156163,0.3575507,0.001800642,0.007718362],"study_design_scores_gemma":[0.0002599962,0.0004156274,0.00002368035,0.0003754509,0.000004032593,5.366941e-7,0.7433496,0.0007375151,0.0005259687,0.2477611,0.006396885,0.000149713],"study_design_candidate":"qualitative","study_design_consensus":"qualitative","genre_codex":"commentary","genre_gemma":"empirical","genre_scores_codex":[0.06587205,0.004537596,0.04292654,0.8539986,0.0004933998,0.01043175,0.0001190801,0.00008876986,0.02153218],"genre_scores_gemma":[0.9615937,0.003225911,0.02348465,0.002107216,0.000739027,0.00457211,0.000006970798,0.00009834122,0.004172048],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8957217,"threshold_uncertainty_score":0.9930978,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"evaluation","study_design":"theoretical_or_conceptual","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","research_integrity"],"domain":"methods","study_design":"theoretical_or_conceptual","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"agree"},{"id":"W2089994419","doi":"10.1007/s11948-014-9600-6","title":"Scientists Admitting to Plagiarism: A Meta-analysis of Surveys","year":2014,"lang":"en","type":"review","venue":"Science and Engineering Ethics","topic":"Academic integrity and plagiarism","field":"Social Sciences","cited_by":132,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal","funders":"","keywords":"Meta-analysis; Psychology; Survey research; Recall; Philosophy of science; Publication bias; Peer review; Medical education; Medicine; Political science; Applied psychology; Law; Pathology","retraction":null,"screen_n_in":null,"score":{"opus":0.2288966979256258,"gpt":0.4437039106443108,"spread":0.214807212718685,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","research_integrity"],"consensus_categories":["metaresearch","research_integrity"],"category_scores_codex":[0.07764409,0.0002697625,0.002180726,0.001514805,0.0007418274,0.0002194498,0.001130967,0.001555659,0.00004745154],"category_scores_gemma":[0.0197689,0.0002174099,0.0006969963,0.00735109,0.0009652114,0.0002109121,0.0002197498,0.003403087,0.0000140053],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001429828,"about_ca_system_score_gemma":0.001771373,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00181898,"about_ca_topic_score_gemma":0.000548003,"domain_scores_codex":[0.9954641,0.0007758982,0.0005951638,0.0006131256,0.001953616,0.0005980919],"domain_scores_gemma":[0.9953203,0.003020309,0.0002668872,0.0003536761,0.0006572473,0.0003816001],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.000001291686,0.00005101263,0.00002416362,0.01455665,0.040513,0.00001135363,0.06918689,0.002194946,0.00001297082,0.2922157,0.0009697952,0.5802622],"study_design_scores_gemma":[0.00001023195,0.00001066939,0.00000955829,0.0003944373,0.04345205,9.02896e-7,0.0001379364,0.0005644653,0.000001422827,0.00004726762,0.9550853,0.0002857264],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.0000138275,0.98156,0.0147156,0.0006529504,0.0006509732,0.0003666332,0.00008786611,0.00007212668,0.00188004],"genre_scores_gemma":[0.000695914,0.9958811,0.002447598,0.0001967417,0.0001262427,0.00002576354,0.00001108822,0.00001794322,0.0005975979],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.9541155,"threshold_uncertainty_score":0.9997405,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"methods","study_design":"meta_analysis","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"high"},{"model":"gpt","categories":["research_integrity"],"domain":null,"study_design":"meta_analysis","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W4386306053","doi":"10.2196/51584","title":"Best Practices for Using AI Tools as an Author, Peer Reviewer, or Editor","year":2023,"lang":"en","type":"editorial","venue":"Journal of Medical Internet Research","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":105,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Victoria; JMIR Publications","funders":"","keywords":"Chatbot; Computer science; Publishing; Scientific publishing; Generative grammar; Data science; World Wide Web; Artificial intelligence; Political science","retraction":null,"screen_n_in":null,"score":{"opus":0.693170544728797,"gpt":0.6910681960244153,"spread":0.002102348704381751,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","research_integrity","insufficient_payload"],"consensus_categories":["metaresearch","research_integrity"],"category_scores_codex":[0.03871451,0.0002975559,0.001124459,0.0008422498,0.0001751369,0.0004563536,0.00123752,0.002506826,0.002797076],"category_scores_gemma":[0.6053051,0.0001992717,0.0003587129,0.0006198622,0.000353111,0.0006258988,0.0002185498,0.008345807,0.0003691042],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008894872,"about_ca_system_score_gemma":0.01947895,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006883003,"about_ca_topic_score_gemma":0.0007595627,"domain_scores_codex":[0.9830087,0.0009987855,0.002179354,0.0005206966,0.01240488,0.0008875646],"domain_scores_gemma":[0.8107825,0.07871965,0.006115306,0.002017484,0.09572633,0.006638729],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.002380294,0.0005071567,0.00001200366,0.001842536,0.0001761609,0.0004676526,0.0007482832,2.930108e-7,0.00001057257,0.00003371106,0.9176924,0.0761289],"study_design_scores_gemma":[0.0002214621,0.006599505,8.521753e-7,0.01012903,0.0002386211,0.0001689718,0.003002647,0.0009594869,0.0001820291,0.00043271,0.9778954,0.0001692257],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"editorial","genre_gemma":"editorial","genre_scores_codex":[0.002011502,0.00158571,0.0004360154,0.04589859,0.9490376,0.0008558461,0.0000335564,0.00002139568,0.0001197806],"genre_scores_gemma":[0.0001395508,0.005374028,0.000681482,0.000320145,0.949342,0.00005989476,0.0001576658,0.0001061589,0.04381905],"genre_candidate":"editorial","genre_consensus":"editorial","teacher_disagreement_score":0.5665906,"threshold_uncertainty_score":0.9997302,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch"],"domain":"evaluation","study_design":"not_applicable","genre":"editorial","about_ca_system":false,"about_ca_topic":false,"confidence":"high"},{"model":"gpt","categories":["research_integrity"],"domain":null,"study_design":"not_applicable","genre":"editorial","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W2268329921","doi":"10.15388/infedu.2014.10","title":"Collaboration, Collusion and Plagiarism in Computer Science Coursework","year":2014,"lang":"en","type":"article","venue":"Informatics in Education","topic":"Academic integrity and plagiarism","field":"Social Sciences","cited_by":97,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Manitoba","funders":"University of Waterloo","keywords":"Coursework; Collusion; Context (archaeology); Academic dishonesty; Computer science; Mathematics education; Engineering ethics; Cheating; Psychology; Engineering","retraction":null,"screen_n_in":null,"score":{"opus":0.006463584133639145,"gpt":0.3096939182671761,"spread":0.303230334133537,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003881174,0.00005646181,0.00008070939,0.000277884,0.0003111353,0.0001567458,0.0001840695,0.0002851964,0.00001096811],"category_scores_gemma":[0.0008043383,0.00005924874,0.000005532207,0.0008957494,0.0003400109,0.0007414352,0.00003689712,0.0004576052,0.00001575612],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002403933,"about_ca_system_score_gemma":0.001279852,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0007870718,"about_ca_topic_score_gemma":0.001920091,"domain_scores_codex":[0.9990182,0.00009559467,0.0002945056,0.00007663834,0.0003296118,0.000185476],"domain_scores_gemma":[0.9993713,0.0001673861,0.0001071249,0.00009221601,0.0001951591,0.00006678449],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"not_applicable","study_design_scores_codex":[0.000009975129,0.0001150794,0.06161882,0.00003722968,0.00000126773,1.045631e-7,0.442059,0.0009215943,0.00001569409,0.312668,0.007334311,0.1752189],"study_design_scores_gemma":[0.001155459,0.0001279753,0.1601398,0.0007640398,0.00001131838,0.000006975675,0.2092646,0.1538233,0.0001821695,0.1376664,0.3360752,0.000782709],"study_design_candidate":"qualitative","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9578862,0.00004425657,0.01015655,0.003995527,0.001496637,0.0003289032,0.000001207232,0.00001989048,0.02607086],"genre_scores_gemma":[0.9868416,0.000114562,0.01168822,0.001027717,0.0002174058,0.00001101823,0.000005725572,0.000002277213,0.00009147491],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3287409,"threshold_uncertainty_score":0.2416094,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["research_integrity"],"domain":null,"study_design":"not_applicable","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"high"},{"model":"gpt","categories":["research_integrity"],"domain":null,"study_design":"not_applicable","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"agree"},{"id":"W3006156524","doi":"10.1136/bmjopen-2019-035561","title":"Defining predatory journals and responding to the threat they pose: a modified Delphi consensus process","year":2020,"lang":"en","type":"article","venue":"BMJ Open","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":97,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Ottawa; Ottawa Hospital","funders":"Social Sciences and Humanities Research Council of Canada; Canadian Institutes of Health Research; Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung; Ottawa Hospital Anesthesia Alternate Funds Association; Institute of Health Services and Policy Research; University of Ottawa; Institute of Musculoskeletal Health and Arthritis; Natural Sciences and Engineering Research Council of Canada; Ottawa Hospital Research Institute; National Science Foundation","keywords":"Outreach; Snowball sampling; Delphi method; Publishing; Checklist; Public relations; Library science; Medicine; Political science; Psychology; Law","retraction":null,"screen_n_in":null,"score":{"opus":0.7858935230383207,"gpt":0.6547070806373642,"spread":0.1311864424009566,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","scholarly_communication"],"consensus_categories":["metaresearch","bibliometrics"],"category_scores_codex":[0.03954222,0.0001405398,0.0003389383,0.01377735,0.0006030073,0.008913535,0.004422141,0.00007675587,0.0002524096],"category_scores_gemma":[0.08963586,0.00007579236,0.0000602112,0.08829046,0.00008188947,0.0004916695,0.003158334,0.0003015747,0.0003109448],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003409904,"about_ca_system_score_gemma":0.0004762844,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00006838201,"about_ca_topic_score_gemma":0.00003389104,"domain_scores_codex":[0.9920571,0.000536052,0.0006957619,0.0007127378,0.005506241,0.0004921317],"domain_scores_gemma":[0.9916403,0.005009608,0.0002962572,0.0006942268,0.00162327,0.0007363859],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"observational","study_design_scores_codex":[0.004967972,0.000191508,0.1973189,0.00008048205,0.0001811282,0.0004377181,0.03539332,0.004631078,0.003111247,0.0207831,0.3689229,0.3639806],"study_design_scores_gemma":[0.008940578,0.003167517,0.4326535,0.0007237735,0.000090552,0.0007423301,0.06685396,0.1473633,0.004403548,0.1005632,0.2316962,0.002801507],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9013997,0.001192291,0.0009614991,0.05913519,0.0002580162,0.004841887,0.00009786884,0.00002530522,0.03208828],"genre_scores_gemma":[0.9951742,0.00002938729,0.0008091661,0.002458923,0.0001013108,0.0001367069,7.719067e-7,0.00001307676,0.001276507],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3611791,"threshold_uncertainty_score":0.9974006,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","scholarly_communication"],"domain":"evaluation","study_design":"qualitative","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","research_integrity","scholarly_communication"],"domain":"methods","study_design":"qualitative","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"split"},{"id":"W2625269005","doi":"10.11613/bm.2017.030","title":"Ethical issues in publishing in predatory journals","year":2017,"lang":"en","type":"review","venue":"Biochemia Medica","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":97,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Public Health Ontario; University of Toronto","funders":"","keywords":"Publishing; Misrepresentation; Publication; Deception; Public relations; Political science; Library science; Sociology; Law; Computer science","retraction":null,"screen_n_in":null,"score":{"opus":0.8941266353903872,"gpt":0.7234011886426125,"spread":0.1707254467477747,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","metaepi_narrow","bibliometrics","scholarly_communication","open_science","research_integrity","insufficient_payload"],"consensus_categories":["metaresearch","bibliometrics","research_integrity"],"category_scores_codex":[0.1373145,0.0003975683,0.002532168,0.1038878,0.0001685755,0.0122657,0.01187921,0.003114029,0.001435708],"category_scores_gemma":[0.522408,0.0002551949,0.0005101176,0.1074857,0.0004509402,0.001125718,0.002496486,0.006580627,0.00045156],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005455957,"about_ca_system_score_gemma":0.003179752,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002103249,"about_ca_topic_score_gemma":0.00006388654,"domain_scores_codex":[0.9741449,0.00129049,0.00283465,0.001552011,0.01902187,0.001156033],"domain_scores_gemma":[0.9818915,0.01127158,0.001624653,0.002547512,0.001562822,0.001101993],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00000173114,0.00005011307,0.0005362217,0.0005061114,0.00001007741,0.0001456388,0.00004652319,2.385062e-8,6.428021e-7,0.00003117385,0.1395514,0.8591204],"study_design_scores_gemma":[0.000248944,0.0000204274,0.0003436936,0.006270661,0.00001089388,0.00002270825,0.00004853031,0.00001919521,0.000005418361,0.0008772882,0.9918603,0.0002719855],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.00004736704,0.986076,0.00001017941,0.003608942,0.001024155,0.0004252271,0.00005346555,0.00001412149,0.008740528],"genre_scores_gemma":[0.0001578872,0.996204,0.0002174515,0.0001418826,0.0006053997,0.0000538875,0.00002064615,0.00002710381,0.002571774],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.8588484,"threshold_uncertainty_score":0.99999,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"evaluation","study_design":"not_applicable","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","research_integrity","scholarly_communication"],"domain":"evaluation","study_design":"design_other","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W1883095706","doi":"10.1111/jlme.12288","title":"Returning a Research Participant's Genomic Results to Relatives: Analysis and Recommendations","year":2015,"lang":"en","type":"article","venue":"The Journal of Law Medicine & Ethics","topic":"Ethics in Clinical Research","field":"Medicine","cited_by":94,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Dalhousie University","funders":"Eunice Kennedy Shriver National Institute of Child Health and Human Development; National Cancer Institute; National Human Genome Research Institute; Wellcome Trust","keywords":"Participant observation; Psychology; Social psychology; Sociology; Social science","retraction":null,"screen_n_in":null,"score":{"opus":0.9094334095054257,"gpt":0.6971338435979062,"spread":0.2122995659075195,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","research_integrity"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.1665347,0.0001426089,0.0007153242,0.0006347437,0.0004403032,0.00003278071,0.000478104,0.0004437634,0.0000558094],"category_scores_gemma":[0.2718719,0.00008025049,0.0001067424,0.001718331,0.001927118,0.0001192546,0.0003209712,0.0137396,0.00002402878],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002152113,"about_ca_system_score_gemma":0.001779941,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0008113317,"about_ca_topic_score_gemma":0.001984794,"domain_scores_codex":[0.9904951,0.003577596,0.0015765,0.0002542861,0.003604312,0.0004921866],"domain_scores_gemma":[0.8815385,0.1074033,0.0006194103,0.0009672088,0.007993066,0.001478533],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.01647097,0.0005210806,0.007445386,0.0004734759,0.00514241,0.0004685232,0.7406046,0.0007277976,0.01281741,0.1731734,0.03987732,0.002277566],"study_design_scores_gemma":[0.02259071,0.03469294,0.0410779,0.01339334,0.01445755,0.0014321,0.2037051,0.001815398,0.001765276,0.4466092,0.2176495,0.0008109534],"study_design_candidate":"qualitative","study_design_consensus":null,"genre_codex":"commentary","genre_gemma":"empirical","genre_scores_codex":[0.274205,0.001365471,0.001263786,0.7076376,0.0002445904,0.0003566587,0.000006591465,0.00001479979,0.01490549],"genre_scores_gemma":[0.9859632,0.003034634,0.003742665,0.004672692,0.0007627544,0.000003621813,0.0000049532,0.00002391553,0.001791587],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.7117582,"threshold_uncertainty_score":0.9885358,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["research_integrity"],"domain":null,"study_design":"theoretical_or_conceptual","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"high"},{"model":"gpt","categories":[],"domain":null,"study_design":"theoretical_or_conceptual","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"},{"id":"W2887657800","doi":"10.1097/acm.0000000000002412","title":"Ethical Shades of Gray: International Frequency of Scientific Misconduct and Questionable Research Practices in Health Professions Education","year":2018,"lang":"en","type":"article","venue":"Academic Medicine","topic":"Academic integrity and plagiarism","field":"Social Sciences","cited_by":92,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"","keywords":"Scientific misconduct; Misconduct; Psychology; Research ethics; Medical education; Gray (unit); Public relations; Medicine; Alternative medicine; Political science; Psychiatry; Law; Pathology","retraction":null,"screen_n_in":null,"score":{"opus":0.2293234273902231,"gpt":0.5565311650287235,"spread":0.3272077376385004,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","sts","research_integrity"],"consensus_categories":[],"category_scores_codex":[0.0253873,0.00006143221,0.0001882492,0.0004366202,0.0003769903,0.000007696217,0.0004059719,0.0008029641,0.0003984661],"category_scores_gemma":[0.02934556,0.00004905369,0.00001041418,0.0007418461,0.003841345,0.0002872318,0.00006359145,0.003697915,0.000005709488],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001173674,"about_ca_system_score_gemma":0.002442566,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0351973,"about_ca_topic_score_gemma":0.002094522,"domain_scores_codex":[0.996519,0.00142109,0.0005464078,0.0002528834,0.000999226,0.0002613487],"domain_scores_gemma":[0.9969721,0.001480505,0.0005136475,0.0001141924,0.000788413,0.0001310885],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.00002448148,0.00003820711,0.07711083,0.000124319,0.000009376504,3.52319e-7,0.141815,1.015883e-7,0.007510806,0.7308824,0.03911689,0.003367164],"study_design_scores_gemma":[0.0006295893,0.0003336036,0.02011402,0.005544062,0.00001543067,0.00001017487,0.2124532,0.0001039507,0.001521905,0.6055532,0.1535455,0.0001753288],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7041979,0.002668146,0.00006366779,0.2692865,0.001567769,0.0004177843,0.000006745798,0.00001364753,0.02177778],"genre_scores_gemma":[0.9934726,0.001704212,0.0005007791,0.0006347347,0.0007883325,0.00001296422,0.00001519357,0.000003854809,0.002867303],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2892747,"threshold_uncertainty_score":0.9988697,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"methods","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","research_integrity"],"domain":"methods","study_design":"design_other","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"split"},{"id":"W4386541616","doi":"10.1002/leap.1578","title":"Risks of abuse of large language models, like <scp>ChatGPT</scp>, in scientific publishing: Authorship, predatory publishing, and paper mills","year":2023,"lang":"en","type":"article","venue":"Learned Publishing","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":87,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Thompson Rivers University","keywords":"Publishing; Mill; Scientific publishing; Key (lock); Public relations; Computer science; Political science; Law; Engineering; Computer security; Mechanical engineering","retraction":null,"screen_n_in":null,"score":{"opus":0.2084196632622204,"gpt":0.4104631433775228,"spread":0.2020434801153024,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","metaepi_narrow","scholarly_communication"],"consensus_categories":["scholarly_communication"],"category_scores_codex":[0.008828962,0.0002687126,0.0005862408,0.001634216,0.00023406,0.003680201,0.0006471984,0.000745654,0.00008204586],"category_scores_gemma":[0.01905698,0.0002656536,0.0001344357,0.003156042,0.0002616288,0.017544,0.0002742603,0.001698549,0.00001805895],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001478321,"about_ca_system_score_gemma":0.0006580333,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00893541,"about_ca_topic_score_gemma":0.002423481,"domain_scores_codex":[0.9955184,0.0003646541,0.001179526,0.0008339625,0.001141576,0.0009618078],"domain_scores_gemma":[0.9960217,0.001137769,0.0005504957,0.0009025291,0.0009051728,0.0004823552],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"qualitative","study_design_scores_codex":[0.000150588,0.001221345,0.2168823,0.003019982,0.0001633088,0.00007502364,0.4531141,0.001061479,0.008363557,0.01650324,0.2211933,0.07825174],"study_design_scores_gemma":[0.002885084,0.0007100066,0.1460057,0.004761049,0.000332299,0.00008002915,0.4822993,0.07179786,0.03007465,0.06000087,0.1999984,0.001054763],"study_design_candidate":"qualitative","study_design_consensus":"qualitative","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9833472,0.002151461,0.00008980657,0.005415246,0.001282449,0.0005969223,0.0000952884,0.0002121178,0.006809464],"genre_scores_gemma":[0.9936768,0.0002691052,0.0004024411,0.0007351103,0.0003282057,0.00004488059,0.0005887766,0.00006487169,0.003889746],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.07719698,"threshold_uncertainty_score":0.9999796,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"methods","study_design":"theoretical_or_conceptual","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","research_integrity","scholarly_communication"],"domain":"evaluation","study_design":"theoretical_or_conceptual","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W1930603020","doi":"10.1371/journal.pone.0133639","title":"A Scoping Review of Empirical Research Relating to Quality and Effectiveness of Research Ethics Review","year":2015,"lang":"en","type":"review","venue":"PLoS ONE","topic":"Ethics in medical practice","field":"Health Professions","cited_by":82,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Western University; Ottawa Public Health; University of British Columbia; Ottawa Hospital; University of Ottawa","funders":"Canadian Institutes of Health Research","keywords":"Research ethics; Quality (philosophy); Engineering ethics; Empirical research; Management science; Medicine; Data science; Computer science; Engineering; Epistemology; Philosophy","retraction":null,"screen_n_in":null,"score":{"opus":0.9603257086203676,"gpt":0.8109373013019492,"spread":0.1493884073184184,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","metaepi_narrow","research_integrity"],"consensus_categories":["metaresearch","research_integrity"],"category_scores_codex":[0.6080248,0.0003779302,0.006849281,0.0005508049,0.001025656,0.000008893285,0.001058442,0.002194005,0.0002088463],"category_scores_gemma":[0.8074649,0.0003021323,0.0002355348,0.003754603,0.001235064,0.0001355851,0.002519712,0.04944252,0.0003197608],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000798576,"about_ca_system_score_gemma":0.01439429,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0003415441,"about_ca_topic_score_gemma":0.00004080516,"domain_scores_codex":[0.5903288,0.3918098,0.005409218,0.001262814,0.009890146,0.001299242],"domain_scores_gemma":[0.5025173,0.4776269,0.00199225,0.002199227,0.01486338,0.0008008763],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"systematic_review","study_design_gemma":"systematic_review","study_design_scores_codex":[0.00002960934,0.0003860552,0.00006823581,0.9795656,0.0001869378,0.000005231914,0.0006068138,2.554248e-9,0.000002243117,0.001860642,0.001852254,0.01543636],"study_design_scores_gemma":[0.0001627247,0.0002498677,0.0000090284,0.9228745,0.0007975773,0.000001618529,0.000125797,2.007182e-7,0.000001382182,0.001998007,0.07356798,0.0002112856],"study_design_candidate":"systematic_review","study_design_consensus":"systematic_review","genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.000008011612,0.9640253,0.0000127877,0.006569544,0.00009633834,0.01832822,0.00003976961,0.00003529495,0.01088479],"genre_scores_gemma":[0.000001330442,0.9931864,0.002414811,0.001513147,0.0002491575,0.002308093,0.00003759289,0.00009903431,0.0001904286],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.19944,"threshold_uncertainty_score":0.9999431,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"evaluation","study_design":"systematic_review","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","research_integrity"],"domain":"methods","study_design":"systematic_review","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"agree"},{"id":"W2948251768","doi":"10.1007/s11948-019-00112-4","title":"Misconduct and Misbehavior Related to Authorship Disagreements in Collaborative Science","year":2019,"lang":"en","type":"article","venue":"Science and Engineering Ethics","topic":"Academic integrity and plagiarism","field":"Social Sciences","cited_by":81,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal","funders":"National Center for Advancing Translational Sciences; Fonds de Recherche du Québec - Santé; National Institutes of Health","keywords":"Scientific misconduct; Psychology; Multidisciplinary approach; Hostility; Research integrity; Social psychology; Social science; Public relations; Sociology; Political science; Medicine; Alternative medicine","retraction":null,"screen_n_in":null,"score":{"opus":0.03739747619945206,"gpt":0.3625665209387409,"spread":0.3251690447392889,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01143263,0.00008294264,0.0001027707,0.0002810225,0.0005252549,0.0001860599,0.0003505719,0.0003082743,0.00001737405],"category_scores_gemma":[0.004117061,0.00007853215,0.00000642268,0.002440132,0.001244848,0.0006901373,0.000102273,0.001335995,0.00001677243],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001507781,"about_ca_system_score_gemma":0.0009801249,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0006809524,"about_ca_topic_score_gemma":0.0002091248,"domain_scores_codex":[0.9981682,0.00005118117,0.0001318737,0.000342516,0.0008706301,0.000435623],"domain_scores_gemma":[0.9991176,0.0002280847,0.00002436515,0.000102098,0.0002650265,0.0002628526],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"observational","study_design_scores_codex":[0.00001037516,0.00002631635,0.04812482,0.00005326291,0.000003333169,0.000007246331,0.4654069,0.000506114,0.09536622,0.3873597,0.00009150758,0.003044151],"study_design_scores_gemma":[0.002260427,0.0007538129,0.5986566,0.002299466,0.00005511915,0.00002909281,0.2068031,0.01945039,0.0267264,0.02623115,0.1132205,0.003513903],"study_design_candidate":"qualitative","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9903194,0.00007497041,0.00002098553,0.004725791,0.0004804177,0.0002364427,0.000003248356,0.00002873336,0.004110056],"genre_scores_gemma":[0.9983009,0.00006755073,0.000803302,0.0002412898,0.00001452268,0.000004687176,2.812226e-7,0.000003485286,0.0005639857],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5505318,"threshold_uncertainty_score":0.5804309,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"incentives","study_design":"theoretical_or_conceptual","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","research_integrity"],"domain":"methods","study_design":"design_other","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"split"},{"id":"W1951343740","doi":"10.1002/asi.23421","title":"Costly collaborations: The impact of scientific fraud on co‐authors' careers","year":2015,"lang":"en","type":"article","venue":"Journal of the Association for Information Science and Technology","topic":"Academic integrity and plagiarism","field":"Social Sciences","cited_by":81,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Université du Québec à Montréal; Université de Montréal","funders":"Canada Research Chairs","keywords":"Misconduct; Scientific misconduct; Productivity; Ranking (information retrieval); Psychology; Research integrity; Rigour; Political science; Public relations; Medicine; Computer science; Law; Alternative medicine; Economics; Epistemology","retraction":null,"screen_n_in":null,"score":{"opus":0.03220766005286423,"gpt":0.3526338973852482,"spread":0.320426237332384,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.01506883,0.00004744861,0.0001098991,0.000484917,0.001137311,0.000228289,0.0006604445,0.0002738978,0.000002339367],"category_scores_gemma":[0.02247673,0.00002580351,0.00005767532,0.002453999,0.0009203951,0.001587831,0.000033604,0.0004403762,0.00000495924],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007309431,"about_ca_system_score_gemma":0.002358588,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00007716806,"about_ca_topic_score_gemma":0.00003258243,"domain_scores_codex":[0.9981709,0.00008578545,0.0003651787,0.00004687895,0.001155647,0.0001756308],"domain_scores_gemma":[0.9931308,0.0002831606,0.001181709,0.0001093915,0.005234896,0.00006003885],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00009976016,0.00007651366,0.1038505,0.000008577986,0.0000995259,1.269579e-7,0.08364552,0.0008015722,0.0009403002,0.5731024,0.2232153,0.01415992],"study_design_scores_gemma":[0.002074982,0.0009026371,0.01958446,0.00009317099,0.00008232368,0.00001576603,0.1483107,0.00153213,0.009549761,0.09466316,0.7229313,0.0002596481],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9447411,0.00005563021,0.0002546347,0.03762115,0.001497457,0.000504414,0.00006623453,0.00002006367,0.01523928],"genre_scores_gemma":[0.9994891,0.00002127813,0.00008568891,0.0001425898,0.00004732206,0.000003298298,0.00000105208,0.000001060735,0.0002085543],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.499716,"threshold_uncertainty_score":0.9857574,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"incentives","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","research_integrity"],"domain":"incentives","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"agree"},{"id":"W3213227864","doi":"10.1371/journal.pone.0258646","title":"Return of individual research results from genomic research: A systematic review of stakeholder perspectives","year":2021,"lang":"en","type":"review","venue":"PLoS ONE","topic":"Ethics in Clinical Research","field":"Medicine","cited_by":78,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"McGill University; Montreal Neurological Institute and Hospital; University of Calgary","funders":"Medical Research Future Fund; Wellcome Trust; National Center for Advancing Translational Sciences; Economic and Social Research Council; National Institutes of Health; Norges Forskningsråd; National Heart, Lung, and Blood Institute; Australian Government; Wellcome; Canadian Institutes of Health Research; Genome Canada; Medical Research Council","keywords":"Biobank; Stakeholder; Qualitative research; Empirical research; Systematic review; Psychology; MEDLINE; Medicine; Public relations; Social science; Political science; Bioinformatics; Biology; Sociology","retraction":null,"screen_n_in":null,"score":{"opus":0.9465482916184245,"gpt":0.6397066401700148,"spread":0.3068416514484097,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","metaepi_narrow","research_integrity"],"consensus_categories":["metaresearch","research_integrity"],"category_scores_codex":[0.09578559,0.0005142905,0.01053237,0.001408191,0.0001526945,0.00007066526,0.002311282,0.001576309,0.0006481032],"category_scores_gemma":[0.4523728,0.0003921409,0.001106032,0.003417833,0.002053227,0.00006890827,0.002217839,0.01541854,0.000193284],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008822845,"about_ca_system_score_gemma":0.009142973,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002021543,"about_ca_topic_score_gemma":0.00004234263,"domain_scores_codex":[0.9591183,0.01492144,0.006000235,0.001924112,0.0168216,0.001214278],"domain_scores_gemma":[0.7888857,0.1787715,0.002378964,0.007654014,0.0214759,0.0008338739],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"systematic_review","study_design_gemma":"systematic_review","study_design_scores_codex":[0.00008705129,0.003023909,0.000001818634,0.9863025,0.00553112,0.0001850188,0.001443167,2.950215e-9,0.00004256004,0.0008794579,0.001016355,0.001487043],"study_design_scores_gemma":[0.0006722033,0.0007613972,0.000004765266,0.9859289,0.00680729,0.00001158313,0.002038957,7.14215e-7,0.00008017965,0.00085401,0.002592636,0.0002473699],"study_design_candidate":"systematic_review","study_design_consensus":"systematic_review","genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.0001897067,0.9835898,3.811664e-7,0.001079903,0.00003350788,0.00917421,0.001977341,0.0000272905,0.003927805],"genre_scores_gemma":[0.00006003823,0.9916972,0.001588693,0.00004633659,0.000415406,0.0006451468,0.0007649739,0.0001725471,0.004609656],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.3565873,"threshold_uncertainty_score":0.9998531,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"methods","study_design":"systematic_review","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":[],"domain":null,"study_design":"systematic_review","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W1969795835","doi":"10.1145/1227310.1227324","title":"Plagiarism detection using feature-based neural networks","year":2007,"lang":"en","type":"article","venue":"","topic":"Academic integrity and plagiarism","field":"Social Sciences","cited_by":75,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"","keywords":"Plagiarism detection; Computer science; Feature (linguistics); Relevance (law); Detector; Artificial neural network; Similarity (geometry); Artificial intelligence; Feature engineering; Code (set theory); Similarity measure; Feature extraction; Measure (data warehouse); Pattern recognition (psychology); Machine learning; Data mining; Deep learning; Programming language","retraction":null,"screen_n_in":null,"score":{"opus":0.02383288585388381,"gpt":0.3070886043220694,"spread":0.2832557184681856,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002452533,0.00009363863,0.00009666618,0.00007852547,0.000715593,0.00005964297,0.0001640747,0.00126747,0.0001945828],"category_scores_gemma":[0.0002064851,0.00008664904,0.00006747726,0.0003289897,0.0001515271,0.0001821843,0.00001408086,0.001710086,0.00001238635],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001432949,"about_ca_system_score_gemma":0.0000620642,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00745536,"about_ca_topic_score_gemma":0.007079286,"domain_scores_codex":[0.9987832,0.0001461457,0.000138822,0.0001657557,0.0003158797,0.0004501967],"domain_scores_gemma":[0.9993285,0.0003102544,0.00006299621,0.00008368657,0.00008025499,0.0001343184],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001100178,0.000340942,0.05598838,0.00004624467,0.0001567172,0.0002338711,0.02209204,0.07192818,0.01051118,0.1639439,0.02233171,0.6513266],"study_design_scores_gemma":[0.0005912604,0.00006350077,0.003806548,0.00002617864,0.00004995465,0.00001184466,0.004466736,0.8310791,0.006111513,0.002086733,0.1510987,0.0006079465],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1276132,0.0001034087,0.8149264,0.001150742,0.001423162,0.0001690488,7.59314e-7,0.0002041872,0.05440903],"genre_scores_gemma":[0.9938959,0.000005602993,0.001944957,0.001675096,0.001080118,7.769521e-7,0.000002293808,0.000007820102,0.001387472],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8662826,"threshold_uncertainty_score":0.9991541,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["research_integrity"],"domain":null,"study_design":"simulation_or_modeling","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":[],"domain":null,"study_design":"bench_or_experimental","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"split"},{"id":"W1588909286","doi":"10.1080/08989620590957175","title":"Haunted Manuscripts: Ghost Authorship in the Medical Literature","year":2005,"lang":"en","type":"article","venue":"Accountability in Research","topic":"Pharmaceutical industry and healthcare","field":"Pharmacology, Toxicology and Pharmaceutics","cited_by":75,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto; Queen's University; Institute for Clinical Evaluative Sciences","funders":"","keywords":"Credibility; Conflict of interest; Accountability; Psychology; Neutrality; Clinical trial; Public relations; Political science; Law; Medicine","retraction":null,"screen_n_in":null,"score":{"opus":0.8261096242099695,"gpt":0.7045818707680983,"spread":0.1215277534418712,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","research_integrity","insufficient_payload"],"consensus_categories":["research_integrity"],"category_scores_codex":[0.03432653,0.0002426459,0.0003180834,0.0003882947,0.0003948528,0.000119093,0.001531168,0.001858722,0.006836003],"category_scores_gemma":[0.003957538,0.0001785863,0.00008229762,0.002200898,0.0009073133,0.0003924698,0.0002771817,0.01689769,0.0005280765],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005197475,"about_ca_system_score_gemma":0.00049232,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0007046929,"about_ca_topic_score_gemma":0.005496151,"domain_scores_codex":[0.9871873,0.008462169,0.000761406,0.0006318362,0.001401137,0.001556169],"domain_scores_gemma":[0.9937984,0.004660639,0.00004547652,0.000593142,0.0003236327,0.0005786509],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"not_applicable","study_design_scores_codex":[0.003156754,0.005259734,0.8003356,0.0007300302,0.00005722278,0.001432952,0.03219696,0.0001043164,0.0007604074,0.01866697,0.05388927,0.08340971],"study_design_scores_gemma":[0.002812216,0.0001635892,0.1926813,0.0002487519,0.00001361503,0.0001415948,0.00479177,0.005919864,0.001110547,0.01010797,0.781616,0.0003927413],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8451093,0.001776189,0.000001168905,0.1365994,0.0003669391,0.0009717062,0.00005424096,0.00004199779,0.01507904],"genre_scores_gemma":[0.9813944,0.0004286747,0.00001963349,0.01652098,0.000811865,0.0002948576,0.00004629389,0.00001675305,0.0004665912],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7277268,"threshold_uncertainty_score":0.9994371,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"methods","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["research_integrity","scholarly_communication"],"domain":null,"study_design":"design_other","genre":"other","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"},{"id":"W3166045557","doi":"10.1186/s13643-021-01733-2","title":"Dealing with predatory journal articles captured in systematic reviews","year":2021,"lang":"en","type":"letter","venue":"Systematic Reviews","topic":"Meta-analysis and systematic reviews","field":"Decision Sciences","cited_by":73,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Ottawa Public Health; University of Ottawa; McGill University; Ottawa Hospital","funders":"","keywords":"Systematic review; Medicine; MEDLINE; Protocol (science); Health care; Peer review; Alternative medicine; Family medicine; Pathology","retraction":null,"screen_n_in":null,"score":{"opus":0.6956962901616436,"gpt":0.4789212342135676,"spread":0.216775055948076,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","metaepi_narrow","metaepi_broad","scholarly_communication","open_science","research_integrity","insufficient_payload"],"consensus_categories":["metaresearch","metaepi_narrow","metaepi_broad","insufficient_payload"],"category_scores_codex":[0.4841305,0.002491905,0.05974738,0.001913323,0.000353179,0.005570516,0.006777234,0.001110448,0.007079146],"category_scores_gemma":[0.2566205,0.0009285524,0.01031699,0.003990494,0.000138609,0.0007370506,0.0003202227,0.0035758,0.01508175],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004876833,"about_ca_system_score_gemma":0.0004394086,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001554734,"about_ca_topic_score_gemma":0.0001401407,"domain_scores_codex":[0.644936,0.2415301,0.09223322,0.003344249,0.01651797,0.001438427],"domain_scores_gemma":[0.8575884,0.0240444,0.09479874,0.01923543,0.003749213,0.0005837874],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"systematic_review","study_design_gemma":"not_applicable","study_design_scores_codex":[0.000001317445,0.00002346417,0.00008222247,0.5130781,0.0004489168,0.0005486399,0.0006755829,0.00001167409,0.000003222146,0.00005189954,0.4849653,0.0001096706],"study_design_scores_gemma":[0.0002542276,0.00004457154,0.000006473944,0.3885653,0.00292436,0.002128106,0.0008664678,0.0004058942,0.000001633261,0.0005630572,0.6033633,0.0008766351],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"review","genre_gemma":"commentary","genre_scores_codex":[0.0001615175,0.8112237,0.02274858,0.05507779,0.003116369,0.1008147,0.00004659493,0.00003377018,0.006776995],"genre_scores_gemma":[0.003336054,0.05326024,0.04778257,0.5993811,0.01242661,0.04872129,0.0003560355,0.001293331,0.2334428],"genre_candidate":"review","genre_consensus":null,"teacher_disagreement_score":0.7579635,"threshold_uncertainty_score":0.9995944,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","scholarly_communication"],"domain":"methods","study_design":"not_applicable","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","research_integrity"],"domain":"methods","study_design":"theoretical_or_conceptual","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W2107418313","doi":"10.1371/journal.pmed.1001406","title":"Policies and Initiatives Aimed at Addressing Research Misconduct in High-Income Countries","year":2013,"lang":"en","type":"letter","venue":"PLoS Medicine","topic":"Academic integrity and plagiarism","field":"Social Sciences","cited_by":71,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"National Institute of Environmental Health Sciences; Health Canada; Research Councils UK; National Institutes of Health; National Science Foundation","keywords":"Misconduct; Scientific misconduct; High income countries; Research integrity; Political science; Developing country; Public relations; Medicine; Psychology; Economic growth; Law; Alternative medicine; Economics; Pathology","retraction":null,"screen_n_in":null,"score":{"opus":0.175147443283836,"gpt":0.4147393244350316,"spread":0.2395918811511956,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["sts","research_integrity","insufficient_payload"],"consensus_categories":["research_integrity"],"category_scores_codex":[0.003539074,0.0002831012,0.0007299602,0.000685634,0.0008453421,0.00008496534,0.0005331307,0.002563853,0.002400054],"category_scores_gemma":[0.00497316,0.0002173163,0.00002988664,0.0004541031,0.00476356,0.0003772844,0.0001976724,0.008643341,0.0001597652],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0004433721,"about_ca_system_score_gemma":0.0003623668,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.04130245,"about_ca_topic_score_gemma":0.002412538,"domain_scores_codex":[0.9944239,0.001833949,0.0004847689,0.0004552583,0.001846573,0.0009555485],"domain_scores_gemma":[0.9940858,0.0049365,0.000201633,0.0002306638,0.000395019,0.0001503795],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00001354378,0.000007512983,0.004241462,0.000304173,0.00004699385,0.0001364891,0.05935288,3.276666e-8,0.00007189712,0.002486861,0.9332545,0.00008369761],"study_design_scores_gemma":[0.0005157662,0.00008815138,0.004612332,0.003750851,0.00004520237,0.000006246373,0.008066765,0.000004819348,0.0001018221,0.00675955,0.975727,0.0003215108],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"commentary","genre_gemma":"commentary","genre_scores_codex":[0.1089971,0.00465672,5.211765e-7,0.8552002,0.0004328917,0.0005467175,0.00003911866,0.00005203751,0.03007467],"genre_scores_gemma":[0.2138923,0.02291935,0.0001262147,0.6849661,0.03500445,0.0002169825,0.000312227,0.0001153358,0.04244695],"genre_candidate":"commentary","genre_consensus":"commentary","teacher_disagreement_score":0.1702341,"threshold_uncertainty_score":0.998731,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"evaluation","study_design":"not_applicable","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","research_integrity"],"domain":"methods","study_design":"not_applicable","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"agree"},{"id":"W2106029785","doi":"10.1177/1740774508089459","title":"Maintaining confidentiality of interim data to enhance trial integrity and credibility","year":2008,"lang":"en","type":"article","venue":"Clinical Trials","topic":"Ethics in Clinical Research","field":"Medicine","cited_by":71,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Health Research Council of New Zealand; National Institute of Allergy and Infectious Diseases; Canadian Institutes of Health Research","keywords":"Interim; Safeguarding; Interim analysis; Credibility; Confidentiality; Data monitoring committee; Harm; Clinical trial; Medicine; Psychology; Computer science; Computer security; Political science; Social psychology; Nursing; Law","retraction":null,"screen_n_in":null,"score":{"opus":0.9540912610429189,"gpt":0.7680111871722448,"spread":0.186080073870674,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","research_integrity"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.2117597,0.0001854988,0.002470759,0.00008255448,0.00008958994,0.00002637708,0.0008474133,0.0007868806,0.0005672659],"category_scores_gemma":[0.796498,0.0001439068,0.0003916056,0.0002372634,0.001842845,0.0001129621,0.001741062,0.00424773,0.00004976436],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004157194,"about_ca_system_score_gemma":0.001366384,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002972572,"about_ca_topic_score_gemma":0.0001295504,"domain_scores_codex":[0.9832373,0.007013449,0.006416992,0.001536265,0.001355242,0.0004407687],"domain_scores_gemma":[0.4495883,0.5317602,0.002708296,0.010136,0.003034247,0.002772931],"domain_codex":null,"domain_gemma":"methods","domain_candidate":null,"domain_consensus":null,"study_design_codex":"randomized_trial","study_design_gemma":"observational","study_design_scores_codex":[0.7040665,0.008812128,0.1265687,0.001434984,0.001711172,0.0002841157,0.001247025,9.539903e-7,0.009714678,0.01232313,0.02968515,0.1041515],"study_design_scores_gemma":[0.3312933,0.0537416,0.3566649,0.005611424,0.002099153,0.0001626854,0.002123618,0.001149236,0.03298901,0.1700676,0.04202836,0.002068998],"study_design_candidate":"randomized_trial","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9842522,0.0001371399,0.002951343,0.00761163,0.001501599,0.002038819,0.0001573552,0.00004683674,0.001303114],"genre_scores_gemma":[0.9881817,0.001223334,0.006642829,0.001346185,0.002012283,0.00002244527,0.00003734258,0.00001853868,0.0005154078],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5847383,"threshold_uncertainty_score":0.9980495,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"methods","study_design":"not_applicable","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"high"},{"model":"gpt","categories":["metaresearch"],"domain":"methods","study_design":"not_applicable","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W2474558075","doi":"10.1302/2046-3758.56.bjr-2016-0047","title":"Retractions in orthopaedic research","year":2016,"lang":"en","type":"article","venue":"Bone and Joint Research","topic":"Academic integrity and plagiarism","field":"Social Sciences","cited_by":67,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Juravinski Cancer Centre; Hamilton Health Sciences; McMaster University","funders":"","keywords":"Scientific misconduct; Scrutiny; Misconduct; Impact factor; Medicine; Cochrane Library; MEDLINE; Data extraction; Meta-analysis; Alternative medicine; Political science; Pathology; Law","retraction":null,"screen_n_in":null,"score":{"opus":0.257422106957606,"gpt":0.4771097084864145,"spread":0.2196876015288086,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","research_integrity"],"consensus_categories":[],"category_scores_codex":[0.03282303,0.00005063262,0.0001230433,0.0006476461,0.0009934165,0.00006965034,0.0001210469,0.0004726849,0.00079377],"category_scores_gemma":[0.007145543,0.00003513767,0.00003283239,0.0009348821,0.001176571,0.0002629986,0.0001224541,0.002708121,0.0003801442],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001909355,"about_ca_system_score_gemma":0.0005055295,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01486594,"about_ca_topic_score_gemma":0.0102258,"domain_scores_codex":[0.99546,0.001913537,0.0002119749,0.0002620417,0.001366584,0.0007858348],"domain_scores_gemma":[0.9973536,0.001851108,0.00001832658,0.0001512746,0.0004183694,0.0002073424],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0001092525,0.0001546489,0.03266975,0.00003372767,0.0000105557,0.0001668056,0.01672745,7.076844e-8,0.01002748,0.5528023,0.1881237,0.1991743],"study_design_scores_gemma":[0.0003627955,0.00008834717,0.0301172,0.000211344,0.000001017796,0.000007733132,0.01115922,0.000005902195,0.000366308,0.07882831,0.8787443,0.0001075586],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5721458,0.0008645496,0.0000550042,0.09889764,0.0001316241,0.0004330394,0.000007153319,0.0000382339,0.327427],"genre_scores_gemma":[0.9616273,0.004837307,0.00007495941,0.00004948468,0.0003835338,0.00003253341,6.888166e-7,0.000005913338,0.03298825],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6906205,"threshold_uncertainty_score":0.9995927,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"evaluation","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","research_integrity"],"domain":"evaluation","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"agree"},{"id":"W2600935272","doi":"10.1007/s11948-017-9883-5","title":"America COMPETES at 5 years: An Analysis of Research-Intensive Universities’ RCR Training Plans","year":2017,"lang":"en","type":"article","venue":"Science and Engineering Ethics","topic":"Human Resource Development and Performance Evaluation","field":"Psychology","cited_by":67,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"University of California, Los Angeles; University of Illinois at Urbana-Champaign; Stony Brook University; Oregon State University; Montana State University; Georgia State University; Dartmouth College; York University; University of Rochester; Georgetown University; University of Central Florida; Johns Hopkins University; Brandeis University; Colorado State University; University of Connecticut; Virginia Commonwealth University; Tulane University; Brown University; Arizona State University; State University of New York; Louisiana State University; George Washington University; California Institute of Technology; Georgia Institute of Technology; Ohio State University; Rice University; Rensselaer Polytechnic Institute; Florida State University; Yale University; Vanderbilt University; University of Oklahoma; Washington University in St. Louis; University at Buffalo; Massachusetts Institute of Technology; City University of New York; Iowa State University; Carnegie Mellon University; North Carolina State University; Purdue University; Case Western Reserve University; National Science Foundation; Mississippi State University; Emory University; Harvard University","keywords":"Philosophy of science; Training (meteorology); Medical education; Business; Political science; Medicine; Geography; Philosophy","retraction":null,"screen_n_in":null,"score":{"opus":0.2363368745113985,"gpt":0.4519432389538049,"spread":0.2156063644424065,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002482126,0.00005052263,0.0001170437,0.0005143341,0.0005780847,0.00006614458,0.0003094681,0.00006273286,0.0001017342],"category_scores_gemma":[0.0003255578,0.00005117345,0.00001829567,0.0003922785,0.0009269785,0.000193104,0.00008870356,0.0002866002,0.000009575394],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00006437402,"about_ca_system_score_gemma":0.0001379146,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001547858,"about_ca_topic_score_gemma":0.00007487998,"domain_scores_codex":[0.9990004,0.00003196484,0.00008456288,0.0001761436,0.0004796765,0.0002272472],"domain_scores_gemma":[0.9990405,0.0001870511,0.00004883738,0.0002665472,0.000385286,0.00007171042],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"observational","study_design_scores_codex":[0.00005064521,0.00002329106,0.04442591,0.00004867624,0.0003023604,0.00002492834,0.9087532,0.0113056,0.007785372,0.01954384,0.0002109263,0.007525239],"study_design_scores_gemma":[0.0001402819,0.00008267967,0.8861794,0.00003742521,0.0000463293,0.00000165469,0.04754082,0.06162584,0.0001071944,0.00002635282,0.004100545,0.0001114023],"study_design_candidate":"qualitative","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9968814,0.00001526759,0.0000946418,0.0002061236,0.0001326797,0.00003427257,0.00000875803,0.00001737748,0.002609459],"genre_scores_gemma":[0.999228,0.00001453288,0.0002694796,0.00005124186,0.00002571174,0.000001479999,0.000009014037,0.000003679211,0.0003968471],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.8612124,"threshold_uncertainty_score":0.4446219,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch"],"domain":"incentives","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","research_integrity"],"domain":"methods","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W1999935843","doi":"10.1136/bmj.38968.611296.f7","title":"Role of systematic reviews in detecting plagiarism: case of Asim Kurjak","year":2006,"lang":"en","type":"review","venue":"BMJ","topic":"Academic integrity and plagiarism","field":"Social Sciences","cited_by":67,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Sveučilište u Zagrebu; Ryerson University","keywords":"Sanctions; Audience measurement; Face (sociological concept); Systematic review; Psychology; Political science; Internet privacy; Computer science; Sociology; Law; Social science; MEDLINE","retraction":null,"screen_n_in":null,"score":{"opus":0.1071014461526957,"gpt":0.4176863635685728,"spread":0.3105849174158771,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["research_integrity"],"consensus_categories":[],"category_scores_codex":[0.01368724,0.0002627364,0.003189482,0.0002548545,0.0001091331,0.00001616936,0.0004960718,0.002053035,0.00005323018],"category_scores_gemma":[0.007421446,0.0002053652,0.0005457435,0.0006805941,0.0001398898,0.00008725431,0.0000606547,0.001947811,0.00004638594],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001704639,"about_ca_system_score_gemma":0.0004569276,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01713082,"about_ca_topic_score_gemma":0.006672072,"domain_scores_codex":[0.9930174,0.003475046,0.002550817,0.0002558497,0.0003865449,0.0003142868],"domain_scores_gemma":[0.994943,0.002075009,0.002481266,0.0003413061,0.0001002241,0.00005925843],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"systematic_review","study_design_gemma":"not_applicable","study_design_scores_codex":[0.000001732791,0.00004313855,0.00001051294,0.5624787,0.00005320389,0.0001393666,0.006151502,0.000001112219,5.810045e-7,0.003527644,0.002669628,0.4249229],"study_design_scores_gemma":[0.00004647675,0.00001123995,2.402239e-7,0.21595,0.0003962795,0.0001859639,0.002670081,0.0000107536,0.000001470746,0.0003883599,0.7801188,0.0002203768],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.00001828358,0.9706821,0.00002228854,0.00002188749,0.0001857461,0.003505893,0.00002711568,0.00001490824,0.02552179],"genre_scores_gemma":[0.0002328523,0.9982618,0.0003448763,0.00001095202,0.0003170352,0.0001948757,0.00000761601,0.00001844521,0.0006115281],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.7774492,"threshold_uncertainty_score":0.9992425,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"evaluation","study_design":"systematic_review","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"high"},{"model":"gpt","categories":["metaresearch","research_integrity"],"domain":"methods","study_design":"systematic_review","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"agree"},{"id":"W2948845587","doi":"10.1007/s11948-019-00113-3","title":"Researchers’ Perceptions of Ethical Authorship Distribution in Collaborative Research Teams","year":2019,"lang":"en","type":"article","venue":"Science and Engineering Ethics","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":67,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal","funders":"National Center for Advancing Translational Sciences; Fonds de Recherche du Québec - Santé; National Institutes of Health","keywords":"Transparency (behavior); Psychology; Mentorship; Productivity; Affect (linguistics); Perception; Public relations; Engineering ethics; Political science; Medical education; Medicine; Engineering","retraction":null,"screen_n_in":null,"score":{"opus":0.6652650663412806,"gpt":0.638300049848536,"spread":0.02696501649274452,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","research_integrity"],"consensus_categories":["metaresearch","bibliometrics"],"category_scores_codex":[0.217933,0.00008940783,0.0002166197,0.02189431,0.0003291951,0.0009881103,0.001594246,0.0004426679,0.0000812314],"category_scores_gemma":[0.2452208,0.00006818251,0.00003137432,0.2043707,0.001623228,0.000575219,0.000747631,0.003783335,0.00008458319],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003388367,"about_ca_system_score_gemma":0.001936472,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002371044,"about_ca_topic_score_gemma":0.00004307048,"domain_scores_codex":[0.9811432,0.0006162938,0.0004797746,0.0007141858,0.01617103,0.0008755482],"domain_scores_gemma":[0.9689204,0.01672061,0.00007526873,0.000733106,0.0129708,0.0005798555],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"observational","study_design_scores_codex":[0.00008980125,0.0002899559,0.3097545,0.0001891896,0.00001133281,0.00001808666,0.02827841,0.006194236,0.2320815,0.3857255,0.003637594,0.03372987],"study_design_scores_gemma":[0.0005231964,0.0005225927,0.8404738,0.0001374146,0.00000155956,0.000005326934,0.01407483,0.1084145,0.004155393,0.01883186,0.01255476,0.0003047152],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9902124,0.0001494009,0.0009585546,0.006231602,0.0002274142,0.0002695744,0.00003266694,0.00001271602,0.001905719],"genre_scores_gemma":[0.9986612,0.0001876591,0.0006199774,0.00003142942,0.00001906328,0.00000895633,0.000002343059,0.00000505187,0.000464345],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5307194,"threshold_uncertainty_score":0.998515,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"incentives","study_design":"qualitative","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","research_integrity"],"domain":"methods","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"split"},{"id":"W2484305216","doi":"10.1136/bmjopen-2016-011997","title":"Reporting of financial and non-financial conflicts of interest by authors of systematic reviews: a methodological survey","year":2016,"lang":"en","type":"article","venue":"BMJ Open","topic":"Pharmaceutical industry and healthcare","field":"Pharmacology, Toxicology and Pharmaceutics","cited_by":65,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"McMaster University; University of Toronto","funders":"Faculty of Medicine, American University of Beirut; American University of Beirut","keywords":"Conflict of interest; Systematic review; Medicine; MEDLINE; Public interest; Cochrane Library; Meta-analysis; Actuarial science; Family medicine; Finance; Political science; Internal medicine; Economics; Law","retraction":null,"screen_n_in":null,"score":{"opus":0.939984580804016,"gpt":0.6974284889577338,"spread":0.2425560918462822,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.03483582,0.0002067763,0.001903,0.00005993818,0.00007780503,0.000007486102,0.0005291218,0.0006481093,0.0003274097],"category_scores_gemma":[0.05397232,0.0001318509,0.0001251468,0.0002189899,0.0004228776,0.0001214018,0.0004462437,0.0005884055,0.000007743034],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000267617,"about_ca_system_score_gemma":0.000280456,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001769178,"about_ca_topic_score_gemma":0.00007555378,"domain_scores_codex":[0.9895977,0.005773147,0.003808334,0.0003335346,0.0001439988,0.0003433347],"domain_scores_gemma":[0.9907807,0.003863775,0.00442657,0.0003276769,0.00031504,0.0002862773],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.008660629,0.001389367,0.395678,0.1209753,0.0003931514,0.000122198,0.002666005,0.000003291953,0.3905271,0.008646606,0.04115714,0.0297812],"study_design_scores_gemma":[0.009259833,0.002498292,0.2919299,0.0416543,0.0009675447,0.0001971474,0.0002479962,0.0002717145,0.6229917,0.002768375,0.02592089,0.00129227],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9863631,0.001154673,0.0004489401,0.0009271022,0.0006123156,0.009046497,0.0004168339,0.00001180331,0.001018712],"genre_scores_gemma":[0.9975864,0.0003143692,0.0004907335,0.0005932662,0.00006806237,0.0003993767,0.000009387551,0.0000113935,0.0005270149],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2324647,"threshold_uncertainty_score":0.9938396,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"reporting","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"},{"model":"gpt","categories":["metaresearch","research_integrity"],"domain":"reporting","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"agree"},{"id":"W3081363283","doi":"10.1371/journal.pone.0237875","title":"Exploring the ethical issues in research using digital data collection strategies with minors: A scoping review","year":2020,"lang":"en","type":"review","venue":"PLoS ONE","topic":"Focus Groups and Qualitative Methods","field":"Social Sciences","cited_by":59,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Western University","funders":"","keywords":"Gatekeeping; Scrutiny; Harm; Data collection; Research ethics; Informed consent; Social research; Engineering ethics; Social media; Psychology; Public relations; Data science; Computer science; Sociology; Political science; Medicine; Social psychology; Social science; Law; Alternative medicine; Engineering","retraction":null,"screen_n_in":null,"score":{"opus":0.9439772789083111,"gpt":0.6428719519843116,"spread":0.3011053269239995,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01103904,0.000228033,0.001442203,0.0001553001,0.0006626025,0.0006521321,0.0009584909,0.0001740981,0.00003130638],"category_scores_gemma":[0.006589953,0.0001449302,0.0000777622,0.002749114,0.0006572053,0.0009474229,0.0003861453,0.001970867,0.00002100604],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002514004,"about_ca_system_score_gemma":0.002151625,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001814115,"about_ca_topic_score_gemma":0.001872549,"domain_scores_codex":[0.9896591,0.00729348,0.0005856164,0.0006026993,0.001408548,0.0004505412],"domain_scores_gemma":[0.995269,0.003668163,0.000207414,0.000536835,0.0002201432,0.00009842798],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"systematic_review","study_design_scores_codex":[0.00001979529,0.0003038839,0.000006554644,0.250417,0.000364128,0.00004410913,0.0219132,1.934538e-7,7.360489e-7,0.006038484,0.0002272448,0.7206647],"study_design_scores_gemma":[0.00004816573,0.0000877656,4.391869e-7,0.7239957,0.000278948,0.00000114938,0.007788213,0.00001297836,5.440345e-7,0.0002389395,0.2673059,0.0002412182],"study_design_candidate":"systematic_review","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.00002523977,0.9931391,0.00008495745,0.002293997,0.00003670375,0.002401727,0.00001997166,0.00003771918,0.001960611],"genre_scores_gemma":[0.000003920351,0.9938037,0.005252831,0.00003636317,0.000437834,0.0002857409,0.00002992725,0.00003359674,0.0001161192],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.7204235,"threshold_uncertainty_score":0.8562546,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["research_integrity"],"domain":null,"study_design":"systematic_review","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"},{"model":"gpt","categories":[],"domain":null,"study_design":"systematic_review","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"},{"id":"W3003877329","doi":"10.1145/1227504.1227324","title":"Plagiarism detection using feature-based neural networks","year":2007,"lang":"en","type":"article","venue":"ACM SIGCSE Bulletin","topic":"Academic integrity and plagiarism","field":"Social Sciences","cited_by":58,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"","keywords":"Plagiarism detection; Feature (linguistics); Relevance (law); Computer science; Detector; Artificial neural network; Similarity (geometry); Artificial intelligence; Code (set theory); Feature engineering; Measure (data warehouse); Pattern recognition (psychology); Natural language processing; Machine learning; Data mining; Deep learning; Programming language; Linguistics","retraction":null,"screen_n_in":null,"score":{"opus":0.02276400814018628,"gpt":0.2937078578585832,"spread":0.270943849718397,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["research_integrity"],"consensus_categories":["research_integrity"],"category_scores_codex":[0.003364227,0.0001714602,0.0001744391,0.0001054986,0.0009298486,0.00008549496,0.000511174,0.001520884,0.0005739043],"category_scores_gemma":[0.001882436,0.0001731522,0.0001122974,0.0003335614,0.0002427333,0.00007544096,0.00006809972,0.002413736,0.00009312084],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001893251,"about_ca_system_score_gemma":0.000072898,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005230187,"about_ca_topic_score_gemma":0.001152847,"domain_scores_codex":[0.9980099,0.0003154592,0.0002374402,0.0002967288,0.0004618161,0.0006786991],"domain_scores_gemma":[0.9982427,0.001028002,0.000142482,0.0002706193,0.0001187093,0.0001974839],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.003119211,0.0007368583,0.03409246,0.0001185065,0.0003023529,0.0009296937,0.02970461,0.04937646,0.01213811,0.01873458,0.253836,0.5969112],"study_design_scores_gemma":[0.0006703114,0.00007302253,0.00160424,0.00004755921,0.00005682033,0.00001519145,0.001926803,0.02207999,0.003689965,0.001040263,0.9682235,0.000572364],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6236551,0.0008662242,0.3117082,0.02956501,0.004523175,0.0008673683,0.000007944262,0.0006760844,0.02813092],"genre_scores_gemma":[0.9906035,0.00001877358,0.00325556,0.003188491,0.001629478,0.000003415601,0.000006485292,0.00002023632,0.001274075],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7143875,"threshold_uncertainty_score":0.9998878,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["research_integrity"],"domain":null,"study_design":"bench_or_experimental","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"},{"model":"gpt","categories":[],"domain":null,"study_design":"design_other","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W2883405865","doi":"10.1177/1609406918788203","title":"Ethical, Practical, and Methodological Considerations for Unobtrusive Qualitative Research About Personal Narratives Shared on the Internet","year":2018,"lang":"en","type":"article","venue":"International Journal of Qualitative Methods","topic":"Focus Groups and Qualitative Methods","field":"Social Sciences","cited_by":57,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Saskatchewan","funders":"","keywords":"Qualitative research; Confidentiality; Engineering ethics; Popularity; The Internet; Ethical issues; Narrative; Diligence; Due diligence; Psychology; Informed consent; Research ethics; Internet research; Internet privacy; Public relations; Sociology; Social psychology; Political science; Computer science; Medicine; Engineering; Social science","retraction":null,"screen_n_in":null,"score":{"opus":0.9160692370572425,"gpt":0.7703199511312407,"spread":0.1457492859260018,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","sts","insufficient_payload"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.2028706,0.0002662569,0.0006177432,0.0004695307,0.001299744,0.0005183916,0.000576086,0.0003310335,0.001275189],"category_scores_gemma":[0.2785402,0.0001718066,0.0003129877,0.0004401239,0.007810379,0.0004776623,0.0001580547,0.001948488,0.0000160901],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000313098,"about_ca_system_score_gemma":0.0009307466,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001898817,"about_ca_topic_score_gemma":0.0001354939,"domain_scores_codex":[0.8531577,0.1431601,0.001065962,0.0004365251,0.001712665,0.0004669933],"domain_scores_gemma":[0.6989594,0.2894471,0.0009941151,0.0001552432,0.01017113,0.0002730425],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0005086416,0.00007683899,0.000003232749,0.000004681547,0.0002985417,0.00000595459,0.483637,2.18258e-7,0.0008657399,0.5016402,0.01080817,0.002150761],"study_design_scores_gemma":[0.0004324163,0.001187766,0.00005851683,0.00009501144,0.00001883331,0.00001500917,0.4531253,0.00005466014,0.001275246,0.4966537,0.04695153,0.0001319926],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.03198739,0.0003192892,0.6766323,0.281494,0.001432096,0.0006887108,0.000121691,0.00002047016,0.007304007],"genre_scores_gemma":[0.1322703,0.00009043547,0.8612159,0.004088166,0.001425858,0.00009691079,0.000003504309,0.0000243516,0.0007845469],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.2774059,"threshold_uncertainty_score":0.9996713,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["research_integrity"],"domain":null,"study_design":"theoretical_or_conceptual","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"high"},{"model":"gpt","categories":[],"domain":null,"study_design":"qualitative","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"},{"id":"W2465870777","doi":"10.1371/journal.pone.0157142","title":"The Challenge of Timely, Responsive and Rigorous Ethics Review of Disaster Research: Views of Research Ethics Committee Members","year":2016,"lang":"en","type":"article","venue":"PLoS ONE","topic":"Ethics in Clinical Research","field":"Medicine","cited_by":56,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Holland Bloorview Kids Rehabilitation Hospital; McMaster University; Centre for Interdisciplinary Research in Rehabilitation; McGill University","funders":"Canadian Institutes of Health Research; McMaster University","keywords":"Disaster research; Research ethics; Public relations; Psychological intervention; Disaster medicine; Natural disaster; Political science; Medicine; Engineering ethics; Poison control; Suicide prevention; Medical emergency; Engineering; Nursing; Management","retraction":null,"screen_n_in":null,"score":{"opus":0.9057001708361343,"gpt":0.646822186922242,"spread":0.2588779839138923,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","sts","research_integrity"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.1548478,0.0001721389,0.001076441,0.0003089935,0.0003178524,0.00001160538,0.0008661317,0.0009868182,0.0003276478],"category_scores_gemma":[0.4627102,0.00009373813,0.0001535775,0.0008606494,0.009230467,0.0000724166,0.001188658,0.01228581,0.00005103559],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001138145,"about_ca_system_score_gemma":0.001923069,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001557949,"about_ca_topic_score_gemma":0.0006739075,"domain_scores_codex":[0.9705899,0.01342347,0.002091934,0.0007494242,0.0121463,0.00099897],"domain_scores_gemma":[0.5959895,0.3825753,0.0004834122,0.002950775,0.01756482,0.0004362469],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"systematic_review","study_design_gemma":"systematic_review","study_design_scores_codex":[0.01148762,0.011104,0.004679584,0.4095472,0.003198346,0.0001349822,0.04555902,8.553651e-8,0.1138135,0.3646209,0.01579993,0.02005485],"study_design_scores_gemma":[0.00519637,0.01298323,0.002384793,0.5525895,0.0005978866,0.00001342303,0.005484125,0.000029143,0.1036856,0.2866453,0.02988846,0.0005022253],"study_design_candidate":"systematic_review","study_design_consensus":"systematic_review","genre_codex":"commentary","genre_gemma":"review","genre_scores_codex":[0.1139505,0.1229699,0.00001222718,0.7171758,0.00006716783,0.003947023,0.0001149252,0.00002947949,0.041733],"genre_scores_gemma":[0.2821732,0.7076916,0.001092723,0.0006470298,0.0001179113,0.00009419657,0.000003149221,0.00005478875,0.008125442],"genre_candidate":"review","genre_consensus":null,"teacher_disagreement_score":0.7165288,"threshold_uncertainty_score":0.9934658,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"evaluation","study_design":"qualitative","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch"],"domain":"evaluation","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"split"},{"id":"W2583047471","doi":"10.1080/08989621.2017.1287567","title":"Best Practice to Order Authors in Multi/Interdisciplinary Health Sciences Research Publications","year":2017,"lang":"en","type":"article","venue":"Accountability in Research","topic":"Ethics in Clinical Research","field":"Medicine","cited_by":56,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal","funders":"Social Sciences and Humanities Research Council; Social Sciences and Humanities Research Council of Canada; Université de Montréal","keywords":"Order (exchange); Engineering ethics; Declaration; Best practice; Accountability; Attribution; Confusion; Computer science; Political science; Psychology; Public relations; Management science; Sociology; Business; Social psychology; Law; Engineering","retraction":null,"screen_n_in":null,"score":{"opus":0.9220851009964572,"gpt":0.8041078746149898,"spread":0.1179772263814675,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","sts","research_integrity","insufficient_payload"],"consensus_categories":["metaresearch","sts"],"category_scores_codex":[0.240632,0.000195284,0.00054918,0.002518669,0.003158268,0.0009636302,0.003281767,0.000568661,0.0003889091],"category_scores_gemma":[0.477455,0.0001738806,0.00007686738,0.005028972,0.005907601,0.001270809,0.006727421,0.01214787,0.00101392],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002410007,"about_ca_system_score_gemma":0.009573539,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.04582198,"about_ca_topic_score_gemma":0.1308576,"domain_scores_codex":[0.9828082,0.004035822,0.001343624,0.001906417,0.007331734,0.002574213],"domain_scores_gemma":[0.8753092,0.09518246,0.0003301474,0.008828728,0.01818358,0.002165934],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.001652888,0.01030547,0.9273467,0.0009447468,0.00002736658,0.0001012623,0.0122684,0.00004069092,0.0003332892,0.02359899,0.007547678,0.01583254],"study_design_scores_gemma":[0.002052485,0.003248475,0.8936264,0.001830177,0.0000036959,0.0000253766,0.0444391,0.001488087,0.0001041649,0.03983393,0.01305076,0.0002973254],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.4948779,0.0001682878,0.00003242673,0.4637107,0.0001443517,0.002795577,0.00001497862,0.00003121772,0.03822455],"genre_scores_gemma":[0.9812374,0.0005018513,0.01024559,0.0006807427,0.0001600285,0.0008045014,0.000008671418,0.00003658687,0.00632464],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4863594,"threshold_uncertainty_score":0.9997639,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"evaluation","study_design":"theoretical_or_conceptual","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["research_integrity","scholarly_communication"],"domain":null,"study_design":"design_other","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W4234553601","doi":"10.31219/osf.io/m9abx","title":"The Hong Kong Principles for Assessing Researchers: Fostering Research Integrity","year":2019,"lang":"en","type":"preprint","venue":"","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":56,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Ottawa Hospital; University of Ottawa","funders":"","keywords":"Transparency (behavior); Trustworthiness; Engineering ethics; Research integrity; Inclusion (mineral); Manifesto; Political science; Research ethics; Public relations; Open science; Quality (philosophy); Declaration; Computer science; Sociology; Engineering; Internet privacy; Law; Social science; Epistemology","retraction":null,"screen_n_in":null,"score":{"opus":0.9051326629292764,"gpt":0.6570798519375924,"spread":0.248052810991684,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["research_integrity"],"consensus_categories":[],"category_scores_codex":[0.01106848,0.0001932496,0.0003376314,0.0002410596,0.0008181034,0.0006941344,0.000474138,0.0005300517,0.00005534935],"category_scores_gemma":[0.008102661,0.0001281152,0.0001759835,0.0002314653,0.0002758511,0.000136362,0.001309184,0.003599067,0.00009441201],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007526019,"about_ca_system_score_gemma":0.003014455,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00257557,"about_ca_topic_score_gemma":0.0006700942,"domain_scores_codex":[0.996601,0.0004087609,0.0006943117,0.0006299594,0.0008189451,0.0008469876],"domain_scores_gemma":[0.9915125,0.004801473,0.0001509453,0.001155982,0.002151456,0.0002276584],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"qualitative","study_design_scores_codex":[0.0007012075,0.0002623162,0.04721813,0.006426875,0.0001926788,0.000009800137,0.006354212,0.0002163532,0.001299768,0.008389633,0.01264231,0.9162867],"study_design_scores_gemma":[0.0004554292,0.001788136,0.0497301,0.02030854,0.0002508429,0.00006836325,0.2370331,0.1688424,0.09936778,0.1930749,0.2272227,0.001857766],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8724896,0.002387751,0.01524016,0.0592149,0.00426387,0.007325234,0.00001073698,0.0001823633,0.03888541],"genre_scores_gemma":[0.9800255,0.0007100945,0.005357048,0.0001465278,0.00165712,0.0005005622,0.00006032866,0.00005422752,0.01148858],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.9144289,"threshold_uncertainty_score":0.9986997,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"evaluation","study_design":"theoretical_or_conceptual","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","research_integrity"],"domain":"evaluation","study_design":"theoretical_or_conceptual","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"agree"},{"id":"W4400879434","doi":"10.1136/bmjebm-2024-112899","title":"Rapid review method series: interim guidance for the reporting of rapid reviews","year":2024,"lang":"en","type":"article","venue":"BMJ evidence-based medicine","topic":"Meta-analysis and systematic reviews","field":"Decision Sciences","cited_by":55,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Public Health Ontario; University of Toronto; St. Michael's Hospital; University of Alberta; Public Health Agency of Canada; Bruyère; University of Ottawa","funders":"Canadian Institutes of Health Research; Canada Research Chairs","keywords":"Interim; Systematic review; Set (abstract data type); Computer science; Psychology; MEDLINE; Political science","retraction":null,"screen_n_in":null,"score":{"opus":0.9004752747875066,"gpt":0.647176774034603,"spread":0.2532985007529035,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","insufficient_payload"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.6212929,0.0005126928,0.008255812,0.0003724203,0.0001626557,0.0002288833,0.002523698,0.00008886765,0.01805723],"category_scores_gemma":[0.7240558,0.0001770049,0.003524217,0.003375835,0.0002885413,0.0004984011,0.0001321225,0.0002813146,0.0004013348],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00006049069,"about_ca_system_score_gemma":0.0004617695,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0000241509,"about_ca_topic_score_gemma":0.000009566395,"domain_scores_codex":[0.9181783,0.01889427,0.05374631,0.001955569,0.006711117,0.0005144303],"domain_scores_gemma":[0.7314931,0.1577007,0.08257787,0.01996645,0.007655666,0.0006062778],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00001939963,0.000008096014,0.00006162098,0.01072564,0.0001251199,0.00001076435,0.0001664192,0.000007610145,0.0003821898,0.0005594808,0.5713201,0.4166135],"study_design_scores_gemma":[0.0001166431,0.0003165945,0.00008848333,0.09394831,0.001492092,0.0000775343,0.0002625198,0.005205038,0.0002782673,0.001005659,0.8970376,0.0001712299],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.0000123077,0.5960172,0.3266341,0.07120948,0.0009676515,0.00426934,0.000006935545,0.00001305338,0.0008699562],"genre_scores_gemma":[0.006081666,0.6034282,0.3036647,0.03809823,0.004126733,0.006502778,0.0000360153,0.0001355678,0.03792608],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.4164423,"threshold_uncertainty_score":0.9828404,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch"],"domain":"reporting","study_design":"not_applicable","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"high"},{"model":"gpt","categories":["metaresearch","research_integrity"],"domain":"reporting","study_design":"not_applicable","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W2766184587","doi":"10.1186/s40510-017-0196-7","title":"Effectiveness of en masse versus two-step retraction: a systematic review and meta-analysis","year":2017,"lang":"en","type":"review","venue":"Progress in Orthodontics","topic":"Orthodontics and Dentofacial Orthopedics","field":"Dentistry","cited_by":53,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"","keywords":"Medicine; Meta-analysis; Dentistry; Randomized controlled trial; Orthodontics; Surgery; Internal medicine","retraction":null,"screen_n_in":null,"score":{"opus":0.1865449401167475,"gpt":0.4686407095626851,"spread":0.2820957694459376,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","metaepi_broad"],"consensus_categories":[],"category_scores_codex":[0.00898363,0.001013629,0.01703977,0.0007510525,0.0001414306,0.000298963,0.0009248139,0.0008251872,0.00008103392],"category_scores_gemma":[0.002495539,0.0007675448,0.004772052,0.001183769,0.0002733378,0.0002515344,0.0005040009,0.001098623,0.00003192488],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001531028,"about_ca_system_score_gemma":0.0002083872,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00004509601,"about_ca_topic_score_gemma":0.00028263,"domain_scores_codex":[0.9899106,0.004641711,0.002831376,0.001059718,0.00103722,0.0005193493],"domain_scores_gemma":[0.9900385,0.002984572,0.003902062,0.00245925,0.0004080636,0.0002075382],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"systematic_review","study_design_gemma":"meta_analysis","study_design_scores_codex":[0.00002789033,0.0001173109,0.004886171,0.8754627,0.1115774,0.0003554928,0.000009958277,4.472391e-7,3.030312e-9,0.0005300533,0.000008788596,0.007023798],"study_design_scores_gemma":[0.0004719409,0.00006864161,0.0002436886,0.07067263,0.8995192,0.00008805483,0.000004795082,0.00001431119,1.764444e-7,0.000003485295,0.02825827,0.0006548142],"study_design_candidate":"meta_analysis","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.000008786069,0.9933023,0.0004452432,0.00001042968,0.0007690649,0.004751027,0.0003124353,0.00005313996,0.000347554],"genre_scores_gemma":[0.0001547628,0.995366,0.002562627,0.000007044553,0.00006480063,0.001022137,0.0001545738,0.0001199755,0.0005480474],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.80479,"threshold_uncertainty_score":0.9994776,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","metaepi_broad"],"domain":"evaluation","study_design":"meta_analysis","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","research_integrity","scholarly_communication"],"domain":"evaluation","study_design":"meta_analysis","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W2606622294","doi":"10.1016/j.wneu.2017.04.014","title":"Retraction of Neurosurgical Publications: A Systematic Review","year":2017,"lang":"en","type":"review","venue":"World Neurosurgery","topic":"Academic integrity and plagiarism","field":"Social Sciences","cited_by":53,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Hospital for Sick Children; University of Toronto","funders":"","keywords":"Medicine; Medical physics; General surgery","retraction":null,"screen_n_in":null,"score":{"opus":0.1983967056150376,"gpt":0.4359178452412505,"spread":0.237521139626213,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","metaepi_narrow","research_integrity"],"consensus_categories":[],"category_scores_codex":[0.007149249,0.0003660683,0.003628759,0.0005774342,0.0006368819,0.0001736746,0.001182748,0.001071918,0.0003190439],"category_scores_gemma":[0.02124418,0.0003025641,0.001275215,0.0010829,0.0003961476,0.0003954602,0.00009220614,0.00320578,0.0002242156],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001112344,"about_ca_system_score_gemma":0.0008649607,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001156169,"about_ca_topic_score_gemma":0.00005947244,"domain_scores_codex":[0.9928511,0.003240393,0.002024649,0.0005123936,0.0009502571,0.0004212023],"domain_scores_gemma":[0.9909487,0.003999677,0.003505489,0.001109942,0.0002432374,0.0001930069],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"systematic_review","study_design_gemma":"not_applicable","study_design_scores_codex":[0.000001107744,0.00004505333,0.000004902614,0.8784114,0.00004510441,0.00001609745,0.00004213936,8.056714e-9,3.00702e-8,0.008817095,0.07308352,0.03953357],"study_design_scores_gemma":[0.000009491111,0.000002670574,7.186721e-7,0.351714,0.0007662855,0.00002714158,0.000003245622,1.613305e-7,3.697217e-8,0.00004133514,0.6472807,0.0001542597],"study_design_candidate":"systematic_review","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[5.919977e-8,0.9568718,0.000004194071,0.001460872,0.002266077,0.002035701,0.00004404976,0.0001142303,0.03720298],"genre_scores_gemma":[0.000003546899,0.9661994,0.000007160917,0.00056412,0.0004686866,0.0003078537,0.00004379641,0.0000391534,0.03236627],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.5741972,"threshold_uncertainty_score":0.9999427,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"evaluation","study_design":"systematic_review","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"high"},{"model":"gpt","categories":["metaresearch","research_integrity"],"domain":"methods","study_design":"systematic_review","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"agree"},{"id":"W2106350222","doi":"10.1525/jer.2012.7.4.10","title":"Opportunities, Ethical Challenges, and Lessons Learned from Working with Peer Research Assistants in a Multi-Method HIV Community-Based Research Study in Ontario, Canada","year":2012,"lang":"en","type":"article","venue":"Journal of Empirical Research on Human Research Ethics","topic":"Mental Health and Patient Involvement","field":"Health Professions","cited_by":50,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":true},"ca_institutions":"Women's Health In Women's Hands; University of Calgary","funders":"Canadian Institutes of Health Research; Mental Health Commission; University of Calgary","keywords":"Focus group; Qualitative research; Human immunodeficiency virus (HIV); Medical education; Research ethics; Peer support; Best practice; Psychology; Facilitation; Engineering ethics; Medicine; Political science; Nursing; Sociology; Engineering; Family medicine; Social science","retraction":null,"screen_n_in":null,"score":{"opus":0.9881056457603145,"gpt":0.7429312961482957,"spread":0.2451743496120188,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","metaepi_narrow","sts","research_integrity"],"consensus_categories":["metaresearch","research_integrity"],"category_scores_codex":[0.4870091,0.000423605,0.001180119,0.005125636,0.007830123,0.0001721187,0.00204121,0.001722928,0.0005013931],"category_scores_gemma":[0.02630611,0.0003336055,0.00007607789,0.002690127,0.001986838,0.0002949261,0.001872403,0.1656433,0.00002797293],"about_ca_system_candidate":true,"about_ca_system_consensus":true,"about_ca_system_score_codex":0.01488104,"about_ca_system_score_gemma":0.0462791,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.9151278,"about_ca_topic_score_gemma":0.9959769,"domain_scores_codex":[0.7427694,0.2263409,0.002343634,0.0008212079,0.02233961,0.005385238],"domain_scores_gemma":[0.8301665,0.1559528,0.0003357346,0.001319451,0.009313629,0.00291188],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.008582594,0.01024569,0.736236,0.001466913,0.0001071073,0.001639442,0.2167703,0.00002472161,0.00005651335,0.004628386,0.01511326,0.005129046],"study_design_scores_gemma":[0.008233517,0.007971729,0.5070406,0.009075887,0.000007545297,0.000003772883,0.3706867,0.0001088162,0.00001867266,0.005305293,0.09118493,0.0003625707],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.897966,0.001091762,0.00001266679,0.09196415,0.0002023424,0.003144275,0.00004506411,0.00001217062,0.005561555],"genre_scores_gemma":[0.9923169,0.001204168,0.0009901554,0.001588739,0.0003222575,0.0004918479,0.000080493,0.0001004557,0.002904972],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.460703,"threshold_uncertainty_score":0.9999116,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch"],"domain":"methods","study_design":"qualitative","genre":"empirical","about_ca_system":false,"about_ca_topic":true,"confidence":"high"},{"model":"gpt","categories":["research_integrity"],"domain":null,"study_design":"qualitative","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"},{"id":"W4406842793","doi":"10.1038/d41586-025-00212-1","title":"‘Stamp out paper mills’ — science sleuths on how to fight fake research","year":2025,"lang":"en","type":"article","venue":"Nature","topic":"Academic integrity and plagiarism","field":"Social Sciences","cited_by":50,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Bishop's University","funders":"","keywords":"Scientific misconduct; Spotting; Misconduct; Engineering ethics; Political science; Engineering; Computer science; Law; Medicine; Alternative medicine","retraction":null,"screen_n_in":null,"score":{"opus":0.04031818940135897,"gpt":0.4228028907455642,"spread":0.3824847013442052,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["sts","research_integrity"],"consensus_categories":["research_integrity"],"category_scores_codex":[0.005278902,0.00009814664,0.0001198654,0.0003513665,0.001624132,0.0002451956,0.001158502,0.002807168,0.0003059565],"category_scores_gemma":[0.007637953,0.00007976296,0.00004833618,0.001857587,0.0006374621,0.0002642359,0.0001515074,0.01040548,0.0003282108],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000307192,"about_ca_system_score_gemma":0.0008605358,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0003355954,"about_ca_topic_score_gemma":0.001484175,"domain_scores_codex":[0.9969459,0.0002936342,0.00009371812,0.0004287795,0.001572181,0.0006658482],"domain_scores_gemma":[0.9983069,0.0006168683,0.00002123268,0.0002718352,0.0005919809,0.0001911358],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00002199074,0.00001997829,0.0002167521,0.000006202847,0.000004762182,0.000005265267,0.01123444,0.000001213165,0.001682321,0.303945,0.6729299,0.009932146],"study_design_scores_gemma":[0.00007106907,0.00002511527,0.0009560608,0.00008079258,0.000002392585,1.396561e-7,0.003832523,0.000004827958,0.002924868,0.00918745,0.982818,0.00009675042],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"other","genre_gemma":"empirical","genre_scores_codex":[0.02413682,0.0007610692,0.00003757921,0.3052607,0.002215668,0.000394245,0.00002704345,0.00007597485,0.6670908],"genre_scores_gemma":[0.8524667,0.000138528,0.0002659316,0.02258747,0.0005510114,0.00001328036,0.000002770431,0.000005334081,0.123969],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8283299,"threshold_uncertainty_score":0.9996756,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"evaluation","study_design":"qualitative","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["research_integrity","scholarly_communication"],"domain":null,"study_design":"not_applicable","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"split"},{"id":"W4387695005","doi":"10.3346/jkms.2023.38.e333","title":"Causes for Retraction in the Biomedical Literature: A Systematic Review of Studies of Retraction Notices","year":2023,"lang":"en","type":"review","venue":"Journal of Korean Medical Science","topic":"Academic integrity and plagiarism","field":"Social Sciences","cited_by":47,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Ottawa Hospital; University of Ottawa","funders":"H. Lundbeck A/S","keywords":"Misconduct; Subspecialty; Medicine; Meta-analysis; MEDLINE; Systematic review; Confidence interval; Scientific misconduct; Publication bias; Family medicine; Internal medicine; Alternative medicine; Pathology; Political science","retraction":null,"screen_n_in":null,"score":{"opus":0.1576615077346172,"gpt":0.4965828687945747,"spread":0.3389213610599575,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","research_integrity"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.07524016,0.0001774101,0.002194985,0.0005384238,0.0002266647,0.00004269029,0.001732692,0.0009991542,0.00001794118],"category_scores_gemma":[0.1599115,0.00008625112,0.0005195613,0.003561051,0.002197417,0.0003990277,0.00005407096,0.002895503,0.000002412797],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002682695,"about_ca_system_score_gemma":0.003098489,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00007951219,"about_ca_topic_score_gemma":0.00007764728,"domain_scores_codex":[0.9903781,0.001699074,0.002586064,0.0002070006,0.004814689,0.0003150437],"domain_scores_gemma":[0.9866428,0.008053438,0.003849352,0.0001965193,0.001052255,0.0002056703],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"systematic_review","study_design_gemma":"systematic_review","study_design_scores_codex":[0.000006181649,0.00007762997,8.937152e-7,0.9633203,0.0001017964,0.00003844143,0.006356533,3.859787e-8,5.953146e-7,0.002823784,0.00545821,0.02181561],"study_design_scores_gemma":[0.00006278503,0.0001320234,0.000001936129,0.8793256,0.0008167957,0.0001055797,0.003858644,0.000005145892,8.278895e-7,0.0005004917,0.1151074,0.00008276542],"study_design_candidate":"systematic_review","study_design_consensus":"systematic_review","genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.00001025965,0.992979,0.00004498718,0.004193753,0.001667675,0.0009479235,0.00001687713,0.000005398183,0.0001340594],"genre_scores_gemma":[0.00008979032,0.9987404,0.0001368793,0.0002068002,0.0007451501,0.00002044621,0.000003917121,0.000006686558,0.00004996469],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.1096492,"threshold_uncertainty_score":0.9994048,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"evaluation","study_design":"systematic_review","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaepi_broad","research_integrity"],"domain":null,"study_design":"systematic_review","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"}]}