{"meta":{"query_hash":"e87bc5552faa","filters":{"venue":"CUHK Research Data Repository"},"cohort_total":6,"direct_labels_cover":0,"predictions_cover":6,"exported":6,"export_cap":100000,"truncated":false,"label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12"},"permalink":"https://metacan.xera.ac/q/e87bc5552faa","api":"https://metacan.xera.ac/api/v1/cohort?venue=CUHK+Research+Data+Repository"},"results":[{"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto; Queen's University","funders":"","keywords":"Jellyfish; Genome; Gene; Homeobox; Genomics","score_opus":0.44829671979699626,"score_gpt":0.502938805463396,"score_spread":0.05464208566639972,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W6910853651","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.00109031,0.00014355479,0.00006170748,0.00004033194,0.000012765549,0.0000098644705,0.99778074,0.00022321352,0.00063739764],"genre_scores_gemma":[0.0006969533,0.00006835118,0.00015768949,0.000013103229,0.0000020873695,0.00002895248,0.99869734,0.000036795627,0.00029870917],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9990495,0.00007656223,0.00009745317,0.00032665502,0.00022918681,0.00022069496],"domain_scores_gemma":[0.99855953,0.0002724376,0.00014978662,0.00027962078,0.00047535083,0.00026330494],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00072778296,0.0018721279,0.0017865935,0.003965015,0.001770524,0.002002402,0.002274653,0.0018058514,0.0429362],"category_scores_gemma":[0.0033375637,0.0009202329,0.0015978874,0.007428872,0.00053202035,0.0011852776,0.0027783278,0.0020148617,0.0387946],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00067808514,0.000063518375,0.0049236733,0.003399485,0.00020324501,0.0001749874,0.00026099215,0.00086807355,0.004017692,0.0014706224,0.97581303,0.008126476],"study_design_scores_gemma":[0.00042404744,0.00004086217,0.0297562,0.00033596653,0.00012275855,0.00016869373,0.00025193774,0.00048791786,0.001317438,0.0010645571,0.9659656,0.0000640098],"about_ca_topic_score_codex":0.056131624,"about_ca_topic_score_gemma":0.0803755,"teacher_disagreement_score":0.056131624,"about_ca_system_score_codex":0.0018794431,"about_ca_system_score_gemma":0.0033926084,"threshold_uncertainty_score":0.14363593},"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto; Queen's University","funders":"","keywords":"Process (computing); Identification (biology); Product (mathematics)","score_opus":0.5530324821035331,"score_gpt":0.5411763995475676,"score_spread":0.011856082555965486,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W6910866301","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.00040319335,0.00033584016,0.00023373649,0.00018367212,0.00015618256,0.000041694566,0.9937621,0.0035356474,0.0013478625],"genre_scores_gemma":[0.00051599875,0.00008792317,0.0004962401,0.00006559918,0.000018022702,0.00008022026,0.9976846,0.0002103715,0.00084105565],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99791116,0.00035847264,0.0001587038,0.00072353263,0.00045473824,0.00039337674],"domain_scores_gemma":[0.9976761,0.0005168847,0.00013909629,0.0008355692,0.0005388465,0.00029340273],"candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0017577534,0.0073335594,0.0032654714,0.0049720258,0.0023254512,0.003792587,0.0074035497,0.0052097165,0.08626911],"category_scores_gemma":[0.0058461875,0.0015415759,0.003117033,0.006353791,0.0015217621,0.0023396472,0.003309359,0.0033312629,0.1724052],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00009394668,0.000035724886,0.00021096076,0.0004148244,0.000025650817,0.000018058157,0.000011694948,0.00023971211,0.0001594824,0.0001968761,0.99664414,0.0019489572],"study_design_scores_gemma":[0.00078829227,0.00009769025,0.002862092,0.0002928773,0.00008717227,0.00022365875,0.0001417762,0.002581764,0.0023090802,0.0018399227,0.98868215,0.000093477436],"about_ca_topic_score_codex":0.045858584,"about_ca_topic_score_gemma":0.065888725,"teacher_disagreement_score":0.91373086,"about_ca_system_score_codex":0.0032147877,"about_ca_system_score_gemma":0.0036757484,"threshold_uncertainty_score":0.28859895},"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto; Queen's University","funders":"","keywords":"Process (computing); Identification (biology); Product (mathematics)","score_opus":0.22899162059909323,"score_gpt":0.41913940156555085,"score_spread":0.19014778096645762,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W6929673858","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.00008621401,0.000085980006,0.00007496742,0.000050701412,0.000018219776,0.000007428233,0.99828976,0.0008967807,0.00048987503],"genre_scores_gemma":[0.00033342984,0.00006803665,0.00020140877,0.000040463947,0.0000056090744,0.00004161187,0.99875224,0.00014398484,0.000413181],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9987626,0.00019512419,0.000114360686,0.00045022342,0.00026660625,0.00021107376],"domain_scores_gemma":[0.9981407,0.0004984006,0.00017022052,0.0005563282,0.00037080218,0.0002635474],"candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0012983048,0.0037951425,0.0025653462,0.0040424503,0.0013382388,0.003681353,0.005368422,0.003582038,0.10671437],"category_scores_gemma":[0.00520899,0.0012172463,0.0020129536,0.0061881905,0.0008361664,0.0019718213,0.00272544,0.0021379245,0.17846985],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00011178959,0.000023385111,0.00053362024,0.0010247453,0.000049600934,0.00002422928,0.00001981172,0.00036147382,0.00019006441,0.00050019386,0.99549943,0.0016614852],"study_design_scores_gemma":[0.00040421443,0.000030424493,0.002408082,0.00022680429,0.0000573357,0.00007381348,0.000058551585,0.00077507686,0.0007761387,0.0015078333,0.99363387,0.000047911024],"about_ca_topic_score_codex":0.023643035,"about_ca_topic_score_gemma":0.031763848,"teacher_disagreement_score":0.89328563,"about_ca_system_score_codex":0.002190956,"about_ca_system_score_gemma":0.002752585,"threshold_uncertainty_score":0.3569951},"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto; Queen's University","funders":"","keywords":"Process (computing); Identification (biology); Product (mathematics)","score_opus":0.5571518294041781,"score_gpt":0.5380763801755616,"score_spread":0.01907544922861648,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W6966470483","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0003054265,0.00019857599,0.0001565952,0.0001568834,0.00010797503,0.000026870828,0.99588877,0.0020002087,0.0011587212],"genre_scores_gemma":[0.0005070237,0.00006731552,0.0003725784,0.000045824625,0.000015097242,0.000070566784,0.9979512,0.00016203318,0.00080837886],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9979512,0.00040120946,0.00015782494,0.00066575734,0.00041929368,0.0004046409],"domain_scores_gemma":[0.99745685,0.00050740247,0.000149237,0.00089743925,0.00065581495,0.0003333113],"candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0018499424,0.0050804415,0.0029755349,0.004397496,0.0022300188,0.0038670632,0.006334874,0.0039932397,0.079997785],"category_scores_gemma":[0.0060500763,0.0012546392,0.0023849993,0.0067668273,0.0013687115,0.0020573295,0.002942649,0.0025691716,0.16037764],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00007190575,0.000025545965,0.00021565317,0.00030932104,0.00002015048,0.0000130040125,0.000010628374,0.00020389807,0.000084453495,0.00022706485,0.9973687,0.0014496784],"study_design_scores_gemma":[0.00057044165,0.000055604916,0.0030020156,0.00023580545,0.00006119242,0.00012888484,0.00014140455,0.0020340965,0.0012589004,0.001738426,0.9906998,0.000073437],"about_ca_topic_score_codex":0.05927213,"about_ca_topic_score_gemma":0.082832836,"teacher_disagreement_score":0.9200022,"about_ca_system_score_codex":0.0033192714,"about_ca_system_score_gemma":0.00402222,"threshold_uncertainty_score":0.26761925},"labels":[],"label_agreement":null},{"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Dalhousie University","funders":"","keywords":"Process (computing); Identification (biology); Product (mathematics)","score_opus":0.39529508570393396,"score_gpt":0.5052243669251737,"score_spread":0.10992928122123974,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W6967189273","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.00006176462,0.00005074571,0.00004973521,0.000059286,0.00003758033,0.000012533246,0.99848855,0.00064735557,0.00059247203],"genre_scores_gemma":[0.00025133806,0.00004349764,0.00017542929,0.000049691975,0.000011561627,0.00007738387,0.9985524,0.0001755275,0.0006630369],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9983662,0.0002922851,0.00017683324,0.0005435209,0.0002783541,0.00034283294],"domain_scores_gemma":[0.9965192,0.00091263687,0.00023373497,0.000991465,0.00086908526,0.00047378006],"candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0013431119,0.0036719088,0.0026028175,0.0045200093,0.0014633476,0.003933512,0.0050587095,0.0036713558,0.24969104],"category_scores_gemma":[0.0065421714,0.0012715396,0.0017908223,0.008172868,0.00088970736,0.0024898732,0.0030329607,0.0021217845,0.30905715],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000043560376,0.000014551558,0.00014221732,0.00043076312,0.000014283352,0.0000072544008,0.000011409793,0.00009656636,0.00007564516,0.00020627222,0.9982022,0.0007551755],"study_design_scores_gemma":[0.00044220084,0.000040484487,0.001829865,0.00025345094,0.000035580884,0.00005283588,0.000088933004,0.00049922615,0.0006151589,0.0012089659,0.9948843,0.000049005394],"about_ca_topic_score_codex":0.02584269,"about_ca_topic_score_gemma":0.033407588,"teacher_disagreement_score":0.750309,"about_ca_system_score_codex":0.0024635813,"about_ca_system_score_gemma":0.003218105,"threshold_uncertainty_score":0.83529973},"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,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto; Queen's University","funders":"","keywords":"Process (computing); Identification (biology); Product (mathematics)","score_opus":0.5571518294041781,"score_gpt":0.5380763801755616,"score_spread":0.01907544922861648,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W6967200374","genre_codex":"dataset","genre_gemma":"dataset","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0002166726,0.0001344416,0.00011198647,0.000118877135,0.00006753561,0.000019429399,0.9970264,0.0013381205,0.0009665777],"genre_scores_gemma":[0.00046141492,0.000059245944,0.00030727408,0.000039660845,0.000012483339,0.00006163595,0.9981669,0.00013585933,0.00075549097],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99809223,0.00037754883,0.00015830704,0.0006143349,0.00038426134,0.00037322767],"domain_scores_gemma":[0.99739563,0.00048538457,0.0001722574,0.00089799665,0.000719495,0.00032927314],"candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0017629922,0.0041755866,0.0025593818,0.00405964,0.0020475753,0.003579389,0.0056533,0.0033698524,0.0880791],"category_scores_gemma":[0.006201304,0.001091543,0.00199083,0.006720627,0.0012289897,0.0019290514,0.0028314416,0.0021527554,0.17247756],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00007004994,0.000022165528,0.0002559292,0.0003225439,0.00001883645,0.000010922972,0.000011170385,0.00020145302,0.0000678566,0.00026389188,0.99741143,0.0013438124],"study_design_scores_gemma":[0.00046537508,0.000041336898,0.0030397507,0.00022343681,0.000051400908,0.00008980127,0.00012865254,0.0014718948,0.0009816529,0.0015525764,0.9918942,0.00005994481],"about_ca_topic_score_codex":0.058597736,"about_ca_topic_score_gemma":0.078137584,"teacher_disagreement_score":0.9119209,"about_ca_system_score_codex":0.0033492264,"about_ca_system_score_gemma":0.0037096716,"threshold_uncertainty_score":0.29465395},"labels":[],"label_agreement":null}]}