{"meta":{"page":1,"per_page":50,"max_per_page":100,"total":9,"total_is_capped":false,"direct_labels_cover":0,"predictions_cover":9,"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","author_layer_release":"2026-06-26"},"query_hash":"c357e742c7f2","filters":{"venue":"Journal of Open Humanities Data"}},"results":[{"id":"W4312706369","doi":"10.5334/johd.88","title":"The CONLIT Dataset of Contemporary Literature","year":2022,"lang":"en","type":"article","venue":"Journal of Open Humanities Data","topic":"Narrative Theory and Analysis","field":"Arts and Humanities","cited_by":4,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"McGill University","funders":"","keywords":"Elite; Biography; Literature; History; Art","authors":[{"name":"Andrew Piper","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.2284273115647129,"gpt":0.3408972242918083,"spread":0.1124699127270954,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001838102,0.001274797,0.001358192,0.02507523,0.001648565,0.004523055,0.002536519,0.002294369,0.03962712],"category_scores_gemma":[0.0146458,0.0004839173,0.0008549555,0.02888524,0.0006272218,0.002241835,0.003358843,0.001586995,0.04487808],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002188561,"about_ca_system_score_gemma":0.004117466,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01110686,"about_ca_topic_score_gemma":0.03207958,"domain_scores_codex":[0.9971751,0.000445496,0.0005753104,0.0005957326,0.0009179002,0.0002904288],"domain_scores_gemma":[0.9887495,0.00418743,0.001687154,0.001471604,0.002891211,0.001013128],"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":[0.000127195,0.0001060175,0.006906373,0.003140166,0.00006084327,0.0002565262,0.0003044465,0.0003928451,0.0005432754,0.003224214,0.9661364,0.01880166],"study_design_scores_gemma":[0.00009657253,0.0000234851,0.009474726,0.0007281545,0.00003414193,0.000249328,0.0004726537,0.0004517656,0.0003460702,0.000988518,0.9871047,0.0000298817],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"dataset","genre_gemma":"dataset","genre_scores_codex":[0.002238865,0.001124955,0.0002177619,0.0002484567,0.00005692785,0.00005779086,0.9924901,0.0002479977,0.003317005],"genre_scores_gemma":[0.001761138,0.0003046152,0.0007418054,0.00008508625,0.00002769167,0.000188517,0.9959629,0.00003712007,0.0008910757],"genre_candidate":"dataset","genre_consensus":"dataset","teacher_disagreement_score":0.03962712,"threshold_uncertainty_score":0.132566,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4308571509","doi":"10.5334/johd.97","title":"CREMMA Medii Aevi: Literary Manuscript Text Recognition in Latin","year":2023,"lang":"en","type":"article","venue":"Journal of Open Humanities Data","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal","funders":"","keywords":"Latin Americans; Computer science; Representation (politics); Natural language processing; Artificial intelligence; Segmentation; Medieval Latin; Information retrieval; Linguistics; History; Classics; Political science","authors":[{"name":"Thibault Clérice","is_ca":false},{"name":"Malamatenia Vlachou-Efstathiou","is_ca":false},{"name":"Alix Chagué","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.23807547872461,"gpt":0.3631402314952308,"spread":0.1250647527706208,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000636499,0.001102234,0.0004920916,0.00355832,0.0009871394,0.001898034,0.001177825,0.001120873,0.008566272],"category_scores_gemma":[0.002731885,0.0001681824,0.0008820532,0.002242932,0.0005243726,0.001411159,0.001271962,0.0007248707,0.01128483],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009279613,"about_ca_system_score_gemma":0.0009354121,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00735213,"about_ca_topic_score_gemma":0.01752676,"domain_scores_codex":[0.9993241,0.0001161166,0.00009722853,0.0002388841,0.000159885,0.00006382694],"domain_scores_gemma":[0.9988243,0.0002332793,0.0001065979,0.0003107163,0.0004113583,0.0001137551],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0008196988,0.000426958,0.01878261,0.002098014,0.0001613163,0.001457827,0.001321472,0.004214742,0.02934627,0.00385063,0.4893068,0.4482136],"study_design_scores_gemma":[0.0001303962,0.0002097609,0.04111538,0.0004584843,0.00008157677,0.002527266,0.002487289,0.02416502,0.03960425,0.002950195,0.8861455,0.0001247778],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"dataset","genre_gemma":"empirical","genre_scores_codex":[0.3447999,0.01377491,0.04615871,0.003938957,0.003984231,0.0009403466,0.4528052,0.05229851,0.08129923],"genre_scores_gemma":[0.2153117,0.001935884,0.07432689,0.0009791129,0.0006426053,0.0008428118,0.6760392,0.001317443,0.02860445],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.008566272,"threshold_uncertainty_score":0.02865702,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4312215389","doi":"10.5334/johd.94","title":"The TRANSCOMP Dataset of Literary Translations from 120 Languages and a Parallel Collection of English-language Originals","year":2022,"lang":"en","type":"article","venue":"Journal of Open Humanities Data","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"McGill University","funders":"","keywords":"Linguistics; Computer science; Word (group theory); Metadata; Literary translation; Natural language processing; Literature; History; Art; World Wide Web; Philosophy","authors":[{"name":"Matt Erlin","is_ca":false},{"name":"Andrew Piper","is_ca":true},{"name":"Douglas Knox","is_ca":false},{"name":"Stephen Pentecost","is_ca":false},{"name":"Allie Blank","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.05700619531125964,"gpt":0.3531706120189455,"spread":0.2961644167076859,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["open_science"],"consensus_categories":[],"category_scores_codex":[0.00140936,0.001549824,0.001075152,0.009182373,0.001480154,0.002151818,0.001923903,0.001633157,0.02525755],"category_scores_gemma":[0.007009558,0.0004069777,0.001037453,0.01015758,0.001011453,0.001837809,0.003228265,0.001572987,0.03728485],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0008320847,"about_ca_system_score_gemma":0.00178224,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005032044,"about_ca_topic_score_gemma":0.01312874,"domain_scores_codex":[0.998027,0.0004092733,0.0003451486,0.0003999996,0.0006134851,0.0002050759],"domain_scores_gemma":[0.9951062,0.001338028,0.000352719,0.001203313,0.001528825,0.0004709347],"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":[0.0004636926,0.0003879423,0.005401305,0.002717291,0.00008923571,0.0009803771,0.001102062,0.0006481677,0.004231126,0.002372569,0.9373074,0.04429876],"study_design_scores_gemma":[0.0002794582,0.0001190445,0.02242036,0.0004726137,0.00005043485,0.001068203,0.001839859,0.001273947,0.003441569,0.001405845,0.967553,0.00007570979],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"dataset","genre_gemma":"dataset","genre_scores_codex":[0.01692666,0.0006954038,0.001257984,0.0002683159,0.0003063066,0.0001818531,0.9729258,0.001294454,0.006143227],"genre_scores_gemma":[0.00425132,0.0001194938,0.001760456,0.00004961857,0.00005410174,0.0002281405,0.9915614,0.0001475358,0.00182808],"genre_candidate":"dataset","genre_consensus":"dataset","teacher_disagreement_score":0.9980761,"threshold_uncertainty_score":0.08449495,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4319597719","doi":"10.5334/johd.95","title":"MultiHATHI: A Complete Collection of Multilingual Prose Fiction in the HathiTrust Digital Library","year":2023,"lang":"en","type":"article","venue":"Journal of Open Humanities Data","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"McGill University","funders":"","keywords":"Bespoke; Metadata; Computer science; Digital library; Classifier (UML); English language; World Wide Web; Information retrieval; Natural language processing; Artificial intelligence; Linguistics; Literature; Art; Poetry","authors":[{"name":"Sil Hamilton","is_ca":true},{"name":"Andrew Piper","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.1386274117001172,"gpt":0.3461929316426599,"spread":0.2075655199425427,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006231831,0.0008974321,0.0006937363,0.01247117,0.001797711,0.002616699,0.001038891,0.00111565,0.02161973],"category_scores_gemma":[0.004316409,0.0002939558,0.0005953143,0.01386523,0.0007140151,0.002967219,0.00243062,0.0009840286,0.02651446],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001439791,"about_ca_system_score_gemma":0.00199101,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01664353,"about_ca_topic_score_gemma":0.05296014,"domain_scores_codex":[0.9991701,0.0001380207,0.0001052874,0.0001687482,0.0002971261,0.0001206846],"domain_scores_gemma":[0.9973581,0.0007071872,0.0002803342,0.0005306274,0.0007534109,0.0003701865],"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":[0.0002562755,0.0001074742,0.01146179,0.002561678,0.00006383966,0.0005822787,0.002478006,0.0004402304,0.00225604,0.003102506,0.9140636,0.06262641],"study_design_scores_gemma":[0.00003399049,0.00003248051,0.02953377,0.0003715955,0.00003429595,0.0004871944,0.001750596,0.0008932977,0.001887204,0.0009480707,0.9639735,0.00005401648],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"dataset","genre_gemma":"empirical","genre_scores_codex":[0.03420883,0.003376062,0.001115022,0.0006079423,0.0003350352,0.0001047341,0.9317678,0.00207751,0.02640715],"genre_scores_gemma":[0.01891634,0.0006375014,0.002640842,0.0001263639,0.0001287123,0.0001412662,0.9694312,0.0003184161,0.007659328],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.02161973,"threshold_uncertainty_score":0.07232523,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3190643984","doi":"10.5334/johd.42","title":"The Corpus for Idiolectal Research (CIDRE)","year":2021,"lang":"en","type":"article","venue":"Journal of Open Humanities Data","topic":"Authorship Attribution and Profiling","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"Agence Nationale de la Recherche","keywords":"Paratext; Scripting language; Metadata; License; Computer science; World Wide Web; Open source; Style (visual arts); Literature; Linguistics; Art; Programming language; Software; Philosophy","authors":[{"name":"Olga Seminck","is_ca":false},{"name":"Philippe Gambette","is_ca":false},{"name":"Dominique Legallois","is_ca":false},{"name":"Thierry Poibeau","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.6364291660620014,"gpt":0.5070253019086151,"spread":0.1294038641533863,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002936705,0.0007539996,0.001006353,0.01730108,0.003419783,0.003865032,0.001196814,0.001156477,0.04065711],"category_scores_gemma":[0.01092529,0.0003045239,0.0003560031,0.01704996,0.001493519,0.001690161,0.002789537,0.001305154,0.01670372],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.005200068,"about_ca_system_score_gemma":0.006072962,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.07552463,"about_ca_topic_score_gemma":0.07976341,"domain_scores_codex":[0.9968388,0.0008986267,0.000359894,0.0007707024,0.0007050975,0.000426799],"domain_scores_gemma":[0.9928323,0.002385525,0.0005535002,0.001424918,0.002291789,0.0005118435],"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":[0.0002298049,0.00008764519,0.005919826,0.002840462,0.00004139943,0.0007831931,0.01153674,0.0002521784,0.003825201,0.05141644,0.7615604,0.1615067],"study_design_scores_gemma":[0.00001840416,0.00000754861,0.008882189,0.0003788226,0.00001071003,0.0002951523,0.001084762,0.00007215051,0.0004420598,0.001297434,0.9874871,0.00002363209],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"dataset","genre_gemma":"dataset","genre_scores_codex":[0.04115633,0.009503007,0.007062937,0.001828871,0.0006726086,0.0007479881,0.8390331,0.001661129,0.09833395],"genre_scores_gemma":[0.1266719,0.003167764,0.02703303,0.0007422514,0.0004625967,0.002466179,0.8139523,0.001467204,0.02403672],"genre_candidate":"dataset","genre_consensus":"dataset","teacher_disagreement_score":0.07552463,"threshold_uncertainty_score":0.15017,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4389949574","doi":"10.5334/johd.169","title":"3D Modelling of Jerusalem’s Maghrebi Quarter","year":2023,"lang":"en","type":"article","venue":"Journal of Open Humanities Data","topic":"3D Surveying and Cultural Heritage","field":"Earth and Planetary Sciences","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"Université Gustave Eiffel; Università Degli Studi di Modena e Reggio Emila","keywords":"Quarter (Canadian coin); Geology; Archaeology; History","authors":[{"name":"Raphaël Banc-Lévêque","is_ca":false},{"name":"Raffaele Peluso","is_ca":false},{"name":"Marco Cozza","is_ca":false},{"name":"Fabio Bruno","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.3692056332634672,"gpt":0.3160634521328087,"spread":0.05314218113065855,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002582705,0.0004479826,0.0003174631,0.001526176,0.0008749653,0.002544108,0.0004978117,0.0007092174,0.005107529],"category_scores_gemma":[0.0006006128,0.0003633496,0.0004340535,0.00147566,0.000646009,0.0005558379,0.0008461056,0.0003441067,0.001196809],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000678483,"about_ca_system_score_gemma":0.0004740703,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.02209685,"about_ca_topic_score_gemma":0.0393419,"domain_scores_codex":[0.999869,0.00002716846,0.000005990094,0.00003338649,0.00003850707,0.00002600639],"domain_scores_gemma":[0.9999385,0.00002122237,0.000005611336,0.00001365153,0.0000144351,0.000006603087],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0009904858,0.0003363971,0.05751362,0.002269952,0.0001754571,0.007933254,0.02766964,0.45272,0.02485854,0.04507497,0.04622883,0.334229],"study_design_scores_gemma":[0.0001174043,0.0002453237,0.1498396,0.001240048,0.0001610946,0.002842974,0.02734694,0.379787,0.01212742,0.008093309,0.417934,0.0002648944],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8688604,0.003291631,0.02279144,0.0008951828,0.0003123019,0.0002990449,0.007440166,0.001010042,0.09509983],"genre_scores_gemma":[0.9702358,0.001160757,0.01478816,0.00005253611,0.00002946908,0.00006564222,0.002649649,0.0001776364,0.0108405],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.02209685,"threshold_uncertainty_score":0.04393643,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4416570569","doi":"10.5334/johd.383","title":"Digital Humanities Training and Workshops: Anonymized Survey Results","year":2025,"lang":"en","type":"article","venue":"Journal of Open Humanities Data","topic":"Digital Humanities and Scholarship","field":"Arts and Humanities","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"Social Sciences and Humanities Research Council of Canada; St. Francis Xavier University; University of Ottawa","keywords":"Certificate; Qualitative property; Digital humanities; Survey data collection; Digital media; Computer-assisted web interviewing; Qualitative research","authors":[{"name":"Bridget Moynihan","is_ca":false},{"name":"Laura Estill","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.3640158202480061,"gpt":0.3469776897929208,"spread":0.01703813045508523,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.03289508,0.0005010659,0.0004628557,0.004854786,0.002562043,0.003205124,0.001328088,0.00124149,0.023351],"category_scores_gemma":[0.1193235,0.0005017909,0.0003776838,0.01233925,0.001676854,0.001840512,0.002875322,0.001070289,0.006804338],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.005132127,"about_ca_system_score_gemma":0.009884391,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.07159688,"about_ca_topic_score_gemma":0.06323,"domain_scores_codex":[0.939134,0.02438137,0.005446082,0.003944649,0.02277974,0.004314147],"domain_scores_gemma":[0.8998405,0.03924365,0.008717386,0.01630764,0.03431796,0.001572968],"domain_codex":null,"domain_gemma":"evaluation","domain_candidate":"evaluation","domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"observational","study_design_scores_codex":[0.0005351299,0.0007545577,0.1912335,0.002384634,0.0001775832,0.0003668856,0.05383369,0.001467775,0.00131499,0.01825877,0.6267241,0.1029484],"study_design_scores_gemma":[0.00009890856,0.0001397381,0.292547,0.001757988,0.0000656136,0.0002093505,0.05410919,0.001465228,0.001579444,0.004391294,0.6434821,0.0001541235],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"dataset","genre_gemma":"empirical","genre_scores_codex":[0.2841879,0.0006260533,0.01294885,0.007272162,0.0004772834,0.009131196,0.5796376,0.0004036688,0.1053153],"genre_scores_gemma":[0.7190723,0.001259564,0.01586077,0.00521473,0.0002434214,0.04626536,0.1802671,0.0005602002,0.03125658],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9671049,"threshold_uncertainty_score":0.1739678,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4408373139","doi":"10.5334/johd.262","title":"Data from ARTiculating Canadian Identities: The Canada Council Art Bank’s Mycelial Presence and Influence in the Realm of Canadian Fine Art","year":2024,"lang":"en","type":"article","venue":"Journal of Open Humanities Data","topic":"Visual Culture and Art Theory","field":"Arts and Humanities","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"Social Sciences and Humanities Research Council of Canada","keywords":"Realm; Political science; Art; Public administration; Law","authors":[{"name":"Jacoba Krina Van Hoeve","is_ca":false}],"retraction":null,"screen_n_in":null,"score":{"opus":0.2285460099227496,"gpt":0.3058638438127861,"spread":0.07731783389003649,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001526305,0.000719315,0.0005438298,0.009395101,0.006620687,0.005415955,0.002150177,0.0009617598,0.01518828],"category_scores_gemma":[0.01013967,0.0004388441,0.0004800999,0.02289746,0.001639175,0.001610564,0.00427033,0.001444977,0.004271553],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.03895355,"about_ca_system_score_gemma":0.05502653,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.9948463,"about_ca_topic_score_gemma":0.9983121,"domain_scores_codex":[0.9969908,0.000142747,0.0001143904,0.0003308156,0.001634191,0.0007871077],"domain_scores_gemma":[0.9895303,0.0007758496,0.0007561002,0.0008921472,0.006509236,0.001536442],"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":[0.0002957214,0.0001363067,0.1362084,0.0008892717,0.0001015204,0.0004647678,0.02466836,0.0009215358,0.0008385116,0.01293169,0.7534844,0.06905948],"study_design_scores_gemma":[0.00002574175,0.00001328754,0.3670702,0.0006476409,0.00005170751,0.0001001542,0.02744156,0.0008050677,0.0006348629,0.0008907404,0.6021895,0.0001294703],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"dataset","genre_gemma":"dataset","genre_scores_codex":[0.1344921,0.001776265,0.001057022,0.002078791,0.0001431759,0.0003471012,0.7792645,0.0003005265,0.08054053],"genre_scores_gemma":[0.3906623,0.001715178,0.005680348,0.0006652654,0.00007714134,0.0007184109,0.567151,0.0003070752,0.03302344],"genre_candidate":"dataset","genre_consensus":"dataset","teacher_disagreement_score":0.03895355,"threshold_uncertainty_score":0.2826292,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4382727197","doi":"10.5334/johd.109","title":"A Mixed Method Twitter Methodology and Anonymous Corpus","year":2023,"lang":"en","type":"article","venue":"Journal of Open Humanities Data","topic":"Misinformation and Its Impacts","field":"Social Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Brock University","funders":"","keywords":"Vetting; Anonymity; Computer science; Field (mathematics); Work (physics); Process (computing); World Wide Web; Simple (philosophy); Data science; Internet privacy; Information retrieval; Computer security; Engineering; Mathematics","authors":[{"name":"Tim Ribaric","is_ca":true}],"retraction":null,"screen_n_in":null,"score":{"opus":0.7209269913846231,"gpt":0.5227081836891851,"spread":0.198218807695438,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.009382276,0.0008598697,0.0006681507,0.00405641,0.003597868,0.00268319,0.002333709,0.00277027,0.03182924],"category_scores_gemma":[0.02602014,0.0005943737,0.000656559,0.004650998,0.001837903,0.001919795,0.004211611,0.001885595,0.01651814],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001753022,"about_ca_system_score_gemma":0.00344717,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.007475318,"about_ca_topic_score_gemma":0.01883431,"domain_scores_codex":[0.9919943,0.003731006,0.0007205065,0.001461365,0.001625316,0.0004675746],"domain_scores_gemma":[0.977466,0.01041824,0.001808046,0.004914744,0.004779711,0.0006132862],"domain_codex":null,"domain_gemma":"methods","domain_candidate":"methods","domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001463053,0.0022396,0.03287637,0.005064581,0.0001964193,0.002463165,0.01819644,0.00492753,0.008354375,0.0503372,0.7113634,0.1625179],"study_design_scores_gemma":[0.0003137216,0.0003224825,0.02684705,0.0007979864,0.00008045335,0.0006648229,0.008547773,0.006547071,0.004497183,0.01388894,0.9373196,0.0001728383],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"dataset","genre_gemma":"methods","genre_scores_codex":[0.1936145,0.00187117,0.09383463,0.005041266,0.002369886,0.02948325,0.5929961,0.003594638,0.07719461],"genre_scores_gemma":[0.1904972,0.0005839791,0.1716594,0.003511851,0.0009679557,0.08086894,0.4976045,0.002099773,0.05220636],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9906178,"threshold_uncertainty_score":0.1064794,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null}]}