{"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":"codex-gemma-dda1882f352a","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.000016832859,0.004027709,5.2314294e-7,0.00010121969,0.0027549323,0.001354974,0.989063,0.0006706625,0.002010102],"genre_scores_gemma":[0.0000014865453,0.0032511582,0.00006557807,0.000025868529,0.0049905614,0.00029304586,0.96803164,0.0004862117,0.022854425],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9829199,0.0032436845,0.0011753757,0.0036979245,0.006304822,0.0026583339],"domain_scores_gemma":[0.9727183,0.0024651927,0.00040219596,0.022386666,0.0010705658,0.0009570557],"candidate_categories":["metaepi_narrow","sts","scholarly_communication","open_science","research_integrity","insufficient_payload"],"consensus_categories":["open_science"],"category_scores_codex":[0.01280601,0.0008184384,0.0010254164,0.0024982635,0.0013769228,0.0017003163,0.018384133,0.0011037599,0.00030473946],"category_scores_gemma":[0.0054375534,0.0008272286,0.00019125201,0.0029585264,0.0017882688,0.00115485,0.0220223,0.005771884,0.39714324],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00009667847,0.00025248274,0.000020492007,0.0007349297,0.0004067247,0.0061681173,0.000014339084,0.0000012279614,0.0005386648,0.000003869533,0.9914493,0.0003131885],"study_design_scores_gemma":[0.00034712604,0.00012034305,0.000113359354,0.00036051663,0.000118068325,0.00021359975,0.00009244419,0.000045713212,0.000114492555,0.00004999218,0.9976409,0.00078344706],"about_ca_topic_score_codex":0.00922037,"about_ca_topic_score_gemma":0.0013203728,"teacher_disagreement_score":0.3968385,"about_ca_system_score_codex":0.0011629006,"about_ca_system_score_gemma":0.002950958,"threshold_uncertainty_score":0.99992317},"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":"codex-gemma-dda1882f352a","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.000016134376,0.0017550343,0.0000017356942,0.00016906328,0.0048076026,0.0015520903,0.98880166,0.0008355057,0.0020611994],"genre_scores_gemma":[0.0000014981529,0.0015723143,0.00008826495,0.000033245797,0.008032849,0.00036143616,0.9686156,0.00056839595,0.020726377],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.979539,0.003466285,0.0013878329,0.0042843125,0.008037369,0.0032852518],"domain_scores_gemma":[0.9694653,0.0024741432,0.00049944455,0.024918236,0.0013644062,0.0012784837],"candidate_categories":["metaresearch","metaepi_narrow","sts","scholarly_communication","open_science","research_integrity","insufficient_payload"],"consensus_categories":["open_science","research_integrity"],"category_scores_codex":[0.014439261,0.001018484,0.0012007566,0.0033077933,0.0015910241,0.0018422678,0.019300647,0.0014565598,0.00033878465],"category_scores_gemma":[0.008960778,0.0010113323,0.00023060535,0.0042397208,0.0020161557,0.0015355067,0.024541268,0.0075745326,0.43129525],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0001620295,0.00035986418,0.000014557393,0.0009855331,0.00042331233,0.0052220793,0.000013211132,0.0000012998586,0.0017483587,0.000007469186,0.9908844,0.0001778683],"study_design_scores_gemma":[0.00050896214,0.00018616967,0.000098761884,0.00079104566,0.00014549764,0.00024463507,0.00014061455,0.00004988875,0.00045798742,0.00004474064,0.9963972,0.0009345184],"about_ca_topic_score_codex":0.017555527,"about_ca_topic_score_gemma":0.0018688351,"teacher_disagreement_score":0.43095645,"about_ca_system_score_codex":0.0013255201,"about_ca_system_score_gemma":0.0037954887,"threshold_uncertainty_score":0.9998398},"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":"codex-gemma-dda1882f352a","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.00008049182,0.0066152294,0.000010695191,0.00025766436,0.010843759,0.0009844942,0.98035264,0.00032873245,0.00052631967],"genre_scores_gemma":[0.000013016096,0.0023851618,0.000029072284,0.00008312297,0.0007881748,0.00019898382,0.9385068,0.00013534579,0.057860367],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9896173,0.0042803,0.0010124486,0.002250677,0.0006552955,0.0021839857],"domain_scores_gemma":[0.9871835,0.0017987724,0.00032711407,0.0099599995,0.0005771639,0.0001534382],"candidate_categories":["metaepi_narrow","sts","open_science","research_integrity","insufficient_payload"],"consensus_categories":["open_science","research_integrity"],"category_scores_codex":[0.0064939302,0.00071639277,0.0009988273,0.0015524297,0.0014654022,0.00035094467,0.00864859,0.0022199114,0.00081662805],"category_scores_gemma":[0.0027102071,0.0006735339,0.00019963548,0.0011773757,0.0013714774,0.00047248654,0.009814254,0.0053958814,0.1567341],"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":[0.0003064507,0.00020960241,0.0000014187158,0.00029927952,0.00058093946,0.0009737647,0.000021058391,1.9295348e-7,0.10335735,0.000033725435,0.89376014,0.00045604783],"study_design_scores_gemma":[0.00072925526,0.00033356235,0.000017796123,0.0003985222,0.000113018985,0.000531398,0.00019956753,0.000004717977,0.04343963,0.00006377983,0.95356536,0.0006034098],"about_ca_topic_score_codex":0.006145635,"about_ca_topic_score_gemma":0.00034138767,"teacher_disagreement_score":0.15591747,"about_ca_system_score_codex":0.00039867975,"about_ca_system_score_gemma":0.00203932,"threshold_uncertainty_score":0.99983454},"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":"codex-gemma-dda1882f352a","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.000026868192,0.0016390888,0.0000017326688,0.0001329141,0.0037510407,0.0014918139,0.99027854,0.000756564,0.0019214528],"genre_scores_gemma":[0.000005354149,0.001264144,0.00006706254,0.00003069781,0.0053281696,0.00032182995,0.9767295,0.00052804034,0.015725194],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.98137414,0.0036391758,0.0013077353,0.0040183566,0.006728029,0.0029325758],"domain_scores_gemma":[0.9723304,0.002238217,0.00047209888,0.02277229,0.0010564084,0.0011305822],"candidate_categories":["metaepi_narrow","sts","scholarly_communication","open_science","research_integrity","insufficient_payload"],"consensus_categories":["open_science","research_integrity"],"category_scores_codex":[0.012423848,0.00096744834,0.0011772312,0.0024720079,0.0015388333,0.0018450298,0.016349291,0.0013523687,0.00038833224],"category_scores_gemma":[0.007687389,0.0009537622,0.00021678528,0.0031666828,0.0018756398,0.0013140544,0.020596348,0.0061986153,0.3370325],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0001977346,0.0003183366,0.000012792,0.0007797204,0.00043683074,0.0043921606,0.00001565767,0.0000010527029,0.0017377613,0.0000060927277,0.9919124,0.00018943686],"study_design_scores_gemma":[0.00058516895,0.00016894614,0.0001399919,0.00062682957,0.00013502697,0.00024443556,0.00016312582,0.000051611838,0.00030750307,0.000048221442,0.99662995,0.0008991868],"about_ca_topic_score_codex":0.014032781,"about_ca_topic_score_gemma":0.0018995221,"teacher_disagreement_score":0.33664417,"about_ca_system_score_codex":0.0011966791,"about_ca_system_score_gemma":0.0032753658,"threshold_uncertainty_score":0.9999441},"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":"codex-gemma-dda1882f352a","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.000037683047,0.017434346,0.0000030707126,0.00014950277,0.0031908315,0.0012049541,0.9749606,0.00027738314,0.0027416192],"genre_scores_gemma":[0.0000064846768,0.003229576,0.0002287191,0.00006616808,0.008597792,0.00028990756,0.97726464,0.00038141155,0.009935273],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9772487,0.0057088584,0.0015816864,0.0048422464,0.0076781167,0.0029403775],"domain_scores_gemma":[0.96320933,0.002419174,0.0006053437,0.029823897,0.0025988682,0.0013433837],"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":[0.010862954,0.0010965554,0.0014842498,0.0022253762,0.0015398986,0.0027241644,0.017189346,0.0015281371,0.0012544087],"category_scores_gemma":[0.010942154,0.0011347273,0.0003178302,0.00423015,0.0015781376,0.0017083612,0.023886936,0.008476214,0.05111434],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":true,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00011632114,0.0006878663,0.00001636701,0.0012419381,0.00045311637,0.011254529,0.00001409629,0.0000010290003,0.0022122487,0.0000062334548,0.98368746,0.00030880293],"study_design_scores_gemma":[0.0005699728,0.000108606415,0.00006302186,0.0006993356,0.00025839327,0.00078529265,0.0002504532,0.000032911274,0.0007896219,0.00001523309,0.9953619,0.001065252],"about_ca_topic_score_codex":0.0043303072,"about_ca_topic_score_gemma":0.00051309075,"teacher_disagreement_score":0.04985993,"about_ca_system_score_codex":0.0018215143,"about_ca_system_score_gemma":0.007199344,"threshold_uncertainty_score":0.9997681},"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":"codex-gemma-dda1882f352a","genre_candidate":"dataset","genre_consensus":"dataset","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.00004275456,0.0014755918,0.0000025927461,0.00013467265,0.0043799654,0.0018643502,0.9888552,0.00089184346,0.002353076],"genre_scores_gemma":[0.000009441971,0.0015147086,0.00008781964,0.00004131307,0.005894573,0.00042331527,0.97209656,0.0006534981,0.019278778],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.97852147,0.0041876705,0.0015924531,0.0046487753,0.007561231,0.0034883763],"domain_scores_gemma":[0.97040075,0.002486828,0.00059849274,0.023847254,0.0012620378,0.0014046613],"candidate_categories":["metaresearch","metaepi_narrow","sts","scholarly_communication","open_science","research_integrity","insufficient_payload"],"consensus_categories":["open_science","research_integrity"],"category_scores_codex":[0.0133391535,0.0011985453,0.0014332982,0.0028506091,0.0018394721,0.0022203608,0.017540753,0.0016124828,0.00069237186],"category_scores_gemma":[0.0085235275,0.0011922112,0.0002838598,0.0035377766,0.002224158,0.0016027818,0.022112198,0.0070180986,0.3859072],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00027523786,0.00040939695,0.000017335757,0.0009777956,0.0005732444,0.0051095863,0.000022730788,0.0000015690068,0.0025323008,0.00000824399,0.9898258,0.00024677938],"study_design_scores_gemma":[0.00077255146,0.00021750234,0.00018643447,0.00079062104,0.00018183982,0.00031220602,0.00021849111,0.00005938102,0.00044953774,0.000053246415,0.9956361,0.0011220705],"about_ca_topic_score_codex":0.015051634,"about_ca_topic_score_gemma":0.0024550105,"teacher_disagreement_score":0.38521484,"about_ca_system_score_codex":0.0014837564,"about_ca_system_score_gemma":0.004006418,"threshold_uncertainty_score":0.9998281},"labels":[],"label_agreement":null}]}