{"meta":{"page":1,"per_page":50,"max_per_page":100,"total":6,"total_is_capped":false,"direct_labels_cover":0,"predictions_cover":6,"direct_label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline (scores rank; they never assert a category)","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12","author_layer_release":"2026-06-26"},"query_hash":"e87bc5552faa","filters":{"venue":"CUHK Research Data Repository"}},"results":[{"id":"W6967189273","doi":"10.48668/ottxne/vi7kp6","title":"Amatrice_CAT5.v20210504.tab","year":2021,"lang":"en","type":"dataset","venue":"CUHK Research Data Repository","topic":"","field":"","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Dalhousie University","funders":"","keywords":"Process (computing); Identification (biology); Product (mathematics)","authors":[],"retraction":null,"screen_n_in":null,"score":{"opus":0.395295085703934,"gpt":0.5052243669251737,"spread":0.1099292812212397,"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","insufficient_payload"],"consensus_categories":["open_science","research_integrity","insufficient_payload"],"category_scores_codex":[0.01086295,0.001096555,0.00148425,0.002225376,0.001539899,0.002724164,0.01718935,0.001528137,0.001254409],"category_scores_gemma":[0.01094215,0.001134727,0.0003178302,0.00423015,0.001578138,0.001708361,0.02388694,0.008476214,0.05111434],"about_ca_system_candidate":true,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001821514,"about_ca_system_score_gemma":0.007199344,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004330307,"about_ca_topic_score_gemma":0.0005130908,"domain_scores_codex":[0.9772487,0.005708858,0.001581686,0.004842246,0.007678117,0.002940377],"domain_scores_gemma":[0.9632093,0.002419174,0.0006053437,0.0298239,0.002598868,0.001343384],"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":[0.0001163211,0.0006878663,0.00001636701,0.001241938,0.0004531164,0.01125453,0.00001409629,0.000001029,0.002212249,0.000006233455,0.9836875,0.0003088029],"study_design_scores_gemma":[0.0005699728,0.0001086064,0.00006302186,0.0006993356,0.0002583933,0.0007852927,0.0002504532,0.00003291127,0.0007896219,0.00001523309,0.9953619,0.001065252],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"dataset","genre_gemma":"dataset","genre_scores_codex":[0.00003768305,0.01743435,0.000003070713,0.0001495028,0.003190832,0.001204954,0.9749606,0.0002773831,0.002741619],"genre_scores_gemma":[0.000006484677,0.003229576,0.0002287191,0.00006616808,0.008597792,0.0002899076,0.9772646,0.0003814115,0.009935273],"genre_candidate":"dataset","genre_consensus":"dataset","teacher_disagreement_score":0.04985993,"threshold_uncertainty_score":0.9997681,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W6929673858","doi":"10.48668/vt8ump/3z9a5c","title":"edible.PacBio.fastq.MD5","year":2023,"lang":"en","type":"dataset","venue":"CUHK Research Data Repository","topic":"Invertebrate Immune Response Mechanisms","field":"Immunology and Microbiology","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto; Queen's University","funders":"","keywords":"Process (computing); Identification (biology); Product (mathematics)","authors":[],"retraction":null,"screen_n_in":null,"score":{"opus":0.2289916205990932,"gpt":0.4191394015655509,"spread":0.1901477809664576,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","sts","open_science","research_integrity","insufficient_payload"],"consensus_categories":["open_science","research_integrity"],"category_scores_codex":[0.00649393,0.0007163928,0.0009988273,0.00155243,0.001465402,0.0003509447,0.00864859,0.002219911,0.000816628],"category_scores_gemma":[0.002710207,0.0006735339,0.0001996355,0.001177376,0.001371477,0.0004724865,0.009814254,0.005395881,0.1567341],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003986798,"about_ca_system_score_gemma":0.00203932,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006145635,"about_ca_topic_score_gemma":0.0003413877,"domain_scores_codex":[0.9896173,0.0042803,0.001012449,0.002250677,0.0006552955,0.002183986],"domain_scores_gemma":[0.9871835,0.001798772,0.0003271141,0.009959999,0.0005771639,0.0001534382],"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":[0.0003064507,0.0002096024,0.000001418716,0.0002992795,0.0005809395,0.0009737647,0.00002105839,1.929535e-7,0.1033574,0.00003372544,0.8937601,0.0004560478],"study_design_scores_gemma":[0.0007292553,0.0003335623,0.00001779612,0.0003985222,0.000113019,0.000531398,0.0001995675,0.000004717977,0.04343963,0.00006377983,0.9535654,0.0006034098],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"dataset","genre_gemma":"dataset","genre_scores_codex":[0.00008049182,0.006615229,0.00001069519,0.0002576644,0.01084376,0.0009844942,0.9803526,0.0003287325,0.0005263197],"genre_scores_gemma":[0.0000130161,0.002385162,0.00002907228,0.00008312297,0.0007881748,0.0001989838,0.9385068,0.0001353458,0.05786037],"genre_candidate":"dataset","genre_consensus":"dataset","teacher_disagreement_score":0.1559175,"threshold_uncertainty_score":0.9998345,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W6910853651","doi":"10.48668/vt8ump","title":"Jellyfish genomes","year":2023,"lang":"en","type":"dataset","venue":"CUHK Research Data Repository","topic":"","field":"","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto; Queen's University","funders":"","keywords":"Jellyfish; Genome; Gene; Homeobox; Genomics","authors":[],"retraction":null,"screen_n_in":null,"score":{"opus":0.4482967197969963,"gpt":0.502938805463396,"spread":0.05464208566639972,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","sts","scholarly_communication","open_science","research_integrity","insufficient_payload"],"consensus_categories":["open_science"],"category_scores_codex":[0.01280601,0.0008184384,0.001025416,0.002498263,0.001376923,0.001700316,0.01838413,0.00110376,0.0003047395],"category_scores_gemma":[0.005437553,0.0008272286,0.000191252,0.002958526,0.001788269,0.00115485,0.0220223,0.005771884,0.3971432],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001162901,"about_ca_system_score_gemma":0.002950958,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00922037,"about_ca_topic_score_gemma":0.001320373,"domain_scores_codex":[0.9829199,0.003243685,0.001175376,0.003697925,0.006304822,0.002658334],"domain_scores_gemma":[0.9727183,0.002465193,0.000402196,0.02238667,0.001070566,0.0009570557],"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":[0.00009667847,0.0002524827,0.00002049201,0.0007349297,0.0004067247,0.006168117,0.00001433908,0.000001227961,0.0005386648,0.000003869533,0.9914493,0.0003131885],"study_design_scores_gemma":[0.000347126,0.0001203431,0.0001133594,0.0003605166,0.0001180683,0.0002135997,0.00009244419,0.00004571321,0.0001144926,0.00004999218,0.9976409,0.0007834471],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"dataset","genre_gemma":"dataset","genre_scores_codex":[0.00001683286,0.004027709,5.231429e-7,0.0001012197,0.002754932,0.001354974,0.989063,0.0006706625,0.002010102],"genre_scores_gemma":[0.000001486545,0.003251158,0.00006557807,0.00002586853,0.004990561,0.0002930459,0.9680316,0.0004862117,0.02285443],"genre_candidate":"dataset","genre_consensus":"dataset","teacher_disagreement_score":0.3968385,"threshold_uncertainty_score":0.9999232,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W6966470483","doi":"10.48668/vt8ump/1elgza","title":"sanderia.PacBio.fastq","year":2023,"lang":"en","type":"dataset","venue":"CUHK Research Data Repository","topic":"","field":"","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto; Queen's University","funders":"","keywords":"Process (computing); Identification (biology); Product (mathematics)","authors":[],"retraction":null,"screen_n_in":null,"score":{"opus":0.5571518294041781,"gpt":0.5380763801755616,"spread":0.01907544922861648,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow","sts","scholarly_communication","open_science","research_integrity","insufficient_payload"],"consensus_categories":["open_science","research_integrity"],"category_scores_codex":[0.01242385,0.0009674483,0.001177231,0.002472008,0.001538833,0.00184503,0.01634929,0.001352369,0.0003883322],"category_scores_gemma":[0.007687389,0.0009537622,0.0002167853,0.003166683,0.00187564,0.001314054,0.02059635,0.006198615,0.3370325],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001196679,"about_ca_system_score_gemma":0.003275366,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01403278,"about_ca_topic_score_gemma":0.001899522,"domain_scores_codex":[0.9813741,0.003639176,0.001307735,0.004018357,0.006728029,0.002932576],"domain_scores_gemma":[0.9723304,0.002238217,0.0004720989,0.02277229,0.001056408,0.001130582],"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":[0.0001977346,0.0003183366,0.000012792,0.0007797204,0.0004368307,0.004392161,0.00001565767,0.000001052703,0.001737761,0.000006092728,0.9919124,0.0001894369],"study_design_scores_gemma":[0.0005851689,0.0001689461,0.0001399919,0.0006268296,0.000135027,0.0002444356,0.0001631258,0.00005161184,0.0003075031,0.00004822144,0.99663,0.0008991868],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"dataset","genre_gemma":"dataset","genre_scores_codex":[0.00002686819,0.001639089,0.000001732669,0.0001329141,0.003751041,0.001491814,0.9902785,0.000756564,0.001921453],"genre_scores_gemma":[0.000005354149,0.001264144,0.00006706254,0.00003069781,0.00532817,0.00032183,0.9767295,0.0005280403,0.01572519],"genre_candidate":"dataset","genre_consensus":"dataset","teacher_disagreement_score":0.3366442,"threshold_uncertainty_score":0.9999441,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W6967200374","doi":"10.48668/vt8ump/c7v1mk","title":"sanderia.PacBio.fastq.gz.MD5","year":2023,"lang":"en","type":"dataset","venue":"CUHK Research Data Repository","topic":"","field":"","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto; Queen's University","funders":"","keywords":"Process (computing); Identification (biology); Product (mathematics)","authors":[],"retraction":null,"screen_n_in":null,"score":{"opus":0.5571518294041781,"gpt":0.5380763801755616,"spread":0.01907544922861648,"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","insufficient_payload"],"consensus_categories":["open_science","research_integrity"],"category_scores_codex":[0.01333915,0.001198545,0.001433298,0.002850609,0.001839472,0.002220361,0.01754075,0.001612483,0.0006923719],"category_scores_gemma":[0.008523528,0.001192211,0.0002838598,0.003537777,0.002224158,0.001602782,0.0221122,0.007018099,0.3859072],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001483756,"about_ca_system_score_gemma":0.004006418,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01505163,"about_ca_topic_score_gemma":0.002455011,"domain_scores_codex":[0.9785215,0.004187671,0.001592453,0.004648775,0.007561231,0.003488376],"domain_scores_gemma":[0.9704008,0.002486828,0.0005984927,0.02384725,0.001262038,0.001404661],"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":[0.0002752379,0.0004093969,0.00001733576,0.0009777956,0.0005732444,0.005109586,0.00002273079,0.000001569007,0.002532301,0.00000824399,0.9898258,0.0002467794],"study_design_scores_gemma":[0.0007725515,0.0002175023,0.0001864345,0.000790621,0.0001818398,0.000312206,0.0002184911,0.00005938102,0.0004495377,0.00005324641,0.9956361,0.00112207],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"dataset","genre_gemma":"dataset","genre_scores_codex":[0.00004275456,0.001475592,0.000002592746,0.0001346726,0.004379965,0.00186435,0.9888552,0.0008918435,0.002353076],"genre_scores_gemma":[0.000009441971,0.001514709,0.00008781964,0.00004131307,0.005894573,0.0004233153,0.9720966,0.0006534981,0.01927878],"genre_candidate":"dataset","genre_consensus":"dataset","teacher_disagreement_score":0.3852148,"threshold_uncertainty_score":0.9998281,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W6910866301","doi":"10.48668/vt8ump/rk7s2t","title":"edible.PacBio.fastq","year":2023,"lang":"en","type":"dataset","venue":"CUHK Research Data Repository","topic":"","field":"","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto; Queen's University","funders":"","keywords":"Process (computing); Identification (biology); Product (mathematics)","authors":[],"retraction":null,"screen_n_in":null,"score":{"opus":0.5530324821035331,"gpt":0.5411763995475676,"spread":0.01185608255596549,"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","insufficient_payload"],"consensus_categories":["open_science","research_integrity"],"category_scores_codex":[0.01443926,0.001018484,0.001200757,0.003307793,0.001591024,0.001842268,0.01930065,0.00145656,0.0003387847],"category_scores_gemma":[0.008960778,0.001011332,0.0002306054,0.004239721,0.002016156,0.001535507,0.02454127,0.007574533,0.4312952],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00132552,"about_ca_system_score_gemma":0.003795489,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01755553,"about_ca_topic_score_gemma":0.001868835,"domain_scores_codex":[0.979539,0.003466285,0.001387833,0.004284312,0.008037369,0.003285252],"domain_scores_gemma":[0.9694653,0.002474143,0.0004994445,0.02491824,0.001364406,0.001278484],"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":[0.0001620295,0.0003598642,0.00001455739,0.0009855331,0.0004233123,0.005222079,0.00001321113,0.000001299859,0.001748359,0.000007469186,0.9908844,0.0001778683],"study_design_scores_gemma":[0.0005089621,0.0001861697,0.00009876188,0.0007910457,0.0001454976,0.0002446351,0.0001406145,0.00004988875,0.0004579874,0.00004474064,0.9963972,0.0009345184],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"dataset","genre_gemma":"dataset","genre_scores_codex":[0.00001613438,0.001755034,0.000001735694,0.0001690633,0.004807603,0.00155209,0.9888017,0.0008355057,0.002061199],"genre_scores_gemma":[0.000001498153,0.001572314,0.00008826495,0.0000332458,0.008032849,0.0003614362,0.9686156,0.0005683959,0.02072638],"genre_candidate":"dataset","genre_consensus":"dataset","teacher_disagreement_score":0.4309565,"threshold_uncertainty_score":0.9998398,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null}]}