{"meta":{"query_hash":"71da544226a3","filters":{"venue":"Caltech Library"},"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/71da544226a3","api":"https://metacan.xera.ac/api/v1/cohort?venue=Caltech+Library"},"results":[{"id":"W6905035831","doi":"10.14291/tccon.ggg2014.eureka01.r3","title":"TCCON data from Eureka (CA), Release GGG2014.R3","year":2019,"lang":"en","type":"dataset","venue":"Caltech Library","topic":"","field":"","cited_by":40,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"Dalhousie University; Environment and Climate Change Canada; University of Toronto","funders":"","keywords":"Atmosphere (unit); Column (typography); Data set; Atmospheric composition; Spectrometer; Carbon fibers; Calibration","score_opus":0.06334565290831067,"score_gpt":0.2867834150941476,"score_spread":0.22343776218583694,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W6905035831","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":[0.00013350163,0.0061030695,0.000007747588,0.00034180624,0.0024811332,0.001370299,0.9871748,0.0010383818,0.0013492544],"genre_scores_gemma":[0.000010744542,0.001484538,0.0006963085,0.0022613236,0.0032944293,0.00008658459,0.98783976,0.0008179865,0.0035083157],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9927507,0.0006153065,0.0011432132,0.0031447306,0.0011619183,0.0011841393],"domain_scores_gemma":[0.98112905,0.00066864677,0.0008934286,0.016492017,0.000043143733,0.0007737214],"candidate_categories":["metaepi_narrow","open_science","research_integrity","insufficient_payload"],"consensus_categories":["metaepi_narrow","open_science","research_integrity","insufficient_payload"],"category_scores_codex":[0.00034867233,0.001377939,0.0015239774,0.00069579383,0.0001537227,0.0006155164,0.010265539,0.0017504088,0.02340347],"category_scores_gemma":[0.0005967975,0.0013734987,0.00027517413,0.00089085946,0.00030863742,0.0044757626,0.008258643,0.0030663921,0.3892174],"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":[0.00025915,0.00025018866,0.000073119765,0.00031961405,0.0002397292,0.00068415765,0.0000027435483,0.0000023656044,0.00017637438,0.000008619441,0.9976467,0.00033724977],"study_design_scores_gemma":[0.0010059864,0.00007448733,0.00014912334,0.00058479776,0.00043506513,0.000022123208,0.0000092440905,0.00021941788,0.00027400453,0.0004610902,0.99523056,0.0015341125],"about_ca_topic_score_codex":0.0005203845,"about_ca_topic_score_gemma":0.000029604116,"teacher_disagreement_score":0.36581394,"about_ca_system_score_codex":0.000096794116,"about_ca_system_score_gemma":0.001230383,"threshold_uncertainty_score":0.9998971},"labels":[],"label_agreement":null},{"id":"W6923360649","doi":"10.14291/tccon.ggg2020.documentation.r0","title":"The Total Carbon Column Observing Network's GGG2020 Data Version: Data Quality, Comparison with GGG2014, and Future Outlook","year":2025,"lang":"en","type":"article","venue":"Caltech Library","topic":"Sociology and Education Studies","field":"Social Sciences","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Environment and Climate Change Canada; University of Toronto","funders":"","keywords":"Column (typography); A priori and a posteriori; Carbon fibers; Work (physics); Carbon flux; Data set","score_opus":0.11061005700987941,"score_gpt":0.39470623063699983,"score_spread":0.2840961736271204,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W6923360649","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.47043914,0.074330725,0.00008190025,0.31206304,0.0076360772,0.0010264096,0.00026532914,0.0006260892,0.13353129],"genre_scores_gemma":[0.9673386,0.0063291085,0.0013811667,0.0032445656,0.0055967784,0.000035009136,0.0008980824,0.00002469613,0.015151944],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9985622,0.00039821383,0.00019383409,0.0003897414,0.00016866952,0.00028734983],"domain_scores_gemma":[0.9979458,0.00090604887,0.00008951982,0.00095784717,0.000026720298,0.0000741034],"candidate_categories":["sts"],"consensus_categories":[],"category_scores_codex":[0.0008005334,0.00010289164,0.00020013377,0.000016386251,0.0017749276,0.00016377975,0.001173212,0.0001376212,0.000040512547],"category_scores_gemma":[0.00014225593,0.00007537042,0.000013140663,0.00034272383,0.0006894505,0.00058996276,0.0017208675,0.00027964645,0.0000057369784],"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":[0.000054192933,0.000021108086,0.28223678,0.000017879323,0.00008499008,0.0000012949562,0.0030501469,0.0000021492733,7.53379e-7,0.0116575,0.6992185,0.0036546562],"study_design_scores_gemma":[0.0001955801,0.00001690806,0.12247308,0.00003233304,0.000041781892,4.6746993e-7,0.04752169,0.00020314046,0.0000018729969,0.0014098918,0.82797885,0.00012441727],"about_ca_topic_score_codex":0.0003906657,"about_ca_topic_score_gemma":0.0002793522,"teacher_disagreement_score":0.49689952,"about_ca_system_score_codex":0.00001642551,"about_ca_system_score_gemma":0.00040673063,"threshold_uncertainty_score":0.99952465},"labels":[],"label_agreement":null},{"id":"W6940426629","doi":"10.7907/f2ag-se23","title":"Cyclogenesis and Effect of Turbulence on Fronts in the Northern Great Plains: Part I. Cyclogenesis in the Northern Great Plains. II. Turbulence in the Frontogenetical Field","year":2025,"lang":"en","type":"article","venue":"Caltech Library","topic":"Mycorrhizal Fungi and Plant Interactions","field":"Agricultural and Biological Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"","keywords":"Cyclogenesis; Tropical cyclogenesis; Extratropical cyclone; Cyclone (programming language); Tropical cyclone; Table (database); Synoptic scale meteorology; Warm front","score_opus":0.006864149552296023,"score_gpt":0.2047188847134817,"score_spread":0.19785473516118568,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W6940426629","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.9877126,0.0026920752,0.0000013712253,0.0054126293,0.0001293124,0.000664677,0.00030739518,0.000021293517,0.0030586338],"genre_scores_gemma":[0.99522656,0.0017489934,0.000009736881,0.0024812317,0.00011062753,0.00015224356,0.0001567762,0.0000022802333,0.00011151868],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.99783844,0.0008137713,0.00036011668,0.00039248343,0.00025912662,0.00033606341],"domain_scores_gemma":[0.99313265,0.006515107,0.00007128799,0.00023137752,0.000010849712,0.0000387058],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0005007897,0.00024905888,0.00029612574,0.00005335696,0.00021341967,0.00007076448,0.00093575567,0.00019072201,0.00022095836],"category_scores_gemma":[0.00023009516,0.000076046425,0.0001091976,0.0006173044,0.00004880958,0.00015197537,0.00011764219,0.00038435246,0.000018046687],"study_design_candidate":"observational","study_design_consensus":"observational","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00036174536,0.00015677112,0.9586323,0.00003621156,0.000013167405,0.00007253738,0.00070527865,0.000053628642,0.0008036423,0.00026353085,0.0022255664,0.036675613],"study_design_scores_gemma":[0.0004788095,0.001713889,0.95631224,0.0008925036,0.00006908744,0.000073773044,0.0014158319,0.0005354226,0.0054552634,0.00222236,0.030326203,0.0005046188],"about_ca_topic_score_codex":0.0018374195,"about_ca_topic_score_gemma":0.008005552,"teacher_disagreement_score":0.036170993,"about_ca_system_score_codex":0.000015899328,"about_ca_system_score_gemma":0.000017829288,"threshold_uncertainty_score":0.44672862},"labels":[],"label_agreement":null},{"id":"W6942490859","doi":"10.14291/tccon.ggg2014.easttroutlake01.r0/1348207","title":"TCCON data from East Trout Lake, SK (CA), Release GGG2014.R0","year":2017,"lang":"en","type":"dataset","venue":"Caltech Library","topic":"Mycorrhizal Fungi and Plant Interactions","field":"Agricultural and Biological Sciences","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"Environment and Climate Change Canada; University of Toronto","funders":"Ontario Ministry of Research and Innovation","keywords":"Trout; Atmosphere (unit); Data set; Water column; Hydrology (agriculture); Absorption (acoustics)","score_opus":0.04256040852875715,"score_gpt":0.24901091187582197,"score_spread":0.20645050334706483,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W6942490859","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":[0.0013613139,0.0011080984,7.855332e-7,0.0012680503,0.0016116468,0.00028796773,0.99081486,0.00016719576,0.003380075],"genre_scores_gemma":[0.00013596476,0.004028698,0.000052491796,0.00095325004,0.0036415781,0.000022305001,0.98772883,0.0000044330227,0.0034324606],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9977282,0.00014452942,0.00040698436,0.0009809255,0.00029540714,0.00044390783],"domain_scores_gemma":[0.99686784,0.0012525857,0.00036329858,0.0011492327,0.00002187269,0.0003451676],"candidate_categories":["insufficient_payload"],"consensus_categories":["insufficient_payload"],"category_scores_codex":[0.000099764504,0.00044327846,0.00050502404,0.000032638134,0.0004642662,0.0006000964,0.0034616175,0.0006154579,0.13585488],"category_scores_gemma":[0.00035576997,0.0001851961,0.00017433986,0.00013087207,0.000031594274,0.0013956281,0.0012743926,0.000923404,0.01961611],"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":[0.00008151119,0.00009081773,0.000433792,0.000020251795,0.00004219403,0.0001328271,0.0000016347208,1.05964276e-7,0.00038266077,0.00000927429,0.9951089,0.0036960253],"study_design_scores_gemma":[0.000051339175,0.000080294194,0.0006118832,0.00022461388,0.00008323717,0.000025511335,0.000013463418,0.00003590882,0.000039503757,0.000391881,0.9979702,0.00047212213],"about_ca_topic_score_codex":0.0027575637,"about_ca_topic_score_gemma":0.0021057334,"teacher_disagreement_score":0.11623878,"about_ca_system_score_codex":0.000008980559,"about_ca_system_score_gemma":0.000058292193,"threshold_uncertainty_score":0.9811472},"labels":[],"label_agreement":null},{"id":"W6942661636","doi":"10.14291/tccon.ggg2020.documentation.r1","title":"The Total Carbon Column Observing Network's GGG2020 Data Version: Data Quality, Comparison with GGG2014, and Future Outlook","year":2025,"lang":"en","type":"article","venue":"Caltech Library","topic":"Mycorrhizal Fungi and Plant Interactions","field":"Agricultural and Biological Sciences","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":"Environment and Climate Change Canada; University of Toronto","funders":"","keywords":"Column (typography); A priori and a posteriori; Carbon fibers; Work (physics); Carbon flux; Data set","score_opus":0.03747287796784141,"score_gpt":0.26044495586485283,"score_spread":0.22297207789701143,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W6942661636","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.839122,0.027944557,0.000023231989,0.067608066,0.0030321341,0.0009382418,0.010836433,0.0005188245,0.04997651],"genre_scores_gemma":[0.89073384,0.011619305,0.0013498424,0.004098174,0.007902486,0.000034215765,0.06867644,0.000010402886,0.015575299],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99878967,0.00013963046,0.00022281823,0.0004453456,0.00014986238,0.00025267067],"domain_scores_gemma":[0.9976101,0.0017271664,0.00008137844,0.00047514762,0.000015686379,0.000090527894],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00022305551,0.00013833829,0.00018769287,0.0000064797596,0.000528745,0.00030701447,0.0012028445,0.0000986329,0.0003911895],"category_scores_gemma":[0.00004532432,0.00004884506,0.000019888228,0.0003440157,0.000025790738,0.00059389527,0.0017170042,0.00029775413,0.000019527834],"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":[0.00024759618,0.000042741314,0.16612217,0.000026102643,0.000057634927,0.00000979789,0.000021854461,0.000009283178,0.0007962868,0.0025721444,0.7990533,0.031041091],"study_design_scores_gemma":[0.00006280432,0.00007669082,0.0348512,0.00009037151,0.000032057476,0.00001340632,0.0005792897,0.005301139,0.00004901684,0.00031011165,0.9584715,0.00016241035],"about_ca_topic_score_codex":0.0004375239,"about_ca_topic_score_gemma":0.0003614061,"teacher_disagreement_score":0.1594182,"about_ca_system_score_codex":0.0000054584425,"about_ca_system_score_gemma":0.000029519664,"threshold_uncertainty_score":0.4283251},"labels":[],"label_agreement":null},{"id":"W6960943120","doi":"10.14291/tccon.ggg2020.indianapolis01.r0","title":"TCCON data from Indianapolis (US), Release GGG2020.R0","year":2022,"lang":"en","type":"dataset","venue":"Caltech Library","topic":"","field":"","cited_by":3,"is_retracted":false,"has_abstract":false,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"","keywords":"Hydrology (agriculture); Calibration; Hydrograph; Surface runoff; Data collection","score_opus":0.04171742265221553,"score_gpt":0.2709516004316027,"score_spread":0.22923417777938718,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W6960943120","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":[0.0003016806,0.0042762086,0.0000010608552,0.00053324935,0.0018242536,0.0009437565,0.9896425,0.0013805955,0.0010966976],"genre_scores_gemma":[0.000008442187,0.0015128427,0.00033866905,0.0044962573,0.0023213227,0.00022327017,0.9894779,0.0007724902,0.000848816],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9922192,0.0008680093,0.0011454492,0.0030521203,0.0014708311,0.0012443635],"domain_scores_gemma":[0.9845987,0.0005852057,0.00091703207,0.01290422,0.000021717147,0.0009731137],"candidate_categories":["metaepi_narrow","open_science","research_integrity","insufficient_payload"],"consensus_categories":["open_science","insufficient_payload"],"category_scores_codex":[0.00036579277,0.0012547192,0.0013422898,0.0010610736,0.00043506201,0.0005810067,0.011596346,0.0010883821,0.44942182],"category_scores_gemma":[0.0006162059,0.0014127591,0.00028367274,0.0019271915,0.00033516675,0.0037288468,0.0155855,0.004277515,0.03872084],"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":[0.00021194074,0.00028835895,0.00008930947,0.00018363325,0.00030910695,0.0027376025,0.00001195931,0.0000024665333,0.0000304335,0.000011194109,0.9957212,0.00040275478],"study_design_scores_gemma":[0.0007910279,0.00008246004,0.00020002828,0.00011783046,0.00040401024,0.000031437823,0.000040773888,0.00003823384,0.000110906054,0.0005446909,0.9961386,0.0015000105],"about_ca_topic_score_codex":0.003745988,"about_ca_topic_score_gemma":0.00014834724,"teacher_disagreement_score":0.41070098,"about_ca_system_score_codex":0.00023784845,"about_ca_system_score_gemma":0.001867839,"threshold_uncertainty_score":0.99883217},"labels":[],"label_agreement":null}]}