{"meta":{"page":1,"per_page":50,"max_per_page":100,"total":262,"total_is_capped":false,"direct_labels_cover":262,"predictions_cover":262,"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":"283450e612e9","filters":{"category":"open_science"}},"results":[{"id":"W2028758832","doi":"10.1016/s0140-6736(10)60903-8","title":"An international registry of systematic-review protocols","year":2010,"lang":"en","type":"article","venue":"The Lancet","topic":"Meta-analysis and systematic reviews","field":"Decision Sciences","cited_by":521,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Ottawa Hospital","funders":"National Institute for Health and Care Research","keywords":"Medicine; Computer science","retraction":null,"screen_n_in":null,"score":{"opus":0.7505475693393026,"gpt":0.5892674142758912,"spread":0.1612801550634113,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","open_science","insufficient_payload"],"consensus_categories":["metaresearch","insufficient_payload"],"category_scores_codex":[0.09903283,0.0001218203,0.003438641,0.00005135027,0.00004429752,0.0004708082,0.007977618,0.0000359237,0.006925179],"category_scores_gemma":[0.04988371,0.00004036322,0.0006346548,0.0003702229,0.00006140624,0.0001627304,0.00007727091,0.0001637879,0.00112761],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000004045208,"about_ca_system_score_gemma":0.00002854121,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000003480945,"about_ca_topic_score_gemma":0.00001383572,"domain_scores_codex":[0.9873357,0.003868084,0.00520132,0.0003307846,0.003144667,0.0001194666],"domain_scores_gemma":[0.9823982,0.001172291,0.004618836,0.01098145,0.0007763844,0.00005288919],"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":[0.00002106834,0.000207084,0.03141287,0.1485324,0.000640763,0.00000524328,0.000656404,0.00001146289,0.003569071,0.09331319,0.7190724,0.002558021],"study_design_scores_gemma":[0.001036418,0.0001395896,0.02423659,0.07192647,0.001651501,0.0001811948,0.0005905969,0.007907512,0.0007359895,0.0299885,0.8609088,0.0006968519],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"other","genre_gemma":"empirical","genre_scores_codex":[0.1408229,0.01093169,0.01967539,0.07371772,0.004405637,0.2561955,0.0008109328,0.0001142744,0.4933259],"genre_scores_gemma":[0.9168285,0.0003760864,0.02333628,0.007845412,0.002427917,0.02198249,0.00002812314,0.00003987004,0.02713529],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.7760056,"threshold_uncertainty_score":0.9996501,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch"],"domain":"reproducibility","study_design":"not_applicable","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["scholarly_communication","open_science"],"domain":null,"study_design":"not_applicable","genre":"dataset","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"split"},{"id":"W2093745477","doi":"10.1038/sdata.2014.49","title":"An open science resource for establishing reliability and reproducibility in functional connectomics","year":2014,"lang":"en","type":"article","venue":"Scientific Data","topic":"Functional Brain Connectivity Studies","field":"Neuroscience","cited_by":497,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"McGill University; Douglas Mental Health University Institute; Université de Montréal; Institut Universitaire de Gériatrie de Montréal","funders":"Eunice Kennedy Shriver National Institute of Child Health and Human Development; National Institute on Drug Abuse; National Institute of Biomedical Imaging and Bioengineering; National Institute of Neurological Disorders and Stroke; National Institute of Mental Health; National Key Research and Development Program of China","keywords":"Connectomics; Reliability (semiconductor); Reproducibility; Computer science; Neuroimaging; Connectome; Open science; Artificial intelligence; Functional connectivity; Psychology; Neuroscience; Statistics; Mathematics","retraction":null,"screen_n_in":null,"score":{"opus":0.1423463828689606,"gpt":0.3469814610188182,"spread":0.2046350781498576,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","sts","scholarly_communication"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.04595133,0.0001399891,0.0002112855,0.0002132811,0.001525177,0.002060755,0.004274282,0.00003792934,0.00001623727],"category_scores_gemma":[0.3108275,0.0001324092,0.0000140344,0.001319847,0.002403673,0.004385597,0.005526684,0.0001713256,0.000007350997],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001183759,"about_ca_system_score_gemma":0.0003268488,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001282189,"about_ca_topic_score_gemma":0.0004813072,"domain_scores_codex":[0.9916981,0.0003175534,0.0003218209,0.006628861,0.0006015148,0.0004321669],"domain_scores_gemma":[0.980438,0.008017742,0.0001496038,0.01095768,0.0002518007,0.0001851568],"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":[0.0009550698,0.001402519,0.03802828,0.0001948047,0.000006193299,0.000002710221,0.001423122,0.002011527,0.6608348,0.04915056,0.2010748,0.0449156],"study_design_scores_gemma":[0.002000842,0.0003539762,0.1142181,0.00005487966,0.00001748294,0.00002283995,0.0006455712,0.2092962,0.06193011,0.06711411,0.5436053,0.0007405782],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.985036,0.00001339333,0.002085725,0.005758423,0.002693885,0.001260108,0.001488212,0.00009328441,0.001570948],"genre_scores_gemma":[0.9956545,6.985337e-7,0.00250692,0.001009734,0.000111871,0.00004510213,0.0001865978,0.00001257746,0.0004720448],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5989047,"threshold_uncertainty_score":0.9997747,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","open_science"],"domain":"reproducibility","study_design":"not_applicable","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"},{"model":"gpt","categories":["metaresearch","open_science"],"domain":"reproducibility","study_design":"not_applicable","genre":"dataset","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"agree"},{"id":"W2996762233","doi":"10.1016/j.softx.2021.100658","title":"Open data from the first and second observing runs of Advanced LIGO and Advanced Virgo","year":2021,"lang":"en","type":"article","venue":"SoftwareX","topic":"Pulsars and Gravitational Waves Research","field":"Physics and Astronomy","cited_by":494,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Division of Human Resource Development; Dirección General de Cultura y Patrimonio, Generalitat Valenciana; Conselleria d'Educació, Investigació, Cultura i Esport; Australian Research Council; Science and Technology Facilities Council; Centre Eau Terre Environnement, Institut National de la Recherche Scientifique; Istituto Nazionale di Fisica Nucleare; Universitat de les Illes Balears; Ministry of Education, India; National Research Foundation of Korea; Hungarian Scientific Research Fund; Generalitat Valenciana; Industry Canada; Kavli Foundation; National Natural Science Foundation of China; National Research, Development and Innovation Office; Agencia Estatal de Investigación; Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung; Govern de les Illes Balears; Nederlandse Organisatie voor Wetenschappelijk Onderzoek; Agence Nationale de la Recherche; Centre National de la Recherche Scientifique; Council of Scientific and Industrial Research, India; National Research Foundation; Nemzeti Kutatási Fejlesztési és Innovációs Hivatal; Abdus Salam International Centre for Theoretical Physics; Instituto Nazionale di Fisica Nucleare; European Commission; Russian Foundation for Basic Research; ICTP South American Institute for Fundamental Research; Canadian Institute for Advanced Research; Ministero dello Sviluppo Economico; Institut des Origines de Lyon; Russian Science Foundation; Ministerio de Ciencia y Tecnología; Leverhulme Trust; Scottish Funding Council; European Regional Development Fund; Scottish Universities Physics Alliance; Conseil Régional, Île-de-France; Ontario Ministry of Economic Development and Innovation; Science and Engineering Research Board; National Science Foundation; Royal Society","keywords":"LIGO; Sky; Sensitivity (control systems); Open source; Gravitational wave; Series (stratigraphy)","retraction":null,"screen_n_in":null,"score":{"opus":0.03586966795001467,"gpt":0.3476143321169346,"spread":0.3117446641669199,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001083901,0.00008065844,0.0001461615,0.000008711186,0.000158673,0.0001143117,0.0005767588,0.0000168325,0.0004689432],"category_scores_gemma":[0.00005531286,0.00006332302,0.00001742192,0.00009203182,0.00006274667,0.0003280299,0.001796339,0.0001013294,0.000004278553],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000004054273,"about_ca_system_score_gemma":0.00007510575,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00009917006,"about_ca_topic_score_gemma":0.0001253479,"domain_scores_codex":[0.9993063,0.00002732765,0.0001334701,0.0002879768,0.0001128627,0.0001321232],"domain_scores_gemma":[0.9988724,0.0004774459,0.00005553713,0.0004748287,0.00006936589,0.00005040421],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0003600053,0.0005757462,0.5207837,0.0002833067,0.0009426381,0.00003513237,0.004493098,0.0002126928,0.04738257,0.0199685,0.03363154,0.3713312],"study_design_scores_gemma":[0.002350599,0.00006665634,0.5899283,0.0002202376,0.00004097499,0.000001780031,0.002715399,0.0006818019,0.004072703,0.151681,0.2478839,0.0003565598],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.987853,0.002320982,0.0002651549,0.001223008,0.00007046589,0.0002734824,0.00658334,0.000004888245,0.001405668],"genre_scores_gemma":[0.9881102,0.00006447875,0.006769764,0.0002089914,0.00008053217,0.00001464775,0.001813108,0.0000151691,0.002923099],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3709746,"threshold_uncertainty_score":0.5134599,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["open_science"],"domain":null,"study_design":"not_applicable","genre":"dataset","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["open_science"],"domain":null,"study_design":"not_applicable","genre":"dataset","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"agree"},{"id":"W3024722698","doi":"10.1038/s41597-020-0486-7","title":"The TRUST Principles for digital repositories","year":2020,"lang":"en","type":"article","venue":"Scientific Data","topic":"Research Data Management Practices","field":"Computer Science","cited_by":390,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Ocean Networks Canada Society; University of Victoria","funders":"National Institute of Biomedical Imaging and Bioengineering; National Institute of Diabetes and Digestive and Kidney Diseases; Economic and Social Research Council","keywords":"World Wide Web; Computer science; Metadata; Scholarly communication; Data science; Internet privacy; Knowledge management; Publishing; Political science","retraction":null,"screen_n_in":null,"score":{"opus":0.2872569636919766,"gpt":0.3757555810503705,"spread":0.08849861735839387,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication","open_science"],"consensus_categories":["scholarly_communication","open_science"],"category_scores_codex":[0.002147771,0.00007645638,0.00006558477,0.00003154868,0.0009497717,0.03837242,0.01526573,0.00001304801,0.000001967453],"category_scores_gemma":[0.006170863,0.00005060746,0.00002346914,0.0006003391,0.0002592145,0.04000018,0.01025894,0.00008066493,0.00007849642],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000011219,"about_ca_system_score_gemma":0.0001221455,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000003639181,"about_ca_topic_score_gemma":0.000009160065,"domain_scores_codex":[0.9978414,0.00003923405,0.0002062888,0.0009667238,0.0006246713,0.0003216393],"domain_scores_gemma":[0.9943673,0.0004363721,0.0001152713,0.004846899,0.0001048473,0.0001293331],"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":[0.0000182429,0.00002992879,0.0001968107,0.00003707585,0.0000348169,0.00001005414,0.0002488848,0.000007382684,0.0001887925,0.4258383,0.5179107,0.05547902],"study_design_scores_gemma":[0.00007565916,0.00002441888,0.0001751642,0.000003098945,0.000003187548,0.000001173154,0.00007910952,0.04619589,0.0002360177,0.0005813912,0.9525578,0.00006707858],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0005563246,0.0004394346,0.9210995,0.06060238,0.003790907,0.0009658859,0.002989911,0.0003062158,0.009249465],"genre_scores_gemma":[0.3362477,0.000683527,0.3351711,0.002322791,0.004216848,0.0004623708,0.02062322,0.000130461,0.300142],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.5859284,"threshold_uncertainty_score":0.9977459,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","open_science"],"domain":"reproducibility","study_design":"theoretical_or_conceptual","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"high"},{"model":"gpt","categories":["scholarly_communication","open_science"],"domain":null,"study_design":"theoretical_or_conceptual","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W3037441421","doi":"10.7554/elife.54967","title":"A community-maintained standard library of population genetic models","year":2020,"lang":"en","type":"article","venue":"eLife","topic":"Evolution and Genetic Dynamics","field":"Biochemistry, Genetics and Molecular Biology","cited_by":256,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"McGill University","funders":"National Institute of General Medical Sciences; Dirección General de Asuntos del Personal Académico, Universidad Nacional Autónoma de México; Simons Center for Quantitative Biology, Cold Spring Harbor Laboratory; National Institutes of Health; University of California Institute for Mexico and the United States; Consejo Nacional de Ciencia y Tecnología; Robertson Foundation; National Human Genome Research Institute; Villum Fonden; National Evolutionary Synthesis Center","keywords":"Python (programming language); Computer science; Data science; Population; Resource (disambiguation); Inference; Artificial intelligence; Programming language","retraction":null,"screen_n_in":null,"score":{"opus":0.01426696380355115,"gpt":0.232018490238618,"spread":0.2177515264350668,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00006058688,0.00007775696,0.0001062783,0.00001666784,0.00005134346,0.000008091336,0.0001471951,0.00007373212,0.00002653111],"category_scores_gemma":[0.0000425838,0.00008051265,0.0000544684,0.00007441637,0.00003666906,0.000003691237,0.0001337772,0.00008200557,0.000003714958],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000003866754,"about_ca_system_score_gemma":0.00003729479,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001688756,"about_ca_topic_score_gemma":0.00001059805,"domain_scores_codex":[0.9993905,0.0001049105,0.0001788914,0.0001110234,0.0001157505,0.00009894571],"domain_scores_gemma":[0.9995962,0.000005047015,0.0000593879,0.0002151782,0.0000391013,0.00008509905],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.002080341,0.0005007823,0.09016239,0.0004590954,0.0003837704,0.000009293246,0.002635462,0.2960126,0.4883672,0.01008005,0.09667826,0.01263078],"study_design_scores_gemma":[0.006940594,0.006133679,0.2163144,0.00010902,0.0001661779,0.0000325931,0.002071246,0.4858447,0.1691329,0.01053898,0.1008758,0.001839895],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9586414,0.0003653968,0.03849063,0.0006863337,0.00003387714,0.0001160934,0.00007507339,0.00002978521,0.001561452],"genre_scores_gemma":[0.9904517,0.0001422706,0.007345712,0.001340999,0.00008586944,0.000003293826,0.0004676985,0.00001863241,0.000143842],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3192343,"threshold_uncertainty_score":0.3283211,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["open_science"],"domain":null,"study_design":"simulation_or_modeling","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"high"},{"model":"gpt","categories":[],"domain":null,"study_design":"not_applicable","genre":"software","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W2048204427","doi":"10.1038/513030a","title":"Satellites: Make Earth observations open access","year":2014,"lang":"en","type":"article","venue":"Nature","topic":"Space exploration and regulation","field":"Physics and Astronomy","cited_by":241,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Natural Resources Canada; University of British Columbia; Canadian Forest Service","funders":"","keywords":"Astrobiology; Earth (classical element); Remote sensing; Geology; Astronomy; Physics","retraction":null,"screen_n_in":null,"score":{"opus":0.04629024388331381,"gpt":0.3451974733286163,"spread":0.2989072294453025,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00007844525,0.00006347222,0.00007076935,0.00002726895,0.0001063259,0.0003732105,0.000363649,0.0001491576,0.0004693685],"category_scores_gemma":[0.000008705993,0.00005231784,0.00002961037,0.0001824257,0.00001117567,0.0005016145,0.0001162586,0.0003655432,0.00006453383],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000003744557,"about_ca_system_score_gemma":0.00002261041,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002419942,"about_ca_topic_score_gemma":0.00002342132,"domain_scores_codex":[0.9996005,0.00002572285,0.0000759833,0.0001246528,0.0000891569,0.00008401326],"domain_scores_gemma":[0.9996487,0.00002559939,0.00004929944,0.0001711727,0.00006846929,0.00003669447],"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":[0.000009855603,0.00007512356,0.1322406,0.000005513496,0.00003189562,1.196825e-7,0.0002529102,0.0004554982,0.001333495,0.7742662,0.02635192,0.06497684],"study_design_scores_gemma":[0.0003592836,0.00000950007,0.1913068,0.00001288567,0.000007808739,8.965802e-8,0.00008822516,0.001605714,0.003053364,0.02158269,0.7818468,0.0001268595],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5131612,0.0002377815,0.04151083,0.03030393,0.002384463,0.001365396,0.00007145126,0.0002282875,0.4107367],"genre_scores_gemma":[0.9875864,0.000001241165,0.001827039,0.001034299,0.000441317,0.00001638756,0.0002120882,0.000009187534,0.008872085],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7554948,"threshold_uncertainty_score":0.5139256,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["open_science","scholarly_communication"],"domain":null,"study_design":"not_applicable","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"},{"model":"gpt","categories":["open_science"],"domain":null,"study_design":"not_applicable","genre":"editorial","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W3134007699","doi":"10.1177/1745691620979806","title":"Estimating the Prevalence of Transparency and Reproducibility-Related Research Practices in Psychology (2014–2017)","year":2021,"lang":"en","type":"article","venue":"Perspectives on Psychological Science","topic":"Meta-analysis and systematic reviews","field":"Decision Sciences","cited_by":184,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Fonds de Recherche du Québec - Santé; Stiftung Charité; National Institute on Alcohol Abuse and Alcoholism; Einstein Stiftung Berlin; Laura and John Arnold Foundation","keywords":"Transparency (behavior); Credibility; Psychology; Confidence interval; Medicine; Clinical psychology; Social psychology; Political science; Law","retraction":null,"screen_n_in":null,"score":{"opus":0.8303977243332353,"gpt":0.6710300330829925,"spread":0.1593676912502429,"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":[0.3731061,0.000193056,0.001184311,0.0004303302,0.0004181263,0.0004345586,0.003742714,0.000101187,0.004156955],"category_scores_gemma":[0.3992323,0.00008170154,0.0003017064,0.007155955,0.00319495,0.0004016326,0.0002400073,0.0007256275,0.0003058846],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004235767,"about_ca_system_score_gemma":0.00009722513,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001543104,"about_ca_topic_score_gemma":0.00001307397,"domain_scores_codex":[0.9646918,0.01839627,0.004332855,0.00643433,0.005604085,0.0005406974],"domain_scores_gemma":[0.9702287,0.0122155,0.003299846,0.01145486,0.002620866,0.0001801872],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0002685632,0.005461874,0.6356961,0.0003300307,0.0001309211,0.000191086,0.06401878,0.0006903328,0.03464501,0.08822694,0.01190653,0.1584339],"study_design_scores_gemma":[0.0001957317,0.0002981776,0.9122776,0.00005804512,0.00002147954,0.00009668678,0.005852564,0.002069968,0.00007956713,0.0784025,0.0005227869,0.0001248653],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9479204,0.004228794,0.0005178048,0.01094838,0.0004474581,0.0005919368,0.000005176702,0.00000567789,0.03533441],"genre_scores_gemma":[0.9923798,0.0004333655,0.005687431,0.0001418952,0.00002843924,0.00003278854,1.462613e-7,0.000003458522,0.00129265],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2765816,"threshold_uncertainty_score":0.9995178,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","bibliometrics"],"domain":"reproducibility","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"high"},{"model":"gpt","categories":["metaresearch","open_science"],"domain":"reproducibility","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W4282962995","doi":"10.1038/s41597-022-01401-7","title":"A large, curated, open-source stroke neuroimaging dataset to improve lesion segmentation algorithms","year":2022,"lang":"en","type":"article","venue":"Scientific Data","topic":"Acute Ischemic Stroke Management","field":"Medicine","cited_by":149,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Research Institute for Aging; Heart and Stroke Foundation; University of Waterloo; Sunnybrook Health Science Centre; Ontario Brain Institute; University of Toronto; University of British Columbia","funders":"Eunice Kennedy Shriver National Institute of Child Health and Human Development; National Institute of Neurological Disorders and Stroke; Medical Research Council; National Institute of General Medical Sciences; Wellcome Trust; U.S. Department of Veterans Affairs","keywords":"Generalizability theory; Segmentation; Computer science; Artificial intelligence; Neuroimaging; Atlas (anatomy); Pattern recognition (psychology); Stroke (engine); Machine learning; Medicine; Mathematics; Statistics","retraction":null,"screen_n_in":null,"score":{"opus":0.06612173964472086,"gpt":0.3552075638803168,"spread":0.2890858242355959,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["open_science","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.001986715,0.0001699973,0.0002064057,0.0002688881,0.0007407909,0.0005445679,0.00234407,0.00001662714,0.0009140257],"category_scores_gemma":[0.0001650056,0.0001701369,0.00002516692,0.000814677,0.00006227687,0.0006860632,0.01677846,0.0002709288,0.000289814],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001608023,"about_ca_system_score_gemma":0.0001317039,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001371185,"about_ca_topic_score_gemma":0.00001358129,"domain_scores_codex":[0.9967853,0.00008566306,0.0003270994,0.001406676,0.000941292,0.0004539597],"domain_scores_gemma":[0.996333,0.00002922162,0.0001338777,0.003238953,0.00005859559,0.0002063541],"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":[0.00005269705,0.0001919558,0.0001941771,0.00002299952,0.00003101597,0.00005929264,0.0001792082,0.00002812342,0.067858,0.000007394835,0.9059753,0.0253998],"study_design_scores_gemma":[0.00148405,0.0001068033,0.000158371,0.00001629821,0.0001007485,0.00003645288,0.003105585,0.01720556,0.003822917,0.000002379175,0.9737935,0.0001673213],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"dataset","genre_gemma":"dataset","genre_scores_codex":[0.1046675,0.0002834095,0.09574472,0.01752212,0.0114383,0.009417011,0.7517484,0.0005270113,0.008651554],"genre_scores_gemma":[0.0663624,0.00001262857,0.0368044,0.009346709,0.0003611394,0.0003702934,0.7949009,0.0001288231,0.0917127],"genre_candidate":"dataset","genre_consensus":"dataset","teacher_disagreement_score":0.08306114,"threshold_uncertainty_score":0.9999993,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["open_science"],"domain":null,"study_design":"not_applicable","genre":"dataset","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":[],"domain":null,"study_design":"not_applicable","genre":"dataset","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W4316495494","doi":"10.12688/wellcomeopenres.18681.1","title":"Pf7: an open dataset of Plasmodium falciparum genome variation in 20,000 worldwide samples","year":2023,"lang":"en","type":"preprint","venue":"Wellcome Open Research","topic":"Malaria Research and Control","field":"Medicine","cited_by":148,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Calgary","funders":"Faculty of Tropical Medicine, Mahidol University; Addis Ababa University; National Institutes of Health; Universiteit Antwerpen; National Institute for Medical Research; Ministère de la Santé; Universidad Nacional de Colombia; Bundesministerium für Gesundheit; World Health Organization; University of Ghana; Bill and Melinda Gates Foundation; Mahidol University; University College London; National Center for Advancing Translational Sciences; National Human Genome Research Institute; Wellcome Trust; National Institute of Allergy and Infectious Diseases; Medical Research Council; London School of Hygiene and Tropical Medicine","keywords":"Plasmodium falciparum; Variation (astronomy); Biology; Genome; Computational biology; Structural variation; Genetics; Malaria; Evolutionary biology; Gene; Immunology","retraction":null,"screen_n_in":null,"score":{"opus":0.2762833932571187,"gpt":0.4550705055030257,"spread":0.178787112245907,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","scholarly_communication","open_science","research_integrity","insufficient_payload"],"consensus_categories":["open_science"],"category_scores_codex":[0.01816784,0.000439904,0.001575739,0.001918708,0.00023526,0.001359556,0.008689972,0.0005284358,0.004493767],"category_scores_gemma":[0.001745661,0.0004068643,0.0001064208,0.001460413,0.0002963298,0.0005691056,0.02100332,0.003414029,0.0007237471],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005269145,"about_ca_system_score_gemma":0.002961498,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0900839,"about_ca_topic_score_gemma":0.01431276,"domain_scores_codex":[0.990455,0.002302214,0.001351933,0.001817248,0.002458382,0.001615211],"domain_scores_gemma":[0.9934971,0.001107666,0.0002981749,0.003586749,0.0006885637,0.0008217653],"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":[0.08741723,0.01809392,0.06750414,0.01517068,0.005315058,0.006949323,0.006867888,0.008528499,0.1450857,0.05927054,0.5502121,0.02958498],"study_design_scores_gemma":[0.01785034,0.003260144,0.6207926,0.0026003,0.0001736963,0.00002724736,0.001124299,0.01474427,0.001210906,0.02892185,0.307861,0.001433397],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.3634628,0.002403042,0.001929642,0.16551,0.002311046,0.1287945,0.1951501,0.0004509519,0.139988],"genre_scores_gemma":[0.8747709,0.001327725,0.005172191,0.0003642733,0.0007022879,0.005269598,0.0807943,0.0002836944,0.03131505],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5532884,"threshold_uncertainty_score":0.9998383,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":[],"domain":null,"study_design":"not_applicable","genre":"dataset","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["open_science"],"domain":null,"study_design":"not_applicable","genre":"dataset","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"split"},{"id":"W2001314811","doi":"10.1038/nchembio0709-436","title":"Open access chemical and clinical probes to support drug discovery","year":2009,"lang":"en","type":"article","venue":"Nature Chemical Biology","topic":"Pharmacogenetics and Drug Metabolism","field":"Pharmacology, Toxicology and Pharmaceutics","cited_by":132,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Structural Genomics Consortium; University of Toronto","funders":"Canadian Institutes of Health Research; Wellcome Trust","keywords":"Drug discovery; Pharmaceutical industry; Productivity; Data science; Business; Drug industry; Computational biology; Computer science; Biology; Biotechnology; Bioinformatics; Economics","retraction":null,"screen_n_in":null,"score":{"opus":0.134351923613858,"gpt":0.5619974727123195,"spread":0.4276455490984615,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","research_integrity"],"consensus_categories":["research_integrity"],"category_scores_codex":[0.0009726431,0.0003523838,0.0006642779,0.00008946157,0.0001007943,0.0001813136,0.00224311,0.001389735,0.000446682],"category_scores_gemma":[0.0004870805,0.0002885174,0.0001467608,0.0002815676,0.0003460156,0.0003064798,0.001916398,0.002484904,0.00007566068],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003542399,"about_ca_system_score_gemma":0.0001333867,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000006378462,"about_ca_topic_score_gemma":0.000001946804,"domain_scores_codex":[0.9973547,0.0003024321,0.00061134,0.0009277374,0.0001134371,0.0006903628],"domain_scores_gemma":[0.9982563,0.0004858274,0.000136933,0.000309757,0.0001048558,0.0007063344],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.0009919782,0.000704637,0.01416758,0.00001522682,0.0001395384,0.00002857764,0.0000732976,0.000001018624,0.7873629,0.005056832,0.05536571,0.1360927],"study_design_scores_gemma":[0.002005464,0.0001150356,0.002729224,0.000006801678,0.0001322516,0.00003150999,0.000006475084,0.00003343862,0.5688101,0.007195745,0.4185392,0.0003947646],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9767361,0.001414241,0.00002079866,0.01110415,0.001459136,0.0007347016,0.0001496914,0.00007400609,0.008307201],"genre_scores_gemma":[0.9497935,0.0006766926,0.0007257066,0.04703065,0.001094645,0.00003906109,0.0001442027,0.00002081119,0.0004747339],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3631735,"threshold_uncertainty_score":0.9999567,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["open_science"],"domain":null,"study_design":"not_applicable","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["open_science","scholarly_communication"],"domain":null,"study_design":"design_other","genre":"other","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"split"},{"id":"W2144889117","doi":"10.1186/gm262","title":"Towards a data sharing Code of Conduct for international genomic research","year":2011,"lang":"en","type":"article","venue":"Genome Medicine","topic":"Ethics in Clinical Research","field":"Medicine","cited_by":112,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"McGill University","funders":"Wellcome Trust","keywords":"Biobank; Data sharing; Context (archaeology); Genomics; Data science; Engineering ethics; Knowledge management; Computer science; Medicine; Bioinformatics; Biology; Engineering; Genetics; Genome","retraction":null,"screen_n_in":null,"score":{"opus":0.958708070608172,"gpt":0.7003965024361307,"spread":0.2583115681720413,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0187038,0.0001212919,0.0004701631,0.000394321,0.00007509421,0.000006949256,0.0022833,0.0002105415,0.002567478],"category_scores_gemma":[0.03852342,0.00009269559,0.00005732399,0.0002952456,0.001080761,0.00007482247,0.001744884,0.001297337,0.00005334186],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001419768,"about_ca_system_score_gemma":0.0006512338,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0004930306,"about_ca_topic_score_gemma":0.0001230145,"domain_scores_codex":[0.9965,0.00006679207,0.0007739889,0.0007295107,0.00147067,0.0004590987],"domain_scores_gemma":[0.9926489,0.002485879,0.0001297148,0.002029075,0.002403793,0.0003026182],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.009348216,0.002042695,0.05293424,0.005823156,0.002514904,0.0003663152,0.02169597,0.000003435071,0.6035852,0.2180013,0.03506281,0.04862175],"study_design_scores_gemma":[0.0247478,0.01126903,0.2511528,0.003164681,0.0005721833,0.0001893332,0.006517313,0.004037544,0.01511169,0.3612263,0.3212414,0.0007699398],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5842754,0.006193476,0.0111171,0.09017685,0.003906605,0.005493024,0.001221614,0.0001820534,0.2974338],"genre_scores_gemma":[0.9697581,0.00143781,0.01868813,0.0006374232,0.001893418,0.00006372659,0.0005102215,0.00005825837,0.006952896],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5884735,"threshold_uncertainty_score":0.9983443,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["open_science"],"domain":null,"study_design":"not_applicable","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["open_science"],"domain":null,"study_design":"theoretical_or_conceptual","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"split"},{"id":"W2989484692","doi":"10.1007/s11192-019-03301-x","title":"Do articles in open access journals have more frequent altmetric activity than articles in subscription-based journals? An investigation of the research output of Finnish universities","year":2019,"lang":"en","type":"article","venue":"Scientometrics","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":105,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"Turun Yliopisto","keywords":"Altmetrics; Discipline; Scientometrics; Bibliometrics; Social media; Library science; Computer science; Sociology; Social science; World Wide Web","retraction":null,"screen_n_in":null,"score":{"opus":0.838152280754456,"gpt":0.6455667087672443,"spread":0.1925855719872117,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","scholarly_communication","open_science"],"consensus_categories":["metaresearch","bibliometrics"],"category_scores_codex":[0.1082317,0.0002111957,0.0006495731,0.2690909,0.0003236766,0.007054362,0.0114311,0.0001994283,0.0004386413],"category_scores_gemma":[0.04319692,0.0001450322,0.000194003,0.6152785,0.0009399091,0.005927231,0.003343812,0.0009296827,0.00002469403],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009683496,"about_ca_system_score_gemma":0.001984661,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004475062,"about_ca_topic_score_gemma":0.0009897391,"domain_scores_codex":[0.9737656,0.002431782,0.001498649,0.001011109,0.02021208,0.001080846],"domain_scores_gemma":[0.9783849,0.008512352,0.001221544,0.001751825,0.009483315,0.0006460279],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0000714087,0.0004190861,0.9661016,0.00002050067,0.000007404975,0.000005940251,0.001483423,0.003067941,0.007809614,0.0001659838,0.0003029797,0.0205441],"study_design_scores_gemma":[0.001086377,0.0002137827,0.9348478,0.0001193549,0.000004348318,0.000001658204,0.008116978,0.01944543,0.02996376,0.005993111,0.00004977768,0.0001576139],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9955313,0.0008126932,0.0002076451,0.001303783,0.0004433122,0.0009106401,0.00006028094,0.000006311641,0.0007239901],"genre_scores_gemma":[0.9989226,0.0001418545,0.0003432622,0.00006284908,0.0000231433,0.000008554824,0.000001501654,0.00001389719,0.0004822791],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3461877,"threshold_uncertainty_score":0.9939764,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","bibliometrics","open_science"],"domain":"evaluation","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["bibliometrics","open_science"],"domain":null,"study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W3136742816","doi":"10.1016/j.respol.2021.104226","title":"The fall of the innovation empire and its possible rise through open science","year":2021,"lang":"en","type":"article","venue":"Research Policy","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":88,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"McGill University","funders":"Bill and Melinda Gates Foundation","keywords":"Incentive; Productivity; Intellectual property; General partnership; Open innovation; Economics; Open science; Knowledge creation; Public economics; Business; Economic growth; Political science; Market economy; Management; Law; Operations management","retraction":null,"screen_n_in":null,"score":{"opus":0.8515611675883188,"gpt":0.720693447368452,"spread":0.1308677202198668,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","sts","scholarly_communication","open_science"],"consensus_categories":["metaresearch","bibliometrics","open_science"],"category_scores_codex":[0.07696775,0.00008472808,0.0001743508,0.0182587,0.002545938,0.007745758,0.009931601,0.00006254338,0.00006138488],"category_scores_gemma":[0.321319,0.00003976992,0.00004765451,0.4948857,0.001941669,0.001280575,0.01325281,0.0004928494,0.0000747888],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001979736,"about_ca_system_score_gemma":0.007047017,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001407805,"about_ca_topic_score_gemma":0.000168996,"domain_scores_codex":[0.982298,0.000953059,0.0005771977,0.0006708198,0.01453713,0.000963764],"domain_scores_gemma":[0.9599999,0.009186872,0.0002527005,0.001968057,0.02827748,0.0003149513],"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":[0.00002907369,0.000153328,0.03476098,0.00001398899,0.00001292193,0.000007645207,0.0009934036,0.000009386551,0.03160311,0.7232306,0.03531572,0.1738698],"study_design_scores_gemma":[0.0006446131,0.0001689742,0.3388201,0.00004062134,0.00000171205,0.00002205175,0.002443034,0.00262952,0.1219528,0.3270071,0.2061066,0.0001628724],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.893646,0.00177158,0.00002971022,0.04207265,0.0001502972,0.0005310362,0.00002801428,0.000004890895,0.06176579],"genre_scores_gemma":[0.9848312,0.0008296206,0.0001343753,0.0002678487,0.00008702253,0.00001843407,3.452256e-7,0.000006394176,0.01382481],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.476627,"threshold_uncertainty_score":0.9987526,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","open_science"],"domain":"incentives","study_design":"theoretical_or_conceptual","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["open_science"],"domain":null,"study_design":"theoretical_or_conceptual","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W2136741769","doi":"10.1080/02681102.2015.1081868","title":"Viscous Open Data: The Roles of Intermediaries in an Open Data Ecosystem","year":2015,"lang":"en","type":"article","venue":"Information Technology for Development","topic":"E-Government and Public Services","field":"Social Sciences","cited_by":88,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"International Development Research Centre","keywords":"Intermediary; Open data; Business; Ecosystem; Environmental resource management; Knowledge management; Computer science; World Wide Web; Marketing; Environmental science; Ecology","retraction":null,"screen_n_in":null,"score":{"opus":0.1319685321627065,"gpt":0.3819785333272386,"spread":0.2500100011645321,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["open_science"],"consensus_categories":[],"category_scores_codex":[0.003669295,0.00007491888,0.0001575124,0.0001381527,0.0002160012,0.0004596711,0.01090221,0.000123926,0.00001765003],"category_scores_gemma":[0.0004272423,0.00005555814,0.000004187132,0.0003337891,0.0002116564,0.007721395,0.007050588,0.00007670867,0.00001902522],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001258743,"about_ca_system_score_gemma":0.0009812133,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0005753241,"about_ca_topic_score_gemma":0.09442689,"domain_scores_codex":[0.9988743,0.00004403071,0.0004707223,0.0001455321,0.0002800994,0.0001852808],"domain_scores_gemma":[0.9986237,0.00006665379,0.0003153466,0.0008086778,0.0001450199,0.00004055773],"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":[0.0001593641,0.0001509944,0.02419291,0.00008664742,0.00009598296,6.264702e-7,0.1511762,0.00001314641,0.000004216891,0.3231123,0.03027828,0.4707294],"study_design_scores_gemma":[0.0005252569,0.00003842076,0.0004427772,0.00003527969,0.000004142978,5.623567e-7,0.1422239,0.0005827509,0.0001878229,0.002896492,0.8529686,0.00009393564],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8275195,0.0003190894,0.007995857,0.0856018,0.002240911,0.01155581,0.001559698,0.0004496069,0.06275777],"genre_scores_gemma":[0.9899163,0.00002167126,0.008100948,0.0003035451,0.00002771067,0.0002114544,0.001294869,0.000004995818,0.0001185352],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.8226904,"threshold_uncertainty_score":0.9944493,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["open_science"],"domain":null,"study_design":"qualitative","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["open_science"],"domain":null,"study_design":"theoretical_or_conceptual","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"},{"id":"W3194760139","doi":"10.1038/s41597-021-00941-8","title":"Open-access quantitative MRI data of the spinal cord and reproducibility across participants, sites and manufacturers","year":2021,"lang":"en","type":"article","venue":"Scientific Data","topic":"Advanced Neuroimaging Techniques and Applications","field":"Medicine","cited_by":77,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Philips (Canada); McGill University; Centre Hospitalier Universitaire de Sherbrooke; International Collaboration On Repair Discoveries; University of British Columbia; Université de Montréal; Université de Sherbrooke; Centre Hospitalier Universitaire Sainte-Justine; Polytechnique Montréal; Montreal Neurological Institute and Hospital; Mila - Quebec Artificial Intelligence Institute","funders":"National Institute of Neurological Disorders and Stroke; Staatssekretariat für Bildung, Forschung und Innovation; Economic and Social Research Council; Engineering and Physical Sciences Research Council; Natural Sciences and Engineering Research Council of Canada; National Institute of Biomedical Imaging and Bioengineering; Canadian Institutes of Health Research; CRIS Cancer Foundation; Regione Puglia; University College London Hospitals NHS Foundation Trust; Ministero dell’Istruzione, dell’Università e della Ricerca; Ministero della Salute; Concordia University; Agentura Pro Zdravotnický Výzkum České Republiky; National Institutes of Health; Rosetrees Trust; European Commission; Multiple Sclerosis Society; Bundesministerium für Bildung und Forschung; National Imaging Facility; National Institute for Health and Care Research; University of Pennsylvania; SpinalCure Australia; University of Minnesota; National Science Foundation; Compute Canada; Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung; Polytechnique Montréal; Wellcome Trust; Institut de Valorisation des Données; AstraZeneca; Craig H. Neilsen Foundation; McGill University; Canada First Research Excellence Fund; Max-Planck-Gesellschaft","keywords":"Protocol (science); Reproducibility; Computer science; Spinal cord; Documentation; Data mining; Medicine; Statistics; Mathematics; Pathology","retraction":null,"screen_n_in":null,"score":{"opus":0.6245720496163727,"gpt":0.5767339061992035,"spread":0.04783814341716919,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["open_science"],"consensus_categories":[],"category_scores_codex":[0.001738994,0.00007491081,0.0001623053,0.00001396586,0.0002786198,0.0003824537,0.002097608,0.00001643681,0.00001762368],"category_scores_gemma":[0.001616848,0.00005197949,0.000009079192,0.0003406774,0.0008719899,0.000733886,0.01495176,0.0001119396,0.000001182314],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000006068723,"about_ca_system_score_gemma":0.0001030831,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00005158479,"about_ca_topic_score_gemma":0.0001394123,"domain_scores_codex":[0.9976369,0.00004481999,0.0001983786,0.00179074,0.0001634488,0.0001657006],"domain_scores_gemma":[0.9899572,0.00004492495,0.0001104097,0.009720027,0.00009863014,0.00006886664],"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":[0.00120456,0.001297933,0.2341873,0.001108011,0.0001555549,0.00006166913,0.0009659761,0.000004969262,0.2213653,0.005927698,0.4161961,0.117525],"study_design_scores_gemma":[0.000780884,0.0001707515,0.5823219,0.0003249721,0.0001997043,0.0001052075,0.0009842508,0.002793869,0.1386297,0.008398251,0.2650311,0.0002594427],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9771098,0.0008895358,0.00205192,0.009303352,0.0002566038,0.0007403863,0.009503441,0.00003713573,0.0001077869],"genre_scores_gemma":[0.9789385,0.00009919237,0.01814502,0.0002599098,0.00001516891,0.000009316887,0.001756869,0.000009039199,0.0007669806],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3481346,"threshold_uncertainty_score":0.9930151,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","open_science"],"domain":"reproducibility","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"high"},{"model":"gpt","categories":["metaresearch","open_science"],"domain":"reproducibility","study_design":"not_applicable","genre":"dataset","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W2796216221","doi":"10.1177/0963662518761733","title":"Identifying diffusion patterns of research articles on Twitter: A case study of online engagement with open access articles","year":2018,"lang":"en","type":"article","venue":"Public Understanding of Science","topic":"Complex Network Analysis Techniques","field":"Physics and Astronomy","cited_by":76,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Ottawa; Simon Fraser University","funders":"Social Sciences and Humanities Research Council of Canada","keywords":"Open science; Social media; Science communication; Sociology; Scholarly communication; Citizen science; Ask price; World Wide Web; Altmetrics; Online community; Public engagement; Data science; Public relations; Computer science; Political science; Biology; Science education; Publishing; Business","retraction":null,"screen_n_in":null,"score":{"opus":0.50896915729909,"gpt":0.496626232536117,"spread":0.01234292476297294,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002881206,0.00009728035,0.000231788,0.0007436593,0.0005672697,0.0004536173,0.001341408,0.000009673142,0.0001286192],"category_scores_gemma":[0.00003649474,0.00007493598,0.0000300079,0.002437967,0.001170432,0.0008551654,0.001680405,0.0001352011,3.589174e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001150339,"about_ca_system_score_gemma":0.0001372697,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005020816,"about_ca_topic_score_gemma":0.001261922,"domain_scores_codex":[0.9978277,0.0001657172,0.0003713213,0.0003587165,0.0009465833,0.0003299293],"domain_scores_gemma":[0.9984764,0.0001961748,0.0002840087,0.0004961622,0.0004572796,0.00009000579],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"qualitative","study_design_scores_codex":[0.00006484748,0.00308025,0.9437094,0.00003387287,0.00008888314,0.00001453184,0.01517381,0.00002697074,0.01169206,0.02102828,0.00006047397,0.005026561],"study_design_scores_gemma":[0.002666369,0.004180921,0.08483757,0.0008741533,0.0001098891,0.00001165853,0.7495735,0.0130032,0.1070031,0.03728257,0.00001963811,0.0004374775],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.967214,0.000001843534,0.03166658,0.000146102,0.00001781885,0.0004348508,0.000005407892,0.00001310858,0.0005002314],"genre_scores_gemma":[0.9992802,7.697605e-7,0.0006444073,0.000006315426,0.00003245316,0.00001299439,0.000001083108,0.000007875936,0.00001390155],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.8588719,"threshold_uncertainty_score":0.7590005,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["scholarly_communication","open_science"],"domain":null,"study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["bibliometrics","scholarly_communication"],"domain":null,"study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W4392914547","doi":"10.1016/j.eclinm.2024.102551","title":"Health data sharing attitudes towards primary and secondary use of data: a systematic review","year":2024,"lang":"en","type":"review","venue":"EClinicalMedicine","topic":"Ethics in Clinical Research","field":"Medicine","cited_by":56,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"","keywords":"CINAHL; PsycINFO; Medicine; Data sharing; MEDLINE; Data extraction; Family medicine; Qualitative research; Health care; Public health; Medical education; Alternative medicine; Psychological intervention; Nursing; Pathology","retraction":null,"screen_n_in":null,"score":{"opus":0.9055626380658633,"gpt":0.7122511745234683,"spread":0.193311463542395,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","metaepi_narrow","metaepi_broad","open_science","research_integrity"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.05130988,0.0007612832,0.01509079,0.0003673184,0.00007372333,0.00007561388,0.004078801,0.001012754,0.0001788557],"category_scores_gemma":[0.1316013,0.0004576911,0.0005106792,0.0009040512,0.001068449,0.0003863435,0.0141226,0.008906658,0.0001407054],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002265247,"about_ca_system_score_gemma":0.008982263,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001483685,"about_ca_topic_score_gemma":0.00007686352,"domain_scores_codex":[0.9840281,0.0009995457,0.008571987,0.003174273,0.002553192,0.0006728891],"domain_scores_gemma":[0.9285204,0.04595831,0.002939353,0.02056152,0.0006062152,0.001414165],"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":[0.00001300505,0.0001657036,0.00001198828,0.7859933,0.001262537,0.0001838103,0.000009364134,1.001301e-9,1.123326e-8,0.0002172782,0.02719049,0.1849525],"study_design_scores_gemma":[0.0003396286,0.0004222092,0.00001979479,0.670038,0.01201654,0.0003178749,0.000003196187,0.00001833502,2.659734e-9,0.0001131598,0.3165144,0.0001968746],"study_design_candidate":"systematic_review","study_design_consensus":"systematic_review","genre_codex":"review","genre_gemma":"review","genre_scores_codex":[2.067622e-7,0.9811065,0.00004729566,0.01004127,0.000509563,0.005964923,0.001835102,0.0001066545,0.0003884785],"genre_scores_gemma":[5.67793e-7,0.9767191,0.004251485,0.007556071,0.0008746603,0.0001018626,0.007773719,0.0001735027,0.002549015],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.2893239,"threshold_uncertainty_score":0.9997875,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["open_science"],"domain":null,"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":"W2607649737","doi":"10.1108/tg-12-2015-0053","title":"Conceptualizing citizen participation in open data use at the city level","year":2017,"lang":"en","type":"article","venue":"Transforming Government People Process and Policy","topic":"E-Government and Public Services","field":"Social Sciences","cited_by":53,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"HEC Montréal","funders":"","keywords":"Open data; Open government; Conceptualization; e-participation; Public relations; Government (linguistics); Construct (python library); Knowledge management; Citizen science; Business; Political science; Computer science","retraction":null,"screen_n_in":null,"score":{"opus":0.2134426164595318,"gpt":0.4431207390166106,"spread":0.2296781225570789,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["sts","scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.001257662,0.0001307154,0.0001785632,0.0000145143,0.001967276,0.00104549,0.001732817,0.00007081149,0.0001607856],"category_scores_gemma":[0.0004146059,0.0001011303,0.0000209332,0.0001212944,0.0002924826,0.003057608,0.0006549243,0.0001025181,0.000003776521],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000229874,"about_ca_system_score_gemma":0.0002170404,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.1165633,"about_ca_topic_score_gemma":0.5638405,"domain_scores_codex":[0.9982849,0.00007572777,0.0002373917,0.0003230736,0.0006834826,0.0003954593],"domain_scores_gemma":[0.9989285,0.0002009773,0.0002054257,0.0005258395,0.00002569561,0.0001135744],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00009097217,0.00007834834,0.8222034,0.00006358605,0.00002618653,0.000001271474,0.1096076,0.000005395431,0.00003531312,0.05076271,0.0002498838,0.01687531],"study_design_scores_gemma":[0.002084095,0.00005431389,0.8332725,0.0001562469,0.00006520759,0.000001314097,0.048237,0.0006012035,0.0004660671,0.003261034,0.1112778,0.0005232149],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9441911,0.00007863378,0.00004184321,0.01662586,0.00006575121,0.0005383333,0.0002979552,0.00001593203,0.03814456],"genre_scores_gemma":[0.9955727,0.0005356208,0.00002018592,0.0007234915,0.0001812477,0.00006107263,0.00002244905,0.00001026593,0.0028729],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4472772,"threshold_uncertainty_score":0.9999915,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["open_science"],"domain":null,"study_design":"qualitative","genre":"empirical","about_ca_system":false,"about_ca_topic":true,"confidence":"high"},{"model":"gpt","categories":[],"domain":null,"study_design":"qualitative","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W2043763793","doi":"10.1371/journal.pone.0097886","title":"Sharing Individual Participant Data from Clinical Trials: An Opinion Survey Regarding the Establishment of a Central Repository","year":2014,"lang":"en","type":"article","venue":"PLoS ONE","topic":"Meta-analysis and systematic reviews","field":"Decision Sciences","cited_by":52,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Medical Research Council; University of Birmingham; Queen's University; Cancer Research UK","keywords":"Safeguarding; Clinical trial; Data sharing; Workload; Medicine; Information repository; Systematic review; Resource (disambiguation); MEDLINE; Alternative medicine; Computer science; Nursing; Pathology; Political science","retraction":null,"screen_n_in":null,"score":{"opus":0.9754389651248165,"gpt":0.6107774621438413,"spread":0.3646615029809752,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","scholarly_communication"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.6082692,0.0001918343,0.004514113,0.00009412284,0.0001526647,0.001138763,0.005037323,0.00009396602,0.0008420367],"category_scores_gemma":[0.3519663,0.00008102434,0.0006761889,0.000448914,0.0000864652,0.0004646254,0.0009303371,0.0002311298,0.0001310869],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000009410888,"about_ca_system_score_gemma":0.00006804799,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0003187879,"about_ca_topic_score_gemma":0.0002415642,"domain_scores_codex":[0.7854382,0.1716076,0.02818607,0.00266045,0.0114942,0.00061346],"domain_scores_gemma":[0.8058426,0.136282,0.02410384,0.03109752,0.001860719,0.0008133398],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00006788341,0.001235264,0.9691786,0.00004135166,0.003236508,0.000001193313,0.001006524,0.00003224213,0.0003571811,0.0002442759,0.01725192,0.007347029],"study_design_scores_gemma":[0.000706616,0.0001613638,0.8253631,0.0003635198,0.004140388,9.706517e-7,0.0007806252,0.1622934,0.0003983456,0.00131045,0.004227715,0.0002535824],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.993377,0.0008387049,0.003399059,0.0002726731,0.0005726212,0.0007668837,0.000264242,0.000007552153,0.0005013076],"genre_scores_gemma":[0.9958612,0.00003652031,0.002897696,0.00008767275,0.0006299022,0.00001762677,0.0002141498,0.000009988876,0.0002452528],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.256303,"threshold_uncertainty_score":0.9998981,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","open_science"],"domain":"reproducibility","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","scholarly_communication","open_science"],"domain":"reproducibility","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W4379599895","doi":"10.1111/hex.13779","title":"Exploring the impact of engagement in mental health and substance use research: A scoping review and thematic analysis","year":2023,"lang":"en","type":"review","venue":"Health Expectations","topic":"Mental Health and Patient Involvement","field":"Health Professions","cited_by":49,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"Centre for Addiction and Mental Health; Trillium Health Centre; University of Toronto","funders":"Canadian Institutes of Health Research; Institute of Neurosciences, Mental Health and Addiction; Centre for Addiction and Mental Health","keywords":"Thematic analysis; Mental health; Substance use; Psychology; Thematic map; Applied psychology; Qualitative research; Sociology; Psychotherapist; Psychiatry; Geography; Social science","retraction":null,"screen_n_in":null,"score":{"opus":0.9341418747535614,"gpt":0.6724996206695402,"spread":0.2616422540840212,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["sts"],"consensus_categories":[],"category_scores_codex":[0.005944661,0.0003646002,0.002668477,0.001256186,0.002100478,0.0000191857,0.0002199354,0.00007856861,0.00004259076],"category_scores_gemma":[0.0003665773,0.0002399495,0.0002641287,0.003912226,0.0001185197,0.0002111167,0.0001871837,0.001333565,0.00002902642],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00148695,"about_ca_system_score_gemma":0.003354757,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00624388,"about_ca_topic_score_gemma":0.007882901,"domain_scores_codex":[0.9862319,0.008367386,0.003099157,0.0005765516,0.0005924772,0.001132591],"domain_scores_gemma":[0.9914821,0.005621632,0.001551679,0.0006383364,0.0001290782,0.0005771532],"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":[0.000009663092,0.000117514,0.0008424281,0.8401853,0.000341334,0.000001146366,0.1062458,0.000001543113,6.184651e-9,0.0003330423,0.001189247,0.05073297],"study_design_scores_gemma":[0.0003574864,0.000422657,0.0003985193,0.9570045,0.0002691378,0.00000131675,0.03571717,0.00003618763,7.666015e-9,0.00004465268,0.005519062,0.0002292656],"study_design_candidate":"systematic_review","study_design_consensus":"systematic_review","genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.0008646494,0.9811711,0.000007611891,0.0007525513,0.0001362711,0.01667986,0.0003209838,0.00003623331,0.00003069756],"genre_scores_gemma":[0.0001109204,0.9910741,0.000195967,0.0005067531,0.00003748148,0.007241278,0.0007087441,0.00005175543,0.00007292991],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.1168193,"threshold_uncertainty_score":0.9991987,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch"],"domain":"methods","study_design":"systematic_review","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"high"},{"model":"gpt","categories":["metaresearch","open_science"],"domain":"methods","study_design":"systematic_review","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W4378782503","doi":"10.1146/annurev-genom-110122-084756","title":"Federated Analysis for Privacy-Preserving Data Sharing: A Technical and Legal Primer","year":2023,"lang":"en","type":"review","venue":"Annual Review of Genomics and Human Genetics","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":42,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"McGill University","funders":"National Human Genome Research Institute; National Cancer Institute; National Heart, Lung, and Blood Institute; National Institutes of Health","keywords":"Data sharing; Information privacy; Internet privacy; Computer science; Data Protection Act 1998; Data science; Computer security","retraction":null,"screen_n_in":null,"score":{"opus":0.1519487372290538,"gpt":0.3996860906048327,"spread":0.2477373533757789,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","open_science"],"consensus_categories":["open_science"],"category_scores_codex":[0.002004808,0.0005558472,0.002672846,0.0004768522,0.0002727207,0.0004083738,0.03524897,0.0003739454,0.000004558613],"category_scores_gemma":[0.005766435,0.0004941375,0.0003453064,0.001173622,0.0002094285,0.0002999952,0.2289358,0.0004459782,0.000003497508],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00006654437,"about_ca_system_score_gemma":0.0002601827,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003646558,"about_ca_topic_score_gemma":0.00004105252,"domain_scores_codex":[0.9957032,0.0001001092,0.00156343,0.001830174,0.0003114482,0.0004916551],"domain_scores_gemma":[0.9852834,0.0003494006,0.0009748556,0.01301533,0.0002335927,0.0001434446],"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":[0.000001875493,0.00007046136,0.00002656381,0.1452143,0.002026627,0.00001654282,0.0000197127,4.241327e-7,0.000005242584,0.004290755,0.05271924,0.7956083],"study_design_scores_gemma":[0.0001066724,0.0001052429,0.00002695604,0.01621167,0.002959738,0.00002682691,0.000004594605,0.003702411,0.000001611976,0.0051535,0.9711431,0.0005576868],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.00001199345,0.9778382,0.01830792,0.0006297086,0.00006907827,0.001532248,0.001366618,0.0001893079,0.00005488583],"genre_scores_gemma":[0.000004899063,0.9028638,0.09574673,0.00007862729,0.0000622487,0.00008474556,0.001029038,0.00006352933,0.00006640331],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.9184238,"threshold_uncertainty_score":0.999751,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["open_science"],"domain":null,"study_design":"theoretical_or_conceptual","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":[],"domain":null,"study_design":"not_applicable","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"},{"id":"W2124995107","doi":"10.2471/blt.09.064659","title":"Open access: a giant leap towards bridging health inequities","year":2009,"lang":"en","type":"article","venue":"Bulletin of the World Health Organization","topic":"Health Sciences Research and Education","field":"Health Professions","cited_by":40,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"University of Toronto","funders":"University of Warwick; University of Southampton","keywords":"Bridging (networking); Public health; Health equity; Public relations; Political science; Medicine; Computer science; Computer security; Nursing","retraction":null,"screen_n_in":null,"score":{"opus":0.154399671463718,"gpt":0.5057955029556225,"spread":0.3513958314919045,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["sts","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.005095324,0.0001703062,0.0004845078,0.0002889431,0.002828204,0.0001447849,0.002160054,0.00007156246,0.004150126],"category_scores_gemma":[0.002031071,0.0001269879,0.00003763337,0.002462818,0.00009515479,0.0002643753,0.001059752,0.0006204669,0.0003177066],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001057201,"about_ca_system_score_gemma":0.01025206,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01543954,"about_ca_topic_score_gemma":0.002588237,"domain_scores_codex":[0.9951163,0.001449845,0.00131081,0.0003863588,0.0006545859,0.001082128],"domain_scores_gemma":[0.9968941,0.0002610958,0.001195902,0.0006039603,0.0005668059,0.0004781407],"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":[0.0000265412,0.0001032026,0.009331852,0.000576366,0.000003659909,1.477205e-7,0.003864378,0.00003264675,0.00002319991,0.004623667,0.9650899,0.01632444],"study_design_scores_gemma":[0.0006341153,0.0002137528,0.3969773,0.001214586,0.000003223281,0.000001932977,0.001640582,0.0001027838,0.0001169517,0.000967927,0.5979648,0.0001619908],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"commentary","genre_gemma":"empirical","genre_scores_codex":[0.0100295,0.0005079362,0.0009687957,0.970761,0.001549005,0.003418363,0.00002298527,0.0001106285,0.01263179],"genre_scores_gemma":[0.7497126,0.001062444,0.003152954,0.2008015,0.001115605,0.0000811159,0.00009341521,0.00005798001,0.04392237],"genre_candidate":"commentary","genre_consensus":null,"teacher_disagreement_score":0.7699594,"threshold_uncertainty_score":0.9984699,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["open_science"],"domain":null,"study_design":"not_applicable","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["scholarly_communication"],"domain":null,"study_design":"theoretical_or_conceptual","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"},{"id":"W2772565825","doi":"10.1038/nbt.4029","title":"The international MAQC Society launches to enhance reproducibility of high-throughput technologies","year":2017,"lang":"en","type":"article","venue":"Nature Biotechnology","topic":"Scientific Computing and Data Management","field":"Decision Sciences","cited_by":38,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Princess Margaret Cancer Centre; University of Toronto; Ontario Institute for Cancer Research","funders":"Biotechnology and Biological Sciences Research Council","keywords":"Reproducibility; Throughput; Emerging technologies; Computer science; Telecommunications; Chemistry; Artificial intelligence; Chromatography","retraction":null,"screen_n_in":null,"score":{"opus":0.05289899810702355,"gpt":0.4077293677463387,"spread":0.3548303696393151,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","open_science"],"consensus_categories":[],"category_scores_codex":[0.01103869,0.0001318836,0.0002421023,0.0001553209,0.0007828893,0.0004653905,0.008209341,0.0005605735,0.00002170605],"category_scores_gemma":[0.06130109,0.00007829646,0.0001253,0.0005063119,0.001167184,0.0001677583,0.005798987,0.0006379738,0.00009658049],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00005710564,"about_ca_system_score_gemma":0.00004306081,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00007402708,"about_ca_topic_score_gemma":0.0001693739,"domain_scores_codex":[0.9960164,0.0000365017,0.0005121289,0.002180304,0.0009695899,0.0002850085],"domain_scores_gemma":[0.986064,0.0005401751,0.0005194953,0.01247542,0.0003778076,0.00002310617],"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":[0.00002594762,0.00004612572,0.001544893,0.000002940367,0.00003980842,0.000002051907,0.00007360833,0.0000114835,0.006108864,0.05273252,0.2513792,0.6880326],"study_design_scores_gemma":[0.0000844176,0.00004514004,0.01008048,0.00001292494,0.000005248515,0.000002768251,0.0008210755,0.0002055221,0.2287696,0.09868699,0.6611718,0.0001140923],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.495826,0.0008616589,0.0103361,0.4816203,0.008093616,0.0004320036,0.0000738792,0.0005622214,0.002194157],"genre_scores_gemma":[0.9852257,0.00007913823,0.01228542,0.0002212558,0.00007135484,0.000009837156,0.000002655467,0.000004959682,0.002099643],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6879185,"threshold_uncertainty_score":0.9971567,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","open_science"],"domain":"reproducibility","study_design":"not_applicable","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"high"},{"model":"gpt","categories":["metaresearch"],"domain":"reproducibility","study_design":"not_applicable","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W3091824907","doi":"10.1016/s2666-5247(20)30122-1","title":"A multinational listeriosis outbreak and the importance of sharing genomic data","year":2020,"lang":"en","type":"article","venue":"The Lancet Microbe","topic":"Listeria monocytogenes in Food Safety","field":"Biochemistry, Genetics and Molecular Biology","cited_by":34,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"Canadian Food Inspection Agency","funders":"Centers for Disease Control and Prevention; National Institutes of Health","keywords":"Outbreak; Multinational corporation; Subtyping; Genome; Data sharing; Dissemination; Whole genome sequencing; Data science; Food safety; Food microbiology; Biotechnology; Business; Biology; Genetics; Computer science; Medicine; Virology; Food science; Gene; Telecommunications","retraction":null,"screen_n_in":null,"score":{"opus":0.1269353286331178,"gpt":0.3259008197003254,"spread":0.1989654910672075,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003128755,0.00008422208,0.0001549765,0.000003991498,0.00005915448,0.00002174079,0.0009057795,0.00003712682,0.00001008224],"category_scores_gemma":[0.00009927296,0.00005112813,0.0000301039,0.00003457311,0.0001955375,0.000003876497,0.003373868,0.00005788654,0.000002417115],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000004257604,"about_ca_system_score_gemma":0.00002091701,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001423469,"about_ca_topic_score_gemma":0.00002607943,"domain_scores_codex":[0.999359,0.00004358518,0.000172206,0.0002591588,0.0000506685,0.0001154112],"domain_scores_gemma":[0.9990664,0.00002252791,0.0001113796,0.0007526084,0.00002479951,0.00002228406],"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":[0.0006451741,0.000009664564,0.04075389,0.00004774316,0.0001100901,7.783874e-7,0.0004546336,0.00002131778,0.9535177,0.0005187598,0.003274676,0.0006455377],"study_design_scores_gemma":[0.01016031,0.0001958425,0.08778147,0.00005360331,0.0002216761,0.000130689,0.0003816907,0.004964993,0.2749411,0.0005457292,0.6199602,0.0006627757],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.988557,0.002169571,0.0002828591,0.007909392,0.00005801684,0.0001841777,0.0002774358,0.000007276052,0.0005542935],"genre_scores_gemma":[0.9899385,0.0003104405,0.00102029,0.00808116,0.0004270114,0.000007099907,0.0001417034,0.00001315472,0.00006066144],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6785767,"threshold_uncertainty_score":0.4205285,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["open_science"],"domain":null,"study_design":"not_applicable","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":[],"domain":null,"study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"},{"id":"W4290546285","doi":"10.29173/cais1262","title":"The oligopoly’s shift to open access publishing","year":2022,"lang":"en","type":"article","venue":"Proceedings of the Annual Conference of CAIS / Actes du congrès annuel de l ACSI","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":34,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":true,"about_ca":true},"ca_institutions":"Université de Montréal; Wilfrid Laurier University; University of Ottawa; Bureau de Coopération Interuniversitaire; Université du Québec à Montréal; Dalhousie University","funders":"Social Sciences and Humanities Research Council of Canada; Natural Sciences and Engineering Research Council of Canada; Canadian Institutes of Health Research","keywords":"Oligopoly; Agency (philosophy); Library science; Publishing; Political science; Economics; Business; Computer science; Sociology; Social science; Law","retraction":null,"screen_n_in":null,"score":{"opus":0.4044126975648308,"gpt":0.5025465979353135,"spread":0.0981339003704827,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","bibliometrics","sts","scholarly_communication","open_science"],"consensus_categories":["scholarly_communication","open_science"],"category_scores_codex":[0.02738457,0.0002433622,0.0005475616,0.008188741,0.00142317,0.1867262,0.04907995,0.00009682442,0.0003283845],"category_scores_gemma":[0.3965835,0.0001414199,0.0002420269,0.06696718,0.000481455,0.04411172,0.03986121,0.0007295772,0.00001245516],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001805284,"about_ca_system_score_gemma":0.0009618522,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001410943,"about_ca_topic_score_gemma":0.00006928737,"domain_scores_codex":[0.988986,0.0001544662,0.001090229,0.0007755145,0.008091196,0.0009025547],"domain_scores_gemma":[0.6683719,0.0094945,0.003952639,0.002182894,0.314607,0.001391084],"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":[0.0003898382,0.0002992875,0.5158728,0.0000372063,0.00009056764,0.000002879405,0.01958681,0.00005169081,0.002593682,0.06483036,0.2929225,0.1033224],"study_design_scores_gemma":[0.0004497628,0.0003484255,0.2789426,0.00003063696,0.00001620032,0.00001375565,0.007527357,0.0004464632,0.005942959,0.04936454,0.6566427,0.0002746738],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9176459,0.0002332337,0.00004451999,0.03689743,0.0006082269,0.0008180629,0.0003311945,0.00002061132,0.04340085],"genre_scores_gemma":[0.9952937,0.0000671587,0.0001798134,0.0008619122,0.00008153009,0.0001177908,0.000001711225,0.00001938849,0.003377031],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.369199,"threshold_uncertainty_score":0.9998769,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["scholarly_communication","open_science"],"domain":null,"study_design":"not_applicable","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["scholarly_communication","open_science"],"domain":null,"study_design":"design_other","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"},{"id":"W2037890806","doi":"10.1087/0953151053585028","title":"In a paperless world a new role for academic libraries: providing open access","year":2005,"lang":"en","type":"article","venue":"Learned Publishing","topic":"Research Data Management Practices","field":"Computer Science","cited_by":34,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université du Québec","funders":"Euroscience; University of Southampton","keywords":"Mandate; Publishing; Scholarly communication; The Internet; World Wide Web; Open science; Open access publishing; Political science; Library science; Computer science; Electronic publishing; Public relations","retraction":null,"screen_n_in":null,"score":{"opus":0.2662665729937543,"gpt":0.452188456835734,"spread":0.1859218838419797,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication","open_science"],"consensus_categories":["scholarly_communication","open_science"],"category_scores_codex":[0.003685189,0.0001744673,0.0002304959,0.0007104008,0.000219666,0.1611478,0.02349629,0.0001015028,0.00006901741],"category_scores_gemma":[0.004915321,0.0001724498,0.00004398488,0.002290935,0.00002531411,0.6376082,0.01456761,0.00102665,0.00002429915],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001477488,"about_ca_system_score_gemma":0.0005718308,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001478538,"about_ca_topic_score_gemma":0.00140517,"domain_scores_codex":[0.9970636,0.0002246708,0.0004142456,0.0009315592,0.0005717836,0.0007941103],"domain_scores_gemma":[0.9973896,0.0009936877,0.0002644746,0.001036374,0.00006075118,0.0002551564],"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":[0.00005947921,0.00005232381,0.001529231,0.00004553274,0.00001837216,0.00001254129,0.001254524,0.0007366043,0.0001355536,0.6559423,0.09982763,0.2403859],"study_design_scores_gemma":[0.0008941538,0.00002043395,0.000516828,0.00008286473,0.00000359645,0.000002331276,0.0002023089,0.03389058,0.0003853092,0.02612564,0.9376398,0.0002361071],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"commentary","genre_gemma":"methods","genre_scores_codex":[0.002218334,0.0008533121,0.3267047,0.5732855,0.0003756198,0.002966242,0.000008034044,0.0004010704,0.09318723],"genre_scores_gemma":[0.2282679,0.0006063685,0.5533985,0.04203348,0.00294981,0.001445963,0.0001188194,0.0001625343,0.1710167],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.8378122,"threshold_uncertainty_score":0.9934024,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["scholarly_communication","open_science"],"domain":null,"study_design":"not_applicable","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"high"},{"model":"gpt","categories":["scholarly_communication","open_science"],"domain":null,"study_design":"not_applicable","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"agree"},{"id":"W3022519489","doi":"10.1016/j.buildenv.2020.106848","title":"Current practices and infrastructure for open data based research on occupant-centric design and operation of buildings","year":2020,"lang":"en","type":"article","venue":"Building and Environment","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":33,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"Australian Research Council; Engineering and Physical Sciences Research Council; Natural Sciences and Engineering Research Council of Canada; Energistyrelsen; National Science Foundation","keywords":"Transparency (behavior); Open data; Metadata; Data sharing; Work (physics); Status quo; Computer science; Open science; Open research; Built environment; Big data; Best practice; Open standard; Data science; Architectural engineering; Engineering; Interoperability; Computer security; World Wide Web; Civil engineering","retraction":null,"screen_n_in":null,"score":{"opus":0.3181109531981248,"gpt":0.4080395077723598,"spread":0.08992855457423504,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["open_science"],"consensus_categories":["open_science"],"category_scores_codex":[0.00147314,0.0001127567,0.0001502237,0.00008331952,0.0002128516,0.0003190974,0.007169334,0.00005789456,0.000003480924],"category_scores_gemma":[0.005139116,0.00009754887,0.000006382635,0.0001276613,0.0001271683,0.0007033417,0.04848348,0.0002022891,4.804874e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003001957,"about_ca_system_score_gemma":0.00003526725,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001665748,"about_ca_topic_score_gemma":1.798503e-7,"domain_scores_codex":[0.9984534,0.0001200864,0.000169808,0.0007911927,0.0002602823,0.0002052308],"domain_scores_gemma":[0.9972489,0.0005855038,0.0001540614,0.001914136,0.00001167043,0.00008576185],"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":[0.0003134203,0.0001895652,0.005922878,0.0003777498,0.00005199413,0.000006277064,0.0003611878,0.003573642,0.01573078,0.004547304,0.1186193,0.8503059],"study_design_scores_gemma":[0.0008545076,0.000479317,0.00291571,0.00008808541,0.00001513011,0.00000354807,0.0000483777,0.947863,0.009518444,0.008537015,0.0295047,0.0001721699],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.08218334,0.001535283,0.8920724,0.02323316,0.00004393457,0.0008292376,0.00005321586,0.00004406966,0.000005344013],"genre_scores_gemma":[0.3656636,0.001532631,0.6326327,0.0001072574,0.00001859178,0.00002474642,0.00001318139,0.000006648547,6.43461e-7],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.9442893,"threshold_uncertainty_score":0.9982024,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["open_science"],"domain":null,"study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","open_science"],"domain":"reproducibility","study_design":"design_other","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"},{"id":"W3134086897","doi":"10.2196/26718","title":"Clinical Trial Data Sharing for COVID-19–Related Research","year":2021,"lang":"en","type":"article","venue":"Journal of Medical Internet Research","topic":"COVID-19 Clinical Research Studies","field":"Medicine","cited_by":32,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of British Columbia","funders":"","keywords":"Sample size determination; Context (archaeology); Statistical power; Clinical trial; Pandemic; Sample (material); Coronavirus disease 2019 (COVID-19); Data sharing; Medicine; Psychology; Data science; Computer science; Statistics; Alternative medicine; Mathematics; Geography","retraction":null,"screen_n_in":null,"score":{"opus":0.775337627386123,"gpt":0.7283670039823996,"spread":0.04697062340372338,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","open_science","research_integrity","insufficient_payload"],"consensus_categories":["metaresearch","open_science"],"category_scores_codex":[0.2252758,0.0002303681,0.001489731,0.001023524,0.000316813,0.0002810732,0.005448588,0.0008918537,0.006066888],"category_scores_gemma":[0.8762583,0.0001663373,0.0007219982,0.001764129,0.002609807,0.0002559781,0.009511479,0.01145094,0.0001835223],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009248404,"about_ca_system_score_gemma":0.0206758,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002723534,"about_ca_topic_score_gemma":0.0001151285,"domain_scores_codex":[0.9738699,0.004414458,0.003800919,0.001519419,0.01467882,0.001716513],"domain_scores_gemma":[0.4969913,0.4458672,0.001494456,0.00971392,0.0265805,0.01935255],"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":[0.09364605,0.005210745,0.009594926,0.002011374,0.002307345,0.01863688,0.0006284082,0.000004477863,0.0001745475,0.004013119,0.7762008,0.08757134],"study_design_scores_gemma":[0.09761091,0.01026133,0.0009374879,0.001731089,0.00013522,0.0004238735,0.0009345768,0.0169749,0.0001171277,0.00919586,0.8614764,0.0002012873],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"commentary","genre_gemma":"empirical","genre_scores_codex":[0.2927056,0.006008049,0.009956212,0.6797446,0.004432353,0.003863029,0.00008891509,0.0000856081,0.003115666],"genre_scores_gemma":[0.9223173,0.02098077,0.002896108,0.007164151,0.01322081,0.0001403662,0.0001700826,0.0001507559,0.03295968],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.6725804,"threshold_uncertainty_score":0.9999324,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","open_science"],"domain":"reproducibility","study_design":"theoretical_or_conceptual","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","scholarly_communication","open_science"],"domain":"reproducibility","study_design":"design_other","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"split"},{"id":"W2164878690","doi":"10.1525/jer.2007.2.1.3","title":"Sharing Data and Results in Ethnographic Research: Why This Should Not Be an Ethical Imperative","year":2007,"lang":"en","type":"article","venue":"Journal of Empirical Research on Human Research Ethics","topic":"Qualitative Research Methods and Ethics","field":"Social Sciences","cited_by":30,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"McMaster University","funders":"McMaster University; U.S. Department of Homeland Security","keywords":"Ethnography; Variety (cybernetics); Sociology; Engineering ethics; Research ethics; Relation (database); Politics; Field (mathematics); Epistemology; Public relations; Political science; Law; Computer science; Anthropology","retraction":null,"screen_n_in":null,"score":{"opus":0.9682090312510818,"gpt":0.7973267805208706,"spread":0.1708822507302112,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","metaepi_narrow","sts","scholarly_communication","open_science","research_integrity"],"consensus_categories":["metaresearch","sts","research_integrity"],"category_scores_codex":[0.827274,0.0004003043,0.0009824377,0.008240171,0.008288706,0.001919067,0.006793903,0.003495933,0.000259595],"category_scores_gemma":[0.5299207,0.0003307509,0.0002039295,0.009413186,0.02661361,0.001644669,0.003894482,0.1013171,0.00003113241],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001651551,"about_ca_system_score_gemma":0.007617528,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.01035588,"about_ca_topic_score_gemma":0.05291056,"domain_scores_codex":[0.7042409,0.2211059,0.004668306,0.003341062,0.05808696,0.008556908],"domain_scores_gemma":[0.6036628,0.3578127,0.0005530997,0.003359358,0.02920601,0.005406014],"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":[0.00652858,0.001689018,0.002429714,0.0003091945,0.0001441757,0.002065957,0.3730872,0.000009342431,0.004176948,0.5415599,0.05648246,0.01151746],"study_design_scores_gemma":[0.003050948,0.009522624,0.006893024,0.001591805,0.00001052002,0.00001940544,0.2224759,0.000528169,0.0009159798,0.3387168,0.4156153,0.0006595504],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"commentary","genre_gemma":"empirical","genre_scores_codex":[0.3429327,0.0008853955,0.0003344662,0.6232512,0.0002059125,0.001539769,0.0001215234,0.00004810974,0.03068096],"genre_scores_gemma":[0.9762027,0.01066283,0.004836647,0.004359131,0.002086841,0.00003340275,0.00003171558,0.0001158047,0.001670941],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.63327,"threshold_uncertainty_score":0.9999145,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","open_science"],"domain":"reproducibility","study_design":"theoretical_or_conceptual","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","open_science","scholarly_communication"],"domain":"reproducibility","study_design":"qualitative","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"split"},{"id":"W2909101430","doi":"10.1111/ddi.12885","title":"Open access solutions for biodiversity journals: Do not replace one problem with another","year":2019,"lang":"en","type":"article","venue":"Diversity and Distributions","topic":"Species Distribution and Climate Change","field":"Environmental Science","cited_by":29,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"Memorial University of Newfoundland; McGill University","funders":"","keywords":"Publishing; Biodiversity; Diversity (politics); Process (computing); Library science; Sociology; Political science; Ecology; Computer science; Biology; Law","retraction":null,"screen_n_in":null,"score":{"opus":0.1362599909243479,"gpt":0.2994477482036357,"spread":0.1631877572792878,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["sts","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0003293398,0.0001115752,0.0001461959,0.00002448401,0.00302407,0.0004013802,0.0006520763,0.00006415349,0.0335033],"category_scores_gemma":[0.00002687249,0.0001038603,0.00005266529,0.0002677729,0.0002130852,0.001164092,0.005386208,0.0001050547,0.0005596789],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003420546,"about_ca_system_score_gemma":0.00001691993,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001054644,"about_ca_topic_score_gemma":0.0004900156,"domain_scores_codex":[0.9989983,0.00003312877,0.00009795577,0.0003333375,0.0002054508,0.000331805],"domain_scores_gemma":[0.999472,0.0000486177,0.00009269261,0.0002086472,0.00004731852,0.0001306779],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0004279339,0.0005398185,0.8540031,0.00002948797,0.00008342064,0.000002914764,0.0005538145,0.00005212369,0.0006059766,0.01735153,0.1259653,0.0003845773],"study_design_scores_gemma":[0.002604326,0.0001896879,0.8751332,0.00003012677,0.0001357258,0.00001040419,0.002783305,0.00004871818,0.0004299634,0.0006738371,0.1175572,0.000403538],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9641122,0.00003294687,0.005249992,0.004592359,0.00006541063,0.0009896748,0.006485419,0.00005244756,0.01841958],"genre_scores_gemma":[0.998037,0.0000681653,0.0001503687,0.0004469167,0.000007946765,0.00001470569,0.0003502671,0.000003133094,0.0009215247],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.0339248,"threshold_uncertainty_score":0.9982738,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["scholarly_communication","open_science"],"domain":null,"study_design":"not_applicable","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"},{"model":"gpt","categories":["scholarly_communication"],"domain":null,"study_design":"design_other","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"},{"id":"W2749249904","doi":"10.1108/ajim-01-2017-0024","title":"Incorporating data sharing to the reward system of science","year":2017,"lang":"en","type":"article","venue":"Aslib Journal of Information Management","topic":"Research Data Management Practices","field":"Computer Science","cited_by":27,"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":"National Institutes of Health","keywords":"Computer science; Similarity (geometry); Set (abstract data type); Data science; Originality; Data sharing; Information retrieval; Value (mathematics); Data set; Scientometrics; Scale (ratio); World Wide Web; Psychology; Artificial intelligence","retraction":null,"screen_n_in":null,"score":{"opus":0.1397662065720466,"gpt":0.3809386428785568,"spread":0.2411724363065102,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication","open_science"],"consensus_categories":["scholarly_communication","open_science"],"category_scores_codex":[0.01402995,0.00007537395,0.0001225384,0.0005919315,0.0008094053,0.006534941,0.02401268,0.00001020956,0.000001309437],"category_scores_gemma":[0.001154968,0.00005080791,0.00002636757,0.0005275125,0.0001327391,0.1133729,0.01315638,0.0001355242,0.00003359528],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00009117433,"about_ca_system_score_gemma":0.00008405392,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002929335,"about_ca_topic_score_gemma":0.000003542743,"domain_scores_codex":[0.9973874,0.00003942865,0.0007269098,0.0001592292,0.001468859,0.0002181509],"domain_scores_gemma":[0.9936606,0.00004832353,0.002124318,0.003628525,0.0004347557,0.0001034919],"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":[0.00001934148,0.00001842077,0.0008884902,0.0002927498,0.0000857898,0.00002030082,0.0006875512,0.002480569,0.00003312268,0.858923,0.004789569,0.1317611],"study_design_scores_gemma":[0.002177511,0.0005838192,0.1099003,0.002078064,0.0001408424,0.0001524845,0.01484004,0.4292265,0.001516591,0.00303295,0.4356942,0.0006567282],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.002788888,0.00001332526,0.9571408,0.006557424,0.0005656938,0.0003435592,0.00000435412,0.00001696284,0.03256899],"genre_scores_gemma":[0.8547153,0.00005942255,0.1448351,0.0002167697,0.00005686442,0.000003869903,0.000002608324,0.000002802459,0.0001072955],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.85589,"threshold_uncertainty_score":0.994825,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","bibliometrics"],"domain":"incentives","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"high"},{"model":"gpt","categories":["metaresearch","bibliometrics","open_science"],"domain":"incentives","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W4387874083","doi":"10.1186/s12916-023-03113-0","title":"A 10-year update to the principles for clinical trial data sharing by pharmaceutical companies: perspectives based on a decade of literature and policies","year":2023,"lang":"en","type":"article","venue":"BMC Medicine","topic":"Ethics in Clinical Research","field":"Medicine","cited_by":25,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto; St. Michael's Hospital","funders":"Cancer Council South Australia; National Health and Medical Research Council; Health Canada","keywords":"Data sharing; Medicine; Transparency (behavior); Scope (computer science); Multidisciplinary approach; Clinical trial; Health care; Data management; Process (computing); Pharmaceutical industry; Knowledge management; Public relations; Alternative medicine; Computer science; Political science","retraction":null,"screen_n_in":null,"score":{"opus":0.8122559443649875,"gpt":0.6759175419285655,"spread":0.136338402436422,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.01522919,0.0001736027,0.0006393567,0.000188612,0.0001099797,0.00003325504,0.0009245094,0.0002473411,0.0002218078],"category_scores_gemma":[0.1288617,0.00009836863,0.00009509899,0.0006112448,0.0008247851,0.00003968826,0.0007499264,0.001209731,0.00003531529],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00003127406,"about_ca_system_score_gemma":0.0003082679,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001442144,"about_ca_topic_score_gemma":0.00003588255,"domain_scores_codex":[0.9966326,0.0002335472,0.0008639898,0.0008647543,0.001030207,0.0003748676],"domain_scores_gemma":[0.9670718,0.03018829,0.0001364747,0.001731119,0.0004366022,0.0004357153],"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":[0.3426383,0.002082869,0.01786198,0.00440174,0.0006765768,0.00004657453,0.01143343,0.00006978598,0.001315419,0.06778744,0.546333,0.005352831],"study_design_scores_gemma":[0.2060097,0.02462385,0.03103985,0.009112635,0.0009221425,0.00001683409,0.01041269,0.1645164,0.0004064655,0.004225171,0.5480989,0.0006153298],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.6348772,0.004634578,0.001939504,0.3453059,0.0007597984,0.00735451,0.001214022,0.000266418,0.003648055],"genre_scores_gemma":[0.9561217,0.00681694,0.005957204,0.01118912,0.006387849,0.000289943,0.001016329,0.000130253,0.01209063],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3341168,"threshold_uncertainty_score":0.8784763,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","open_science"],"domain":"reproducibility","study_design":"observational","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"high"},{"model":"gpt","categories":["metaresearch","scholarly_communication","open_science"],"domain":"reproducibility","study_design":"theoretical_or_conceptual","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"split"},{"id":"W2808252916","doi":"10.1038/s41436-018-0017-5","title":"Development of a consent resource for genomic data sharing in the clinical setting","year":2018,"lang":"en","type":"article","venue":"Genetics in Medicine","topic":"Genomics and Rare Diseases","field":"Biochemistry, Genetics and Molecular Biology","cited_by":25,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"McGill University","funders":"National Institute of Child Health and Human Development; National Cancer Institute; National Human Genome Research Institute","keywords":"Data sharing; Informed consent; Resource (disambiguation); Key (lock); Genomic information; Computer science; Medicine; Data science; Internet privacy; Genome; Alternative medicine; Biology; Pathology; Computer security","retraction":null,"screen_n_in":null,"score":{"opus":0.1593591139675099,"gpt":0.419254181272979,"spread":0.2598950673054691,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001877612,0.00008141044,0.0001479457,0.00003890456,0.00003987033,0.000004362551,0.000699739,0.00006551327,0.000004894948],"category_scores_gemma":[0.0004096954,0.00005893284,0.00002064281,0.0000564743,0.0001848273,5.630089e-7,0.0004174914,0.00005879466,8.692442e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000007608592,"about_ca_system_score_gemma":0.0001636528,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000007992849,"about_ca_topic_score_gemma":0.0004521615,"domain_scores_codex":[0.9988379,0.00004562868,0.0005150551,0.0003442572,0.00008835131,0.0001687372],"domain_scores_gemma":[0.9990799,0.00007415494,0.0001013168,0.0006705741,0.00003924688,0.00003474061],"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":[0.000953178,0.0006736176,0.4076506,0.0004277699,0.0002171165,0.00003154722,0.009075265,0.0001731803,0.4109814,0.0002588898,0.03271056,0.1368468],"study_design_scores_gemma":[0.005489359,0.001036266,0.2688025,0.0003608891,0.0000867966,0.00003229547,0.006150122,0.003382004,0.01560635,0.0005809218,0.6980129,0.0004595692],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9965562,0.001570099,0.0008557866,0.0003070685,0.0001074145,0.000267999,0.00001893183,0.00000130858,0.0003152453],"genre_scores_gemma":[0.9781986,0.0002178886,0.01995105,0.0007430906,0.0005727378,0.00001360892,0.0002656811,0.00001227417,0.00002502504],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6653024,"threshold_uncertainty_score":0.2403212,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["open_science"],"domain":null,"study_design":"not_applicable","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":[],"domain":null,"study_design":"design_other","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"split"},{"id":"W2588666165","doi":"10.1186/s13063-017-1816-8","title":"Challenges to complete and useful data sharing","year":2017,"lang":"en","type":"letter","venue":"Trials","topic":"Ethics in Clinical Research","field":"Medicine","cited_by":24,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Hamilton Health Sciences; Population Health Research Institute; McMaster University; St. Joseph’s Healthcare Hamilton; Impact","funders":"","keywords":"Data sharing; Data quality; The Internet; Quality (philosophy); Clinical trial; Data collection; Computer science; Medicine; Data access; Collaboratory; Internet privacy; Data management; Modalities; Data science; World Wide Web; Data mining; Alternative medicine; Business; Database","retraction":null,"screen_n_in":null,"score":{"opus":0.9778865347180541,"gpt":0.7107495091616779,"spread":0.2671370255563762,"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":[0.03719867,0.0002483256,0.001829617,0.0001655058,0.0001293429,0.0001958314,0.002436305,0.001614268,0.0003026323],"category_scores_gemma":[0.2312334,0.000192192,0.0001144162,0.00003073573,0.0002303133,0.00006955503,0.003501636,0.006009587,0.0004671973],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004044146,"about_ca_system_score_gemma":0.0002851666,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00009406997,"about_ca_topic_score_gemma":0.0001352091,"domain_scores_codex":[0.995281,0.0004260022,0.001005691,0.001528128,0.001289642,0.0004695253],"domain_scores_gemma":[0.9500219,0.04115234,0.0004574264,0.007795495,0.0002889586,0.000283908],"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":[0.0001750847,0.00001541809,0.00008978684,0.001461503,0.0002531316,0.0006915292,0.00007319478,1.601316e-8,0.00005875848,0.0003618355,0.9885554,0.00826437],"study_design_scores_gemma":[0.001241513,0.0001977864,0.0008390298,0.001769632,0.0002130501,0.00002584468,0.000007173167,0.000006070807,0.00001352564,0.01079733,0.9847026,0.0001864547],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"commentary","genre_gemma":"commentary","genre_scores_codex":[0.0001289211,0.002842561,0.00001585204,0.9746013,0.0006674359,0.001430862,0.0004302191,0.00004747871,0.01983541],"genre_scores_gemma":[0.0004885313,0.02301244,0.004578599,0.8905301,0.04270871,0.0001046385,0.001170715,0.000187729,0.03721856],"genre_candidate":"commentary","genre_consensus":"commentary","teacher_disagreement_score":0.1940347,"threshold_uncertainty_score":0.9996818,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch"],"domain":"reproducibility","study_design":"not_applicable","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"high"},{"model":"gpt","categories":["metaresearch","open_science"],"domain":"reproducibility","study_design":"not_applicable","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W2998766662","doi":"10.7554/elife.52157","title":"Open exploration","year":2020,"lang":"en","type":"article","venue":"eLife","topic":"Meta-analysis and systematic reviews","field":"Decision Sciences","cited_by":23,"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 Mental Health; National Institute on Drug Abuse; Social Sciences and Humanities Research Council of Canada; Knut och Alice Wallenbergs Stiftelse; National Science Foundation","keywords":"Open science; Open data; Focus (optics); Open research; Data science; Computer science; Engineering ethics; Political science; Sociology; World Wide Web; Engineering; Mathematics","retraction":null,"screen_n_in":null,"score":{"opus":0.9687226010785637,"gpt":0.6272201453176606,"spread":0.3415024557609031,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","scholarly_communication","insufficient_payload"],"consensus_categories":["metaresearch","insufficient_payload"],"category_scores_codex":[0.04053147,0.00008366362,0.001150396,0.00004022896,0.00006409329,0.001529794,0.002301443,0.00002131733,0.02246228],"category_scores_gemma":[0.0337923,0.00003766659,0.0003912585,0.0007430561,0.000008923921,0.000546633,0.0002963805,0.00004527106,0.05047313],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000004388263,"about_ca_system_score_gemma":0.00002852717,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000003535324,"about_ca_topic_score_gemma":0.000005391781,"domain_scores_codex":[0.9901658,0.002634283,0.003397012,0.0004760389,0.003247397,0.00007951816],"domain_scores_gemma":[0.996251,0.0005375215,0.001269647,0.001349209,0.0004497801,0.0001428809],"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":[0.000002166382,0.000007425195,0.002818918,0.000003404961,0.00002443155,0.000001580299,0.0008826936,0.00006058237,0.00009648615,0.003601351,0.9772657,0.01523532],"study_design_scores_gemma":[0.00007008615,0.00001739947,0.001160724,0.000002675769,0.00002041612,4.139338e-7,0.0005485929,0.0070622,0.0001542499,0.001668738,0.9892365,0.00005796051],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"other","genre_gemma":"empirical","genre_scores_codex":[0.07095672,0.000985523,0.3737921,0.1184228,0.0007277241,0.002489307,0.00001858023,0.00002640623,0.4325808],"genre_scores_gemma":[0.9649937,0.00001274757,0.008165293,0.01122045,0.0002527095,0.00003512147,0.000006001291,0.000006799457,0.01530716],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.894037,"threshold_uncertainty_score":0.9995067,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["open_science"],"domain":null,"study_design":"not_applicable","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["insufficient_payload"],"domain":null,"study_design":"not_applicable","genre":"other","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W2791957709","doi":"10.3389/fnins.2018.00047","title":"Open Science Meets Stem Cells: A New Drug Discovery Approach for Neurodegenerative Disorders","year":2018,"lang":"en","type":"review","venue":"Frontiers in Neuroscience","topic":"Pluripotent Stem Cells Research","field":"Biochemistry, Genetics and Molecular Biology","cited_by":23,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":true},"ca_institutions":"Montreal Neurological Institute and Hospital; McGill University","funders":"Québec Consortium for Drug Discovery; Fondation Brain Canada","keywords":"Drug discovery; Induced pluripotent stem cell; Open science; Neuroscience; Computer science; Translational research; Stem cell; Computational biology; Data science; Medicine; Bioinformatics; Biology; Embryonic stem cell; Pathology","retraction":null,"screen_n_in":null,"score":{"opus":0.04700102043857651,"gpt":0.3336071977617853,"spread":0.2866061773232088,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","open_science"],"consensus_categories":[],"category_scores_codex":[0.001504954,0.0005744984,0.0009871178,0.0004547248,0.0004411243,0.0008658267,0.008528531,0.0002028833,7.473514e-7],"category_scores_gemma":[0.0002237229,0.00049101,0.000301457,0.001550434,0.001609414,0.0001024662,0.003849952,0.0003309679,0.000003134981],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001394028,"about_ca_system_score_gemma":0.002893661,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003179743,"about_ca_topic_score_gemma":0.00001895132,"domain_scores_codex":[0.9942583,0.0003298388,0.0005769291,0.003053165,0.000688224,0.001093574],"domain_scores_gemma":[0.9978906,0.00003361687,0.0003425628,0.001314239,0.00008726322,0.0003317512],"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":[0.0001950622,0.0006643414,0.00007214537,0.006382613,0.00004723154,0.00001571821,0.000193962,0.0004475932,0.01679526,0.0001542855,0.2661154,0.7089164],"study_design_scores_gemma":[0.0003527295,0.0002514716,0.000002677084,0.0003809267,0.00003844736,0.000008635275,0.00003559857,0.001200274,0.00356067,0.0000251385,0.9935765,0.0005669181],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.00007245887,0.9280457,0.06222457,0.00006058462,0.003390792,0.0049302,0.0001904959,0.00001614678,0.001069066],"genre_scores_gemma":[0.0000888569,0.9730482,0.01127193,0.000141029,0.000240431,0.0004581346,0.00005807342,0.00008688479,0.01460647],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.7274612,"threshold_uncertainty_score":0.9997541,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["open_science"],"domain":null,"study_design":"not_applicable","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","open_science"],"domain":"reproducibility","study_design":"design_other","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"},{"id":"W2887866346","doi":"10.1136/bmjebm-2018-111040","title":"More than 2 billion pairs of eyeballs: Why aren’t you sharing medical knowledge on Wikipedia?","year":2018,"lang":"en","type":"article","venue":"BMJ evidence-based medicine","topic":"Wikis in Education and Collaboration","field":"Social Sciences","cited_by":22,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Queen's University","funders":"","keywords":"Reading (process); Knowledge translation; Internet privacy; World Wide Web; Public relations; Knowledge sharing; Encyclopedia; Resource (disambiguation); Information Dissemination; Computer science; Medical education; Political science; Library science; Knowledge management; Medicine; Law","retraction":null,"screen_n_in":null,"score":{"opus":0.1161487099793547,"gpt":0.4504509774698217,"spread":0.3343022674904669,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.004630015,0.0001830207,0.0003254802,0.0003110839,0.0003543466,0.00002860558,0.0005918871,0.0002250205,0.003360987],"category_scores_gemma":[0.01405614,0.0001423936,0.00006670058,0.001382626,0.001578699,0.0003006567,0.00004351611,0.0002077471,0.000108119],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002458562,"about_ca_system_score_gemma":0.001977165,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001644031,"about_ca_topic_score_gemma":0.003296679,"domain_scores_codex":[0.9965275,0.0003873924,0.0005920181,0.0004264929,0.001709188,0.0003574172],"domain_scores_gemma":[0.9968806,0.001008123,0.0003086268,0.0004823304,0.0009118203,0.0004084958],"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":[0.0009796361,0.001342618,0.117814,0.0005398632,0.00009314148,0.00003157141,0.1220384,0.0000374425,0.001006995,0.06455932,0.5705924,0.1209645],"study_design_scores_gemma":[0.004839323,0.005114344,0.05896708,0.02820513,0.0003156303,0.000005791457,0.09055906,0.004160958,0.006702699,0.005208592,0.7945506,0.001370803],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.4803763,0.003709541,0.003065245,0.4317496,0.009758262,0.002232896,0.000009810873,0.0004067124,0.06869159],"genre_scores_gemma":[0.9864697,0.0004470129,0.000197874,0.004389865,0.0069116,0.00008565325,0.00001477357,0.00001808262,0.001465396],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5060934,"threshold_uncertainty_score":0.9975501,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["scholarly_communication","open_science"],"domain":null,"study_design":"not_applicable","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"},{"model":"gpt","categories":["scholarly_communication"],"domain":null,"study_design":"not_applicable","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"split"},{"id":"W2602562802","doi":"10.1145/3051088","title":"Guilt-free data reuse","year":2017,"lang":"en","type":"article","venue":"Communications of the ACM","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","cited_by":20,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"Stanford Bio-X; Simons Institute for the Theory of Computing, University of California Berkeley; University of Toronto; University of Pennsylvania; Microsoft Research; Alfred P. Sloan Foundation; National Science Foundation","keywords":"Computer science; Reuse; Inference; Process (computing); Set (abstract data type); Data mining; Data set; Statistical inference; Simple (philosophy); Machine learning; Data validation; Artificial intelligence; Database","retraction":null,"screen_n_in":null,"score":{"opus":0.2370552844700501,"gpt":0.3837375646579778,"spread":0.1466822801879276,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","open_science"],"consensus_categories":["open_science"],"category_scores_codex":[0.0006796418,0.00006940887,0.000099474,0.00003696942,0.0008073205,0.0001850312,0.9961025,0.00005213898,0.000002396468],"category_scores_gemma":[0.9834828,0.00005294835,0.00003196406,0.000133877,0.0006775819,0.001026635,0.9995713,0.0002432043,0.00002261999],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001905802,"about_ca_system_score_gemma":0.00005686426,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000143835,"about_ca_topic_score_gemma":0.00008736785,"domain_scores_codex":[0.9991809,0.00009074721,0.0001875184,0.0002390647,0.0001733556,0.0001283889],"domain_scores_gemma":[0.08847716,0.0003268369,0.0003046054,0.9107986,0.0000746956,0.0000181068],"domain_codex":null,"domain_gemma":"reproducibility","domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[3.653538e-7,0.00002836596,0.001021606,0.000002553645,0.00001009636,9.118173e-8,0.00001958458,1.976385e-7,0.0001862254,0.008927578,0.9889661,0.0008372196],"study_design_scores_gemma":[0.00009670857,0.000004048994,0.006489985,0.00003060067,0.000005756123,0.000002323872,0.000005467532,0.008557901,0.0009605169,0.9609535,0.0228349,0.00005823289],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"commentary","genre_gemma":"methods","genre_scores_codex":[0.001129896,0.0004838362,0.000408948,0.9929394,0.0001720427,0.0001077202,0.00008449482,0.0001836757,0.004490059],"genre_scores_gemma":[0.02285854,0.0003878367,0.9765785,0.0001152743,0.000008220468,0.000007776966,0.00001173138,0.000005363346,0.00002677414],"genre_candidate":"commentary","genre_consensus":null,"teacher_disagreement_score":0.9928241,"threshold_uncertainty_score":0.6209338,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","open_science"],"domain":"reproducibility","study_design":"theoretical_or_conceptual","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["open_science"],"domain":null,"study_design":"not_applicable","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"},{"id":"W3105389930","doi":"10.12688/wellcomeopenres.16378.2","title":"Fostering global data sharing: highlighting the recommendations of the Research Data Alliance COVID-19 working group","year":2021,"lang":"en","type":"preprint","venue":"Wellcome Open Research","topic":"COVID-19 diagnosis using AI","field":"Medicine","cited_by":20,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"McGill University; Laurentian University; Environment and Climate Change Canada","funders":"Biotechnology and Biological Sciences Research Council; Université de Bordeaux; Horizon 2020; Rural Development Administration; European Commission; Agence Universitaire de la Francophonie; Wellcome Trust; Fundação de Amparo à Pesquisa do Estado de São Paulo; National Institute on Drug Abuse; University of Michigan; National Science Foundation","keywords":"Data sharing; Preparedness; Interoperability; Medical research; Metadata; Knowledge management; Political science; Public relations; Data science; Medicine; Computer science; World Wide Web; Alternative medicine","retraction":null,"screen_n_in":null,"score":{"opus":0.8031307428435279,"gpt":0.5907681340689679,"spread":0.21236260877456,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","metaepi_narrow","sts","scholarly_communication","open_science","research_integrity"],"consensus_categories":["metaresearch","open_science"],"category_scores_codex":[0.05606813,0.0004760702,0.001019414,0.0003197131,0.002428669,0.003514814,0.04529882,0.0004777017,0.0005989505],"category_scores_gemma":[0.02646784,0.0003387249,0.000169859,0.003939997,0.001222823,0.0005540189,0.4215621,0.006608725,0.00007207488],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002434999,"about_ca_system_score_gemma":0.006741542,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0266183,"about_ca_topic_score_gemma":0.007248407,"domain_scores_codex":[0.9834565,0.00457417,0.001439354,0.004462591,0.004351419,0.001715951],"domain_scores_gemma":[0.9589548,0.009404747,0.0005578115,0.02916355,0.001241879,0.0006772152],"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":[0.0006635708,0.00157668,0.04560303,0.005394991,0.001195401,0.0004378397,0.002639828,0.000637004,0.001058903,0.005844106,0.8791017,0.05584696],"study_design_scores_gemma":[0.001460232,0.0000916928,0.005981056,0.01121818,0.0001615901,0.0000607463,0.002975133,0.01239823,0.0002623741,0.003102661,0.9618106,0.0004774678],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"commentary","genre_gemma":"empirical","genre_scores_codex":[0.008069743,0.005100039,0.0006247596,0.9562982,0.001900812,0.01023908,0.001593914,0.000125369,0.01604804],"genre_scores_gemma":[0.9345994,0.01125722,0.01645208,0.01690986,0.002595039,0.00155451,0.01060256,0.0004040142,0.005625333],"genre_candidate":"commentary","genre_consensus":null,"teacher_disagreement_score":0.9393884,"threshold_uncertainty_score":0.9999065,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["open_science"],"domain":null,"study_design":"not_applicable","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","open_science","scholarly_communication"],"domain":"reproducibility","study_design":"not_applicable","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"},{"id":"W4317525596","doi":"10.2196/40814","title":"Understanding Public Attitudes and Willingness to Share Commercial Data for Health Research: Survey Study in the United Kingdom","year":2023,"lang":"en","type":"article","venue":"JMIR Public Health and Surveillance","topic":"Ethics in Clinical Research","field":"Medicine","cited_by":20,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":true,"about_ca":false},"ca_institutions":"","funders":"Peter Sowerby Foundation; University College London; Cancer Research UK","keywords":"Public health; Survey research; Environmental health; Survey data collection; Internet privacy; Psychology; Business; Medicine; Computer science; Applied psychology; Nursing; Statistics","retraction":null,"screen_n_in":null,"score":{"opus":0.9634509391386481,"gpt":0.6910705090113438,"spread":0.2723804301273043,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.1421728,0.0001832752,0.0006453003,0.0008284406,0.001106075,0.0004678839,0.0008949802,0.0001821585,0.000008293045],"category_scores_gemma":[0.07253789,0.0001364293,0.00002382214,0.004572057,0.0003456565,0.0001835682,0.000943987,0.001941838,0.000005614914],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003685855,"about_ca_system_score_gemma":0.00368076,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00399753,"about_ca_topic_score_gemma":0.1248838,"domain_scores_codex":[0.9873908,0.006222823,0.001055903,0.00120986,0.002071565,0.002049042],"domain_scores_gemma":[0.9521856,0.0436802,0.0001430846,0.001543415,0.0008626249,0.00158509],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0002403522,0.0003532937,0.9642355,0.001076694,0.00002164239,0.000009158723,0.002746226,4.577878e-7,1.500129e-7,0.004188377,0.02454243,0.002585667],"study_design_scores_gemma":[0.001763555,0.001588188,0.9543583,0.0001211218,2.465335e-7,0.000004013081,0.005256737,0.0006949496,5.425757e-9,0.0005306576,0.03556689,0.0001152845],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.622133,0.0001921449,0.0003230141,0.3721273,0.0001159035,0.004459307,0.0005149349,0.000079249,0.00005513167],"genre_scores_gemma":[0.9866263,0.0008713644,0.00007395519,0.01020972,0.0001760591,0.0003270021,0.001612431,0.00003671166,0.00006641053],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3644933,"threshold_uncertainty_score":0.9352745,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["open_science"],"domain":null,"study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":[],"domain":null,"study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"},{"id":"W2530827738","doi":"10.11613/bm.2016.035","title":"IMPACT Observatory: tracking the evolution of clinical trial data sharing and research integrity","year":2016,"lang":"en","type":"review","venue":"Biochemia Medica","topic":"Research Data Management Practices","field":"Computer Science","cited_by":19,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Université de Montréal; Computer Research Institute of Montréal","funders":"FP7 People: Marie-Curie Actions; European Commission","keywords":"Data sharing; Clinical trial; Business; Process (computing); Psychological intervention; Open data; Data integrity; Medicine; Public relations; Political science; Knowledge management; Computer science; Computer security; Alternative medicine; Nursing; World Wide Web","retraction":null,"screen_n_in":null,"score":{"opus":0.7990351710026269,"gpt":0.6297849765274307,"spread":0.1692501944751962,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","scholarly_communication","open_science","research_integrity"],"consensus_categories":["metaresearch","open_science"],"category_scores_codex":[0.05916949,0.0002734746,0.001123965,0.0002746884,0.0001952934,0.001073978,0.0188699,0.0004513591,0.00002050403],"category_scores_gemma":[0.0434311,0.0001362671,0.0002376609,0.001022636,0.0007553555,0.006347005,0.01766453,0.002496898,0.00002419328],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000284569,"about_ca_system_score_gemma":0.002886231,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002860258,"about_ca_topic_score_gemma":0.000008387251,"domain_scores_codex":[0.9921601,0.001895949,0.001406646,0.001736491,0.002186871,0.0006139798],"domain_scores_gemma":[0.9834166,0.006403439,0.001098513,0.008442394,0.0002919054,0.0003471053],"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":[0.0001275258,0.0000804025,0.0001128933,0.002162084,0.0002137582,0.00001006849,0.00001222242,9.380214e-10,8.782407e-7,0.004327388,0.01173406,0.9812187],"study_design_scores_gemma":[0.001884066,0.0001843366,0.00009128244,0.006596797,0.00009508043,0.000007590844,0.00001855957,0.000224455,0.000001104961,0.000218648,0.9905162,0.0001618216],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[0.000117711,0.9928448,0.003678543,0.000798435,0.0005844795,0.00132088,0.00009195646,0.00004184188,0.0005213539],"genre_scores_gemma":[0.0002334204,0.9974874,0.0009358799,0.000006719479,0.001021196,0.00005602141,0.0000883333,0.00001890355,0.0001521384],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.9810569,"threshold_uncertainty_score":0.999963,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","research_integrity"],"domain":"reproducibility","study_design":"observational","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","open_science","research_integrity"],"domain":"reproducibility","study_design":"observational","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"},{"id":"W4226504205","doi":"10.31222/osf.io/9e526","title":"Teaching Open and Reproducible Scholarship: A Critical Review of the Evidence Base for Current Pedagogical Methods and their Outcomes","year":2022,"lang":"en","type":"review","venue":"","topic":"Meta-analysis and systematic reviews","field":"Decision Sciences","cited_by":18,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Economic and Social Research Council; Queen's University; Department for the Economy; UK Research and Innovation; Queen's University Belfast; John Templeton Foundation","keywords":"Scholarship; Credibility; Promotion (chess); Psychology; Engineering ethics; Medical education; Mathematics education; Political science; Medicine; Engineering","retraction":null,"screen_n_in":null,"score":{"opus":0.9813179241858248,"gpt":0.7638947768303257,"spread":0.2174231473554991,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","metaepi_broad","scholarly_communication","open_science","insufficient_payload"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.5297139,0.0006820084,0.01590762,0.0002334929,0.0005323451,0.001365307,0.006132409,0.0001121121,0.00747299],"category_scores_gemma":[0.6484308,0.0002053494,0.0049496,0.001061328,0.0001777959,0.0005117864,0.004416536,0.0009775612,0.0000212258],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004184778,"about_ca_system_score_gemma":0.0003702641,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001213356,"about_ca_topic_score_gemma":0.000003618962,"domain_scores_codex":[0.868408,0.1079905,0.01597005,0.003664132,0.003579727,0.0003876392],"domain_scores_gemma":[0.7561515,0.2108378,0.01240093,0.0187497,0.001364728,0.000495353],"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":[8.460356e-7,0.000035572,0.00008591464,0.0852066,0.0001243185,2.461782e-7,0.00006619717,9.402108e-9,7.101838e-8,0.006084142,0.004997429,0.9033986],"study_design_scores_gemma":[0.00003807172,0.00003359653,0.00001706851,0.0562207,0.002370225,0.00004350206,0.0001411389,0.00002887638,1.718373e-7,0.004883263,0.9359967,0.0002267353],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"review","genre_gemma":"review","genre_scores_codex":[9.506371e-7,0.9815127,0.008255969,0.002697117,0.000346657,0.00629487,0.00005977427,0.000003738566,0.0008282163],"genre_scores_gemma":[0.000001256021,0.9547147,0.03934084,0.0006405404,0.00004651584,0.0009445004,0.000004375103,0.00002231453,0.004285029],"genre_candidate":"review","genre_consensus":"review","teacher_disagreement_score":0.9309992,"threshold_uncertainty_score":0.9996714,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","open_science"],"domain":"reproducibility","study_design":"not_applicable","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"high"},{"model":"gpt","categories":["metaresearch","open_science"],"domain":"reproducibility","study_design":"systematic_review","genre":"review","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W2261976862","doi":"10.3138/jsp.47.2.147","title":"Publishing Undergraduate Research: Linking Teaching and Research through a Dedicated Peer-Reviewed Open Access Journal","year":2016,"lang":"en","type":"article","venue":"Journal of Scholarly Publishing","topic":"Higher Education Practises and Engagement","field":"Social Sciences","cited_by":17,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":true,"about_ca":false},"ca_institutions":"","funders":"","keywords":"Publishing; Work (physics); Discipline; Process (computing); Peer review; Quality (philosophy); Public relations; Engineering ethics; Sociology; Library science; Political science; Computer science; Engineering; Social science","retraction":null,"screen_n_in":null,"score":{"opus":0.5479534976218382,"gpt":0.5658860954472149,"spread":0.01793259782537671,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","sts","scholarly_communication","open_science","research_integrity"],"consensus_categories":["metaresearch","scholarly_communication"],"category_scores_codex":[0.2771217,0.0001956978,0.0004587441,0.001456371,0.008916861,0.4328944,0.008074312,0.0003524013,0.0003066757],"category_scores_gemma":[0.2235549,0.0001348633,0.000138655,0.002298869,0.0005049698,0.4742536,0.003203299,0.01418024,0.00001868407],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001460103,"about_ca_system_score_gemma":0.00381464,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004814826,"about_ca_topic_score_gemma":0.0004675571,"domain_scores_codex":[0.9725503,0.01585077,0.001464759,0.00049274,0.008272412,0.001368964],"domain_scores_gemma":[0.972412,0.007121974,0.001582034,0.0004714189,0.01718252,0.001229983],"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":[0.00008234582,0.0002161258,0.01288775,0.00004069271,0.0001775089,0.0001032745,0.03156834,0.000002239045,0.0005180805,0.1163933,0.6146766,0.2233338],"study_design_scores_gemma":[0.0009821255,0.0001052866,0.003199314,0.001495617,0.0000257647,0.00009297644,0.01960317,0.000001425377,0.00001849398,0.1295366,0.8447704,0.0001688115],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"commentary","genre_gemma":"empirical","genre_scores_codex":[0.07285976,0.002949694,0.001686898,0.8509142,0.002065924,0.0004296086,0.000003783082,0.00003111213,0.06905896],"genre_scores_gemma":[0.9677097,0.007696472,0.009860041,0.001707144,0.004591519,0.00001496096,0.000002537837,0.00005215143,0.008365502],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8948499,"threshold_uncertainty_score":0.9972925,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["scholarly_communication","open_science"],"domain":null,"study_design":"not_applicable","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["scholarly_communication"],"domain":null,"study_design":"not_applicable","genre":"commentary","about_ca_system":false,"about_ca_topic":false,"confidence":"low"}],"label_agreement":"split"},{"id":"W1953894052","doi":"10.6017/ital.v34i3.5900","title":"Self-Archiving with Ease in an Institutional Repository: Microinteractions and the User Experience","year":2015,"lang":"en","type":"article","venue":"Information Technology and Libraries","topic":"Digital and Traditional Archives Management","field":"Arts and Humanities","cited_by":17,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"MacEwan University; University of Alberta","funders":"","keywords":"Computer science; Usability; Process (computing); World Wide Web; Institutional repository; Human–computer interaction","retraction":null,"screen_n_in":null,"score":{"opus":0.01198402682143977,"gpt":0.1807309643641611,"spread":0.1687469375427214,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0000301129,0.00005027392,0.00004965189,0.0001693913,0.0003019013,0.0004004563,0.00007146392,0.000009119928,0.00000386916],"category_scores_gemma":[0.000009885008,0.00003067226,0.000005268895,0.00004129266,0.001014859,0.004828052,0.00006040942,0.00007190699,0.000003624291],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000004232718,"about_ca_system_score_gemma":0.00002513509,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00001838816,"about_ca_topic_score_gemma":0.00002105841,"domain_scores_codex":[0.9997255,0.000009458621,0.0001059635,0.00005056012,0.00004985277,0.00005867902],"domain_scores_gemma":[0.9998379,0.00002594369,0.00003828716,0.00005762829,0.00001711566,0.00002315815],"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":[0.00006719941,0.00001801294,0.00166537,0.000008397526,0.000007035733,0.000002052353,0.02721014,0.000007234043,1.583375e-7,0.9692053,0.00005953104,0.001749529],"study_design_scores_gemma":[0.002077127,0.0002345431,0.01554565,0.00006939766,0.00001478405,0.0001820781,0.06216539,0.00225203,0.00007128052,0.3511207,0.5660381,0.0002288718],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8561742,0.00003910972,0.0003546918,0.002762897,0.00007141916,0.0001470734,0.000009597617,0.0001362589,0.1403048],"genre_scores_gemma":[0.9988528,0.00000466583,0.0005690994,0.0002629407,0.00001881829,0.00004071988,0.00001669549,0.000001325,0.0002329271],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6180846,"threshold_uncertainty_score":0.3861609,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["scholarly_communication","open_science"],"domain":null,"study_design":"qualitative","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["scholarly_communication","open_science"],"domain":null,"study_design":"design_other","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"split"},{"id":"W2792505333","doi":"10.47622/978-1-928331-56-8","title":"The Social Dynamics of Open Data","year":2017,"lang":"en","type":"book","venue":"African Minds eBooks","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":17,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Comisión Nacional de Investigación Científica y Tecnológica; International Development Research Centre","keywords":"Open data; Data collection; Data science; Dynamics (music); Computer science; Open science; Open research; Open source; Management science; World Wide Web; Sociology; Social science; Engineering; Statistics; Mathematics","retraction":null,"screen_n_in":null,"score":{"opus":0.412274993420936,"gpt":0.4866937168620067,"spread":0.0744187234410707,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["scholarly_communication","open_science"],"consensus_categories":["open_science"],"category_scores_codex":[0.009052572,0.0002455779,0.0006420264,0.0001338508,0.001117386,0.002380326,0.03588445,0.0001852191,0.0001404625],"category_scores_gemma":[0.001443629,0.0001565464,0.0001367997,0.00004037228,0.001263899,0.0002988028,0.02304062,0.0003127539,0.0002247374],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001016116,"about_ca_system_score_gemma":0.0008959612,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000266618,"about_ca_topic_score_gemma":0.00553647,"domain_scores_codex":[0.9958735,0.0002525976,0.0009320966,0.0008521433,0.001780787,0.0003089377],"domain_scores_gemma":[0.9898058,0.001388874,0.001799512,0.00670635,0.0002205496,0.00007894388],"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":[0.00001649984,0.000006854717,7.389218e-7,0.00000450185,0.00005533573,0.000004433602,0.0003061898,2.046688e-8,1.171371e-7,0.1913991,0.4944133,0.313793],"study_design_scores_gemma":[0.0001326681,0.0000257017,0.00002570809,0.0000209553,0.00006052331,6.845275e-7,0.00108841,0.00006383823,0.000001018534,0.2021709,0.7962524,0.000157195],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"other","genre_gemma":"other","genre_scores_codex":[0.000004276157,0.00003692042,0.0003377559,0.001521364,0.0004878107,0.000531372,0.004480064,0.000009103432,0.9925913],"genre_scores_gemma":[0.001753614,0.000007497913,0.0001922743,0.0001394414,0.0002393305,0.00001700341,0.0006538372,0.00002468004,0.9969723],"genre_candidate":"other","genre_consensus":"other","teacher_disagreement_score":0.3136358,"threshold_uncertainty_score":0.9986553,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["open_science","sts"],"domain":null,"study_design":"theoretical_or_conceptual","genre":"other","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["scholarly_communication","open_science"],"domain":null,"study_design":"theoretical_or_conceptual","genre":"other","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W2286638398","doi":"10.15353/joci.v12i2.3221","title":"User Centred Methods for Measuring the Value of Open Data","year":2016,"lang":"en","type":"article","venue":"The Journal of Community Informatics","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":17,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":true,"about_ca":false},"ca_institutions":"","funders":"International Development Research Centre","keywords":"Open data; Computer science; Key (lock); Measure (data warehouse); Data science; Data quality; Quality (philosophy); Open source; Knowledge management; World Wide Web; Data mining; Engineering; Software; Operations management; Computer security; Metric (unit)","retraction":null,"screen_n_in":null,"score":{"opus":0.6407874768984645,"gpt":0.5418677551865244,"spread":0.09891972171194008,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","open_science"],"consensus_categories":["metaresearch"],"category_scores_codex":[0.08041298,0.00007744776,0.0002942163,0.00009059844,0.0004548857,0.0003018768,0.01893046,0.00002436675,0.0000457173],"category_scores_gemma":[0.01328273,0.00002684917,0.00007368773,0.0002362785,0.0001962451,0.002413441,0.006909979,0.0002658816,0.00001166823],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002192432,"about_ca_system_score_gemma":0.00009757075,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00008404275,"about_ca_topic_score_gemma":0.000041856,"domain_scores_codex":[0.9937062,0.00366729,0.001626527,0.00001480904,0.0008349512,0.0001502786],"domain_scores_gemma":[0.9759389,0.01841446,0.001872452,0.003159744,0.0005641405,0.00005032167],"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":[0.0003294932,0.0001592083,0.00006160804,0.00005395592,0.000283169,5.169614e-8,0.09007178,0.0003417117,0.0001714592,0.02590206,0.1515738,0.7310517],"study_design_scores_gemma":[0.001104538,0.0001495487,0.0008530351,0.000134304,0.0001469115,0.00001778577,0.2177025,0.00150707,0.001266248,0.0972589,0.6797711,0.00008803212],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03478066,0.00008916568,0.9494325,0.01084326,0.0003320871,0.0004604736,0.0001619624,0.000004206329,0.003895685],"genre_scores_gemma":[0.8113946,0.0005919851,0.1824139,0.004331606,0.0001058761,0.000003874912,0.00001459729,0.00001722995,0.001126256],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.776614,"threshold_uncertainty_score":0.9950288,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["open_science"],"domain":null,"study_design":"qualitative","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["open_science"],"domain":null,"study_design":"design_other","genre":"methods","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"split"},{"id":"W3159243089","doi":"10.1136/bmjopen-2020-041530","title":"Living labs for patient engagement and knowledge exchange: an exploratory sequential mixed methods study to develop a living lab in paediatric rehabilitation","year":2021,"lang":"en","type":"article","venue":"BMJ Open","topic":"Innovative Approaches in Technology and Social Development","field":"Business, Management and Accounting","cited_by":16,"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":"Canadian Institutes of Health Research","keywords":"Thematic analysis; Knowledge translation; Medicine; Medical education; Data collection; Rehabilitation; Qualitative research; Health care; Qualitative property; Independent living; Nursing; Knowledge management; Gerontology; Computer science","retraction":null,"screen_n_in":null,"score":{"opus":0.1186407241714337,"gpt":0.4095494602372792,"spread":0.2909087360658456,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01793626,0.0001837681,0.000306968,0.0003854499,0.0003641622,0.000291815,0.000311584,0.00008728736,0.00004786946],"category_scores_gemma":[0.004972566,0.0001967246,0.0000217093,0.001891693,0.00003366174,0.0008708191,0.005468539,0.0001691312,0.00001344944],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000147251,"about_ca_system_score_gemma":0.0002247919,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00005462355,"about_ca_topic_score_gemma":0.001599031,"domain_scores_codex":[0.9975095,0.001023985,0.0005000196,0.0005255538,0.0001507117,0.0002902416],"domain_scores_gemma":[0.998332,0.0005459367,0.0001863437,0.0002612141,0.000658642,0.00001590225],"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":[0.0001481939,0.002809662,0.1254985,0.001064198,0.0001578888,0.00002741157,0.3170717,0.0000175304,0.0003948976,0.01904067,0.004719959,0.5290494],"study_design_scores_gemma":[0.001540883,0.0003883496,0.3668244,0.0005200229,0.0000873991,0.000001211865,0.5928224,0.0007350242,0.000335075,0.004211293,0.03143521,0.00109873],"study_design_candidate":"qualitative","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9831989,0.00007773844,0.006479346,0.0004539608,0.0006393446,0.00762783,0.000002155105,0.00007277294,0.001447985],"genre_scores_gemma":[0.8954296,0.000002317172,0.09704514,0.0006304883,0.0002487318,0.006461245,0.00001867311,0.00003289711,0.0001309591],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5279506,"threshold_uncertainty_score":0.8022196,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["open_science"],"domain":null,"study_design":"qualitative","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"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":"W2766932290","doi":"10.3138/jsp.49.1.5","title":"Traditional versus Open Access Scholarly Journal Publishing: An Economic Perspective","year":2017,"lang":"en","type":"article","venue":"Journal of Scholarly Publishing","topic":"scientometrics and bibliometrics research","field":"Decision Sciences","cited_by":16,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":true,"about_ca":false},"ca_institutions":"","funders":"","keywords":"Publishing; Misinformation; Electronic publishing; Scholarly communication; Perspective (graphical); Open access publishing; Public relations; Business; Computer science; Political science; World Wide Web; The Internet; Law","retraction":null,"screen_n_in":null,"score":{"opus":0.8658095751057954,"gpt":0.6303082985146546,"spread":0.2355012765911407,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","metaepi_narrow","bibliometrics","sts","scholarly_communication","open_science","research_integrity","insufficient_payload"],"consensus_categories":["metaresearch","scholarly_communication","open_science"],"category_scores_codex":[0.1650671,0.0003484361,0.0008807163,0.05186339,0.004778166,0.9670092,0.07678702,0.0005337502,0.00231214],"category_scores_gemma":[0.5205153,0.0002600059,0.0005895458,0.01954708,0.0003843054,0.8922455,0.008726421,0.008395823,0.00009304696],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002586645,"about_ca_system_score_gemma":0.005773737,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0009909322,"about_ca_topic_score_gemma":0.0001749232,"domain_scores_codex":[0.9758969,0.00114735,0.002503644,0.001025041,0.0183292,0.00109785],"domain_scores_gemma":[0.9140291,0.006315053,0.009006974,0.002788261,0.06374548,0.004115154],"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":[0.001472929,0.0007944292,0.2327997,0.000004500322,0.000462946,0.000618578,0.00112033,0.0002509641,0.0003966603,0.08729751,0.5100968,0.1646847],"study_design_scores_gemma":[0.007696371,0.001296905,0.5992395,0.00007369957,0.00005103126,0.001062872,0.007062193,0.0006764107,0.0001283424,0.2742805,0.1078768,0.0005553115],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7244895,0.001455922,0.001646372,0.1065625,0.01517182,0.0003437015,0.0001567511,0.00002022382,0.1501533],"genre_scores_gemma":[0.9883361,0.0002879534,0.004801818,0.0007166428,0.004811955,0.000003722095,0.000005366537,0.00004197557,0.0009943914],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.40222,"threshold_uncertainty_score":0.9999852,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["scholarly_communication","open_science"],"domain":null,"study_design":"theoretical_or_conceptual","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"},{"model":"gpt","categories":["scholarly_communication"],"domain":null,"study_design":"theoretical_or_conceptual","genre":"other","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W3122450121","doi":"10.22024/unikent/03/fal.7","title":"What's Feminist About Open Access? A Relational Approach to Copyright in the Academy","year":2011,"lang":"en","type":"article","venue":"University of Kent","topic":"Copyright and Intellectual Property","field":"Business, Management and Accounting","cited_by":16,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"York University","funders":"","keywords":"Intellectual property; Feminist legal theory; Sociology; Scholarship; Context (archaeology); Criticism; Political science; Feminism; Law and economics; Law","retraction":null,"screen_n_in":null,"score":{"opus":0.1695429822840149,"gpt":0.2723435056772202,"spread":0.1028005233932053,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0002975396,0.00008069559,0.0001149713,0.0001520486,0.0001688141,0.0003603746,0.001629962,0.00004693333,0.003473778],"category_scores_gemma":[0.00002380204,0.00005855872,0.00004213943,0.0003871331,0.00007946511,0.003382934,0.000960627,0.0001272885,0.0002481779],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002928997,"about_ca_system_score_gemma":0.00002157617,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002309191,"about_ca_topic_score_gemma":0.000135008,"domain_scores_codex":[0.9993941,0.00001415828,0.0001072765,0.0001869387,0.0001692994,0.0001282543],"domain_scores_gemma":[0.9997027,0.00002274014,0.00008125122,0.0001307526,0.00005157935,0.00001099351],"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":[0.0007471303,0.0008895969,0.01388496,0.0001014842,0.00005611359,0.00001274011,0.01514625,0.00002379782,0.00003684262,0.158995,0.8059394,0.004166662],"study_design_scores_gemma":[0.0005310529,0.00001607555,0.07620198,0.00007547444,0.00002922764,9.498381e-7,0.005433953,0.003104317,0.00002993948,0.002432402,0.9119728,0.0001718212],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"other","genre_gemma":"empirical","genre_scores_codex":[0.05275948,0.00003361569,0.0003616495,0.0015193,0.0001020777,0.0004598758,0.000001222123,0.00001526581,0.9447475],"genre_scores_gemma":[0.9940845,0.00001735306,0.0003061424,0.002335636,0.00007455659,8.900352e-7,0.000009446035,0.000004884298,0.003166589],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9415809,"threshold_uncertainty_score":0.9974372,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["sts","scholarly_communication","open_science"],"domain":null,"study_design":"theoretical_or_conceptual","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["scholarly_communication","open_science"],"domain":null,"study_design":"theoretical_or_conceptual","genre":"other","about_ca_system":false,"about_ca_topic":false,"confidence":"high"}],"label_agreement":"split"},{"id":"W2470439906","doi":"10.1038/ng.3499","title":"Analysis of five years of controlled access and data sharing compliance at the International Cancer Genome Consortium","year":2016,"lang":"en","type":"letter","venue":"Nature Genetics","topic":"Ethics in Clinical Research","field":"Medicine","cited_by":16,"is_retracted":false,"has_abstract":false,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Ontario Institute for Cancer Research; McGill University","funders":"","keywords":"Biology; Genome; Cancer; Compliance (psychology); Genetics; Computational biology; Gene","retraction":null,"screen_n_in":null,"score":{"opus":0.3813966335204105,"gpt":0.5693888942973265,"spread":0.187992260776916,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["research_integrity"],"consensus_categories":["research_integrity"],"category_scores_codex":[0.0015841,0.0001513574,0.0008826627,0.0002283562,0.00003047919,0.00003018592,0.00257524,0.001359714,0.0005599358],"category_scores_gemma":[0.004315619,0.0000967853,0.0001699297,0.0002767442,0.0006335538,0.00002391708,0.002943556,0.003864697,0.000002370738],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00009768952,"about_ca_system_score_gemma":0.0003293933,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00006214584,"about_ca_topic_score_gemma":0.0005037387,"domain_scores_codex":[0.9970449,0.00006077785,0.0006161744,0.0006784148,0.001383852,0.0002159049],"domain_scores_gemma":[0.990978,0.005556546,0.0005358191,0.001869856,0.001006451,0.00005332232],"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":[0.003005635,0.0001242907,0.3489719,0.002625975,0.04635945,0.0002940742,0.0003719905,0.0002743547,0.004444173,0.0001574357,0.5852929,0.008077801],"study_design_scores_gemma":[0.008956296,0.0001442145,0.3275422,0.001758032,0.01126074,0.000011133,0.00002213709,0.007070093,0.0006519015,0.001942427,0.640137,0.0005038331],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"commentary","genre_gemma":"commentary","genre_scores_codex":[0.1157832,0.03246119,0.00007862531,0.8367786,0.001195622,0.001363855,0.00961035,0.0000143827,0.002714154],"genre_scores_gemma":[0.3405667,0.07953026,0.001008491,0.5322894,0.006547614,0.0000818609,0.003468075,0.000135284,0.03637233],"genre_candidate":"commentary","genre_consensus":"commentary","teacher_disagreement_score":0.3044892,"threshold_uncertainty_score":0.9999368,"prediction_status":"machine_predicted_unvalidated"},"labels":[{"model":"gemma","categories":["metaresearch","open_science"],"domain":"reproducibility","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"low"},{"model":"gpt","categories":["metaresearch","open_science"],"domain":"reproducibility","study_design":"observational","genre":"empirical","about_ca_system":false,"about_ca_topic":false,"confidence":"medium"}],"label_agreement":"agree"}]}