{"meta":{"query_hash":"ddb50c5bddf8","filters":{"venue":"Harvard Data Science Review"},"cohort_total":18,"direct_labels_cover":0,"predictions_cover":18,"exported":18,"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/ddb50c5bddf8","api":"https://metacan.xera.ac/api/v1/cohort?venue=Harvard+Data+Science+Review"},"results":[{"id":"W2954249017","doi":"10.1162/99608f92.130f856e","title":"(A) Data in the Life: Authorship Attribution in Lennon-McCartney Songs","year":2019,"lang":"en","type":"article","venue":"Harvard Data Science Review","topic":"Music and Audio Processing","field":"Computer Science","cited_by":4,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Melody; Chord (peer-to-peer); Authorship attribution; Musical; Lyrics; Period (music)","score_opus":0.16064901104838888,"score_gpt":0.3739416594571497,"score_spread":0.2132926484087608,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2954249017","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.9096428,0.006207497,0.026615871,0.012253107,0.0006265852,0.00014197425,0.023441933,0.0003199958,0.020750124],"genre_scores_gemma":[0.9837335,0.000718874,0.006101095,0.00028992112,0.0001612654,0.000047405563,0.006365106,0.000021049256,0.00256183],"study_design_codex":"observational","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.99893063,0.0005012475,0.00011542575,0.00014934552,0.00024631535,0.000057088906],"domain_scores_gemma":[0.9796317,0.012916255,0.0028321675,0.0012622211,0.0028961361,0.0004614601],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0025354428,0.00026085484,0.00017084753,0.0018518382,0.00042791793,0.00094950385,0.0004240941,0.00049002265,0.0033807035],"category_scores_gemma":[0.028382871,0.000079339494,0.0001820945,0.0038110095,0.00063322193,0.00093113235,0.0006438967,0.000598489,0.0008054697],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00069320796,0.000086572654,0.72923404,0.00081842754,0.00012249287,0.0007671591,0.0033494509,0.0039178715,0.0016750153,0.009096727,0.043934125,0.20630494],"study_design_scores_gemma":[0.000051265022,0.00013293079,0.82481253,0.0006867766,0.00009306111,0.00091862964,0.004770938,0.044978485,0.0035192731,0.01771966,0.10224106,0.00007547013],"about_ca_topic_score_codex":0.011955757,"about_ca_topic_score_gemma":0.015896795,"teacher_disagreement_score":0.011955757,"about_ca_system_score_codex":0.000666104,"about_ca_system_score_gemma":0.00031386322,"threshold_uncertainty_score":0.023772359},"labels":[],"label_agreement":null},{"id":"W3172682396","doi":"10.1162/99608f92.8bb28793","title":"The Building Blocks of Statistical Education in the Data Science Ecosystem","year":2021,"lang":"en","type":"article","venue":"Harvard Data Science Review","topic":"Data Analysis with R","field":"Computer Science","cited_by":0,"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":"Ecosystem; Data science; Environmental science; Environmental resource management; Computer science; Ecology; Biology","score_opus":0.060116750503667465,"score_gpt":0.3821871329555472,"score_spread":0.3220703824518797,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3172682396","genre_codex":"commentary","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0031252815,0.1890635,0.20789866,0.5705324,0.005438528,0.00015217502,0.00030387717,0.00074185117,0.022743693],"genre_scores_gemma":[0.22919549,0.21816857,0.35891455,0.15340574,0.027888544,0.001308501,0.00053012185,0.0016212589,0.008967203],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9291757,0.047443807,0.0042869747,0.0043405537,0.012738126,0.0020148982],"domain_scores_gemma":[0.53760195,0.37814575,0.0058402084,0.026455345,0.043938316,0.008018426],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.1671763,0.0010776735,0.0025682796,0.009802377,0.0030833841,0.015327492,0.004240379,0.008033035,0.0058355047],"category_scores_gemma":[0.23670031,0.0016069387,0.0016760245,0.005594398,0.041171473,0.02793367,0.007473532,0.028484935,0.002029876],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00006350913,0.000081699596,0.0010000067,0.0008829995,0.00013332431,0.00006240906,0.0006136157,0.0011179856,0.00024400717,0.84841937,0.025340669,0.122040436],"study_design_scores_gemma":[0.0000515368,0.000072930714,0.0011881759,0.0015356295,0.000045839177,0.00008845757,0.0002585919,0.0019047691,0.0003776784,0.8760873,0.118303865,0.000085174106],"about_ca_topic_score_codex":0.012194465,"about_ca_topic_score_gemma":0.008938782,"teacher_disagreement_score":0.1671763,"about_ca_system_score_codex":0.008620433,"about_ca_system_score_gemma":0.020158477,"threshold_uncertainty_score":0.8841234},"labels":[],"label_agreement":null},{"id":"W4307407759","doi":"10.1162/99608f92.13e1d60e","title":"The Use of Data Science in a National Statistical Office","year":2022,"lang":"en","type":"article","venue":"Harvard Data Science Review","topic":"Data Analysis with R","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"Statistics Canada","funders":"","keywords":"Data science; Computer science","score_opus":0.25460396327041324,"score_gpt":0.39743772442241093,"score_spread":0.1428337611519977,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4307407759","genre_codex":"commentary","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0020944758,0.16200414,0.2050534,0.50194323,0.021154394,0.0003209406,0.0016225182,0.0010736194,0.10473328],"genre_scores_gemma":[0.13539393,0.24998574,0.36463273,0.1696845,0.04383723,0.0015721994,0.0027063238,0.0023335086,0.029853718],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.79518723,0.138755,0.014982324,0.012103217,0.03707491,0.0018972595],"domain_scores_gemma":[0.5702175,0.27803457,0.019696647,0.05236619,0.074508674,0.0051764906],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.15770067,0.0008676118,0.0014297365,0.0073794425,0.003691574,0.011695749,0.0035501905,0.0054299533,0.004681365],"category_scores_gemma":[0.24831685,0.0008169844,0.0010417129,0.011208033,0.023537809,0.015874447,0.0068185795,0.01504665,0.003556926],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000028650218,0.00002464159,0.0010123716,0.0011370508,0.00006332695,0.00007714754,0.00075667386,0.00070084445,0.00024006522,0.7596975,0.10724378,0.1290179],"study_design_scores_gemma":[0.000011600935,0.000030781524,0.00084320316,0.00205377,0.000028333268,0.00011531438,0.00037030756,0.0007384562,0.0002734263,0.11170656,0.88379127,0.000037012906],"about_ca_topic_score_codex":0.008457003,"about_ca_topic_score_gemma":0.0035745115,"teacher_disagreement_score":0.15770067,"about_ca_system_score_codex":0.009824232,"about_ca_system_score_gemma":0.024971092,"threshold_uncertainty_score":0.83401084},"labels":[],"label_agreement":null},{"id":"W4367177363","doi":"10.1162/99608f92.81172cdd","title":"On “A conversation with Sir David Cox”","year":2023,"lang":"en","type":"article","venue":"Harvard Data Science Review","topic":"Dental Education, Practice, Research","field":"Health Professions","cited_by":7,"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":"Conversation; Psychology; Psychoanalysis; Communication","score_opus":0.27616997298837553,"score_gpt":0.5653658902526287,"score_spread":0.2891959172642532,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4367177363","genre_codex":"commentary","genre_gemma":"other","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"other","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.000024683737,0.005429954,0.00009005279,0.9389057,0.054784227,0.000011131372,0.000050461014,0.000012008034,0.00069180626],"genre_scores_gemma":[0.0001985524,0.0014379441,0.000093045084,0.96899533,0.027333349,0.00003380737,0.000019196601,0.000022353299,0.001866318],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.9643561,0.01424276,0.004238283,0.003488154,0.011352372,0.0023223476],"domain_scores_gemma":[0.80055374,0.14912058,0.0084623,0.003987277,0.022790823,0.015085249],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.042816702,0.0023565535,0.004514732,0.0053761443,0.009917335,0.015319002,0.0062395674,0.060454838,0.01767555],"category_scores_gemma":[0.2377684,0.0019353007,0.0029216935,0.0048078103,0.009444307,0.019566685,0.009499338,0.07708996,0.01192726],"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":[0.000015836633,0.000004815027,0.000037220336,0.000067204,0.000011824879,0.000033524517,0.00007296322,0.0000060559837,0.000013913582,0.00085733074,0.99701166,0.0018675684],"study_design_scores_gemma":[0.000046277677,0.000013618711,0.00027406946,0.00075620995,0.000035334964,0.00007832663,0.000477735,0.000032943568,0.000041549247,0.003154612,0.9950394,0.000049857663],"about_ca_topic_score_codex":0.023514748,"about_ca_topic_score_gemma":0.050304946,"teacher_disagreement_score":0.060454838,"about_ca_system_score_codex":0.0073629394,"about_ca_system_score_gemma":0.01522293,"threshold_uncertainty_score":0.22643906},"labels":[],"label_agreement":null},{"id":"W4385297687","doi":"10.1162/99608f92.9ba2bd43","title":"Reinforcing Reproducibility and Replicability: An Introduction","year":2023,"lang":"en","type":"article","venue":"Harvard Data Science Review","topic":"BIM and Construction Integration","field":"Engineering","cited_by":4,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université du Québec à Montréal","funders":"National Science Foundation","keywords":"Reproducibility; Computer science; Statistics; Mathematics","score_opus":0.052198051255793125,"score_gpt":0.3084426474261471,"score_spread":0.25624459617035394,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4385297687","genre_codex":"commentary","genre_gemma":"commentary","domain_codex":null,"domain_gemma":"reproducibility","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"commentary","genre_consensus":"commentary","domain_candidate":"reproducibility","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0014932613,0.24795419,0.12857345,0.4913547,0.08038438,0.00053116475,0.0004235283,0.0005564016,0.048728894],"genre_scores_gemma":[0.08644813,0.30886194,0.126915,0.22419345,0.21170343,0.002186368,0.00060748286,0.0023572915,0.03672697],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.9301143,0.04375318,0.0052328305,0.0050472976,0.014389551,0.0014628682],"domain_scores_gemma":[0.72280943,0.23219325,0.0054934737,0.010431488,0.026842192,0.0022302037],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.06195481,0.0017525761,0.0017968537,0.006650229,0.005796846,0.018042933,0.0038263588,0.010550097,0.01059155],"category_scores_gemma":[0.17329612,0.0012165693,0.00210421,0.006616653,0.031198088,0.028713588,0.0073823,0.020566337,0.0039214804],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000059194583,0.000046249894,0.0005535533,0.0023454148,0.00005138389,0.00018295973,0.005782756,0.00063944055,0.00048676,0.7134391,0.19531134,0.08110183],"study_design_scores_gemma":[0.000015300175,0.000054367607,0.00031875548,0.003015654,0.000022269161,0.00024051862,0.0010400859,0.0006830131,0.0002902753,0.26540995,0.7288497,0.000060095033],"about_ca_topic_score_codex":0.0047907364,"about_ca_topic_score_gemma":0.004032794,"teacher_disagreement_score":0.9380452,"about_ca_system_score_codex":0.0087470915,"about_ca_system_score_gemma":0.010852004,"threshold_uncertainty_score":0.32765228},"labels":[],"label_agreement":null},{"id":"W4391378488","doi":"10.1162/99608f92.e3d09bff","title":"Discerning Audiences Through Like Buttons","year":2024,"lang":"en","type":"article","venue":"Harvard Data Science Review","topic":"Communication in Education and Healthcare","field":"Psychology","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University","funders":"Simon Fraser University","keywords":"Aesthetics; Art; Computer science","score_opus":0.2906058399536966,"score_gpt":0.5471223570002874,"score_spread":0.25651651704659084,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4391378488","genre_codex":"other","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.04264472,0.1232504,0.005054062,0.35244265,0.037069947,0.00023962352,0.0030170365,0.0005328022,0.43574876],"genre_scores_gemma":[0.44004002,0.122105874,0.0069866576,0.14374632,0.05026682,0.0006386405,0.002438454,0.0011808266,0.23259638],"study_design_codex":"not_applicable","study_design_gemma":"observational","domain_scores_codex":[0.9983632,0.0006697315,0.00008154932,0.00024071553,0.0004939847,0.00015080872],"domain_scores_gemma":[0.98570377,0.008720125,0.0012281514,0.0003222426,0.00320967,0.0008160482],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003442109,0.00030641392,0.00036520962,0.002423955,0.0015935829,0.0069273748,0.0005481479,0.0014394424,0.02417994],"category_scores_gemma":[0.017899187,0.00027571723,0.00024712706,0.0014111921,0.0018619949,0.008607834,0.0016076495,0.0024689045,0.0048350785],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00019129393,0.0000490664,0.0047536,0.0011778856,0.000025409834,0.00013691619,0.01377208,0.000024921477,0.00058395823,0.030671585,0.80198425,0.146629],"study_design_scores_gemma":[0.000018190618,0.000039008388,0.012794739,0.0012044169,0.000023706198,0.00015222961,0.00951266,0.0000735509,0.00040663764,0.0045749554,0.97117406,0.000025960666],"about_ca_topic_score_codex":0.0035547584,"about_ca_topic_score_gemma":0.009651462,"teacher_disagreement_score":0.02417994,"about_ca_system_score_codex":0.0008459448,"about_ca_system_score_gemma":0.0009356412,"threshold_uncertainty_score":0.08088994},"labels":[],"label_agreement":null},{"id":"W4391382837","doi":"10.1162/99608f92.d3197524","title":"Advancing Differential Privacy: Where We Are Now and Future Directions for Real-World Deployment","year":2024,"lang":"en","type":"article","venue":"Harvard Data Science Review","topic":"Defense, Military, and Policy Studies","field":"Economics, Econometrics and Finance","cited_by":35,"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 Waterloo","funders":"","keywords":"Differential privacy; Software deployment; Internet privacy; Computer science; Differential (mechanical device); Computer security; Engineering; Data mining; Aerospace engineering; Software engineering","score_opus":0.06757007992727386,"score_gpt":0.32236731613325914,"score_spread":0.25479723620598527,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4391382837","genre_codex":"methods","genre_gemma":"review","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"review","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.008173124,0.18524446,0.41428196,0.35934404,0.0028232234,0.00029354286,0.00045581133,0.0011886979,0.028195085],"genre_scores_gemma":[0.2411494,0.30793467,0.39768857,0.03813275,0.0066104126,0.00068695674,0.00087613956,0.0005896726,0.006331503],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9801216,0.011015789,0.0009839644,0.002106685,0.0044954116,0.00127659],"domain_scores_gemma":[0.90755326,0.06676883,0.0027377845,0.00967878,0.009492037,0.003769332],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.042460304,0.0013688151,0.001970284,0.0017868136,0.0023785196,0.012609806,0.005084927,0.009596006,0.0128620155],"category_scores_gemma":[0.06640749,0.0009276541,0.001721564,0.0033758327,0.0094515225,0.053219195,0.008312952,0.01662257,0.0035061014],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0002203169,0.00020827599,0.0028322062,0.002177682,0.00009278408,0.0002703666,0.0008353622,0.006271123,0.0018644847,0.5428176,0.03557689,0.40683287],"study_design_scores_gemma":[0.00006265954,0.0003740233,0.00090318586,0.0028778592,0.000042622974,0.0005764516,0.0022888042,0.011784304,0.0016890228,0.70900744,0.27028495,0.00010874043],"about_ca_topic_score_codex":0.00161179,"about_ca_topic_score_gemma":0.0017196063,"teacher_disagreement_score":0.042460304,"about_ca_system_score_codex":0.003598472,"about_ca_system_score_gemma":0.006188106,"threshold_uncertainty_score":0.22455424},"labels":[],"label_agreement":null},{"id":"W4392785285","doi":"10.1162/99608f92.35a2f3ff","title":"The Turing Transformation: Artificial Intelligence, Intelligence Augmentation, and Skill Premiums","year":2024,"lang":"en","type":"article","venue":"Harvard Data Science Review","topic":"Economic Development and Digital Transformation","field":"Economics, Econometrics and Finance","cited_by":7,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Creative Destruction Lab; University of Toronto","funders":"","keywords":"Transformation (genetics); Turing; Computer science; Artificial intelligence; Cognitive science; Cognitive psychology; Psychology; Programming language","score_opus":0.11922236168275696,"score_gpt":0.30994067499290423,"score_spread":0.19071831331014727,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4392785285","genre_codex":"other","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.37609243,0.071998425,0.08576479,0.08096451,0.0009158817,0.000123634,0.0009855379,0.00047827556,0.3826765],"genre_scores_gemma":[0.9798655,0.008000037,0.0041198726,0.0016773897,0.00050952076,0.00005109487,0.00014734315,0.00004697734,0.0055821673],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9986985,0.0004486678,0.000086451066,0.00023570853,0.0003664332,0.00016424236],"domain_scores_gemma":[0.9894539,0.006181076,0.0015168085,0.0014610646,0.00081046525,0.000576699],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0033860612,0.00036007393,0.0005489261,0.0022348461,0.00075515447,0.002967042,0.00092216965,0.0011095374,0.008558319],"category_scores_gemma":[0.015986666,0.00019962518,0.00066471775,0.0019412893,0.009287149,0.00651344,0.0024162498,0.002514222,0.00062198774],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00013777566,0.00008289904,0.011976637,0.00027784402,0.00007593008,0.000100076984,0.0004661677,0.004167547,0.00043191697,0.87721163,0.0054184655,0.099653095],"study_design_scores_gemma":[0.000024766834,0.00006797989,0.02594468,0.00021149809,0.00003117417,0.00017957827,0.00020099025,0.00512518,0.00047927356,0.9482409,0.0194688,0.000025197409],"about_ca_topic_score_codex":0.0021566255,"about_ca_topic_score_gemma":0.0014042972,"teacher_disagreement_score":0.008558319,"about_ca_system_score_codex":0.0019165247,"about_ca_system_score_gemma":0.0007697146,"threshold_uncertainty_score":0.028630495},"labels":[],"label_agreement":null},{"id":"W4394820768","doi":"10.1162/99608f92.d949f941","title":"Government Interventions to Avert Future Catastrophic AI Risks","year":2024,"lang":"en","type":"article","venue":"Harvard Data Science Review","topic":"Leadership, Behavior, and Decision-Making Studies","field":"Decision Sciences","cited_by":5,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université de Montréal","funders":"","keywords":"Government (linguistics); Psychological intervention; Business; Risk analysis (engineering); Psychology; Psychiatry","score_opus":0.4249351892336851,"score_gpt":0.5346952383454162,"score_spread":0.1097600491117311,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4394820768","genre_codex":"commentary","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.036142007,0.08959468,0.03817091,0.56899047,0.0072645517,0.0006575038,0.00036605002,0.00041673964,0.25839704],"genre_scores_gemma":[0.75923115,0.062285505,0.021417184,0.1358644,0.004013772,0.0012670768,0.0002608018,0.000111178204,0.0155489985],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9780813,0.014372577,0.0007237998,0.0009751766,0.0038093075,0.0020378283],"domain_scores_gemma":[0.94171447,0.040781014,0.006035818,0.00379984,0.006230528,0.0014383489],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.023650903,0.0005453671,0.00051497994,0.0026570447,0.0021913762,0.0042468063,0.0019273133,0.0055049206,0.0060518887],"category_scores_gemma":[0.09536259,0.00028441355,0.000822522,0.0017893814,0.005464025,0.004435601,0.00428901,0.005455183,0.00080010307],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00009319708,0.00027307728,0.007397887,0.001795157,0.00030185818,0.0003738243,0.0033219368,0.004234254,0.0005875605,0.652654,0.10310789,0.22585936],"study_design_scores_gemma":[0.00012455056,0.0002489643,0.012942099,0.0055445274,0.00022114548,0.00027910297,0.0064628795,0.002234465,0.0013661138,0.3129132,0.65758157,0.00008151213],"about_ca_topic_score_codex":0.0065764124,"about_ca_topic_score_gemma":0.011991392,"teacher_disagreement_score":0.023650903,"about_ca_system_score_codex":0.0044483016,"about_ca_system_score_gemma":0.016213417,"threshold_uncertainty_score":0.1250794},"labels":[],"label_agreement":null},{"id":"W4398764826","doi":"10.1162/99608f92.f86474d0","title":"A Familiar, Invisible Engine Is Driving the AI Revolution","year":2024,"lang":"en","type":"article","venue":"Harvard Data Science Review","topic":"Explainable Artificial Intelligence (XAI)","field":"Computer Science","cited_by":1,"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":"Computer science; Artificial intelligence; Aeronautics; Psychology; Automotive engineering; Engineering","score_opus":0.06234315868469324,"score_gpt":0.3537710257510638,"score_spread":0.2914278670663705,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4398764826","genre_codex":"commentary","genre_gemma":"commentary","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"commentary","genre_consensus":"commentary","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.020566288,0.039888423,0.037631955,0.63553876,0.0110323755,0.000071762974,0.00069918553,0.00086796243,0.2537034],"genre_scores_gemma":[0.61862016,0.046305533,0.03296184,0.18196854,0.019665245,0.00031378507,0.0008469281,0.0012773729,0.09804067],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.9913037,0.0034896263,0.0002901347,0.0013909569,0.0026662091,0.0008593736],"domain_scores_gemma":[0.98219866,0.009427145,0.0012678982,0.0022135526,0.0026440683,0.002248711],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0103401635,0.0008005183,0.0007181512,0.0026848696,0.0049394453,0.014403399,0.0016065516,0.006576578,0.0139559405],"category_scores_gemma":[0.028899197,0.00057992997,0.000789948,0.0025214697,0.021841496,0.024552917,0.0064749396,0.010071082,0.00643333],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000115074916,0.00008612173,0.0018499065,0.00031187848,0.000045106164,0.0001709222,0.002939414,0.0007580924,0.00069705164,0.8106794,0.13217163,0.050175354],"study_design_scores_gemma":[0.000030960837,0.000046692436,0.0017483903,0.00040875966,0.000022913808,0.00019163534,0.0022714708,0.0012367917,0.00046277168,0.42857856,0.56492746,0.00007353896],"about_ca_topic_score_codex":0.006734265,"about_ca_topic_score_gemma":0.004573849,"teacher_disagreement_score":0.014403399,"about_ca_system_score_codex":0.0051303105,"about_ca_system_score_gemma":0.0044526495,"threshold_uncertainty_score":0.0546847},"labels":[],"label_agreement":null},{"id":"W4400451566","doi":"10.1162/99608f92.6a5d7a87","title":"Differentially Private Linear Regression With Linked Data","year":2024,"lang":"en","type":"article","venue":"Harvard Data Science Review","topic":"Privacy-Preserving Technologies in Data","field":"Computer Science","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":"McGill University","funders":"","keywords":"Linear regression; Regression; Statistics; Mathematics","score_opus":0.10649766255760484,"score_gpt":0.3586239172977485,"score_spread":0.25212625474014366,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4400451566","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0046310453,0.0003052757,0.9937968,0.0005257108,0.000020231797,0.000023868184,0.00006303187,0.00015873874,0.000475256],"genre_scores_gemma":[0.47728375,0.0017465506,0.5153559,0.0007438667,0.00038746995,0.00035585603,0.0005900972,0.00021143696,0.0033251664],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9847074,0.010331736,0.00046698953,0.001801318,0.0023051614,0.00038751622],"domain_scores_gemma":[0.96172535,0.028340954,0.00236321,0.005840314,0.0014214873,0.00030862616],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01637228,0.0010964705,0.0017077244,0.0016794933,0.00090822444,0.003529729,0.0027366932,0.0025577934,0.0011018504],"category_scores_gemma":[0.055093843,0.00089705555,0.0013980882,0.0028222196,0.0031583507,0.005609598,0.0052756458,0.0041334657,0.00069302996],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00024338275,0.00009502353,0.0018094564,0.00019215702,0.00014502463,0.0002160595,0.0002963006,0.52099407,0.0022866235,0.373639,0.0017952475,0.09828769],"study_design_scores_gemma":[0.000023238024,0.000033128963,0.00019371364,0.000019490451,0.000012780285,0.000068582776,0.000019691559,0.81178486,0.0014695099,0.18513273,0.0012221235,0.000020062635],"about_ca_topic_score_codex":0.0011660801,"about_ca_topic_score_gemma":0.00069904694,"teacher_disagreement_score":0.01637228,"about_ca_system_score_codex":0.0026921956,"about_ca_system_score_gemma":0.0020763534,"threshold_uncertainty_score":0.08658594},"labels":[],"label_agreement":null},{"id":"W4406435392","doi":"10.1162/99608f92.6f5dfc6f","title":"Data Literacy in Industry: High Time to Focus on Operationalization Through Middle Managers","year":2025,"lang":"en","type":"article","venue":"Harvard Data Science Review","topic":"Data Quality and Management","field":"Decision Sciences","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"","funders":"Dalhousie University","keywords":"Operationalization; Focus (optics); Middle management; Business; Literacy; Psychology; Marketing; Pedagogy","score_opus":0.33414786909287836,"score_gpt":0.49253170853835254,"score_spread":0.15838383944547418,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4406435392","genre_codex":"commentary","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.005490908,0.19239122,0.0057593086,0.7749096,0.005844197,0.00004683624,0.00005870078,0.00006918951,0.015430074],"genre_scores_gemma":[0.2669588,0.24958594,0.012729875,0.43465132,0.016925208,0.00041577063,0.00024145274,0.00030971706,0.018181968],"study_design_codex":"design_other","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9768754,0.011960746,0.0014273878,0.0017089332,0.005608947,0.0024186184],"domain_scores_gemma":[0.86007667,0.08586016,0.006521391,0.005824016,0.027848974,0.013868895],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.060479324,0.00038687413,0.0009838102,0.0034603951,0.0030875395,0.012727686,0.002321777,0.005479942,0.006899583],"category_scores_gemma":[0.08397029,0.00050374906,0.0005988291,0.0026679453,0.008284007,0.02298473,0.00886511,0.013044515,0.00184855],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00013949943,0.00015616896,0.0041244472,0.004344674,0.000068962516,0.00023252734,0.009593167,0.00025368747,0.0007260648,0.2454442,0.19530165,0.539615],"study_design_scores_gemma":[0.00004230674,0.00015946507,0.005649219,0.008376099,0.000033912624,0.00017616246,0.008478012,0.00025659418,0.00046784364,0.07408214,0.9022263,0.00005179878],"about_ca_topic_score_codex":0.007448986,"about_ca_topic_score_gemma":0.00908864,"teacher_disagreement_score":0.060479324,"about_ca_system_score_codex":0.007583472,"about_ca_system_score_gemma":0.024309337,"threshold_uncertainty_score":0.31984907},"labels":[],"label_agreement":null},{"id":"W4406971547","doi":"10.1162/99608f92.7fcd521d","title":"What Are Statistical Assumptions About? An Answer From Perspectivism","year":2025,"lang":"en","type":"preprint","venue":"Harvard Data Science Review","topic":"Statistics Education and Methodologies","field":"Mathematics","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":"Simon Fraser University","funders":"","keywords":"Perspectivism; Epistemology; Computer science; Econometrics; Mathematics; Philosophy","score_opus":0.5070296381043464,"score_gpt":0.5557659503054858,"score_spread":0.04873631220113939,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4406971547","genre_codex":"commentary","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.008041171,0.017261365,0.23289257,0.70163214,0.0031340274,0.000046005287,0.00043161723,0.00018889456,0.03637224],"genre_scores_gemma":[0.7118878,0.022016156,0.097563215,0.13875104,0.022847516,0.00050109555,0.0007472221,0.00080526125,0.0048806826],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.92920244,0.04661575,0.0032471481,0.006664574,0.012921308,0.0013489058],"domain_scores_gemma":[0.7022638,0.25344014,0.009193217,0.016956113,0.015755905,0.00239086],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.07930588,0.0011865771,0.0034657693,0.003461604,0.0043374407,0.012349738,0.0042942613,0.010910218,0.0065242606],"category_scores_gemma":[0.21985768,0.0011913929,0.0020209346,0.0031992854,0.053854894,0.054095503,0.008812854,0.023321563,0.0018611922],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0000302296,0.0000118442695,0.00030967657,0.00014297226,0.00004156681,0.000032992655,0.0013952741,0.0004091069,0.000039816023,0.9847009,0.005882724,0.007002861],"study_design_scores_gemma":[0.000009075843,0.0000041919657,0.00006924812,0.00008769424,0.000004357296,0.000015550491,0.0003104424,0.00028618856,0.000024179317,0.9929427,0.006238849,0.000007565064],"about_ca_topic_score_codex":0.002895013,"about_ca_topic_score_gemma":0.00090468675,"teacher_disagreement_score":0.07930588,"about_ca_system_score_codex":0.0061249435,"about_ca_system_score_gemma":0.0050877933,"threshold_uncertainty_score":0.41941458},"labels":[],"label_agreement":null},{"id":"W4407949312","doi":"10.1162/99608f92.d605e50f","title":"The Problem of Terroir in the Anthropocene","year":2025,"lang":"en","type":"article","venue":"Harvard Data Science Review","topic":"Global Energy and Sustainability Research","field":"Energy","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 British Columbia","funders":"","keywords":"Anthropocene; Terroir; Environmental ethics; Geography; History; Archaeology; Philosophy; Art","score_opus":0.041182875888009673,"score_gpt":0.3828525447909881,"score_spread":0.34166966890297845,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4407949312","genre_codex":"review","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.019299613,0.5477925,0.013392805,0.3264773,0.010247129,0.00006654135,0.0015602243,0.0001519556,0.08101198],"genre_scores_gemma":[0.26289088,0.60196346,0.01633935,0.082598604,0.011020399,0.00028007422,0.0019499353,0.00036476468,0.02259246],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9910166,0.0045575197,0.0008813934,0.0010696921,0.0020304467,0.00044426686],"domain_scores_gemma":[0.9761254,0.015226619,0.0016303436,0.002629158,0.003793523,0.00059493707],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.016514147,0.0003220273,0.000855147,0.0021348067,0.0023565937,0.006160604,0.0018644895,0.001965489,0.0056878105],"category_scores_gemma":[0.026952535,0.00033956423,0.0005726637,0.00671832,0.012212894,0.020500654,0.0046372525,0.0068411822,0.000782225],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00011217565,0.000021123986,0.0052748793,0.0025125053,0.000083974825,0.00029260272,0.011241999,0.00042652953,0.0003403328,0.5327879,0.08246309,0.36444277],"study_design_scores_gemma":[0.0000044819126,0.000015188903,0.003612093,0.0015699563,0.000014026221,0.00015336329,0.004514212,0.00006813851,0.00010464973,0.04531059,0.94461423,0.000019023684],"about_ca_topic_score_codex":0.014228476,"about_ca_topic_score_gemma":0.017261526,"teacher_disagreement_score":0.016514147,"about_ca_system_score_codex":0.003776154,"about_ca_system_score_gemma":0.0066602025,"threshold_uncertainty_score":0.08733624},"labels":[],"label_agreement":null},{"id":"W4409640743","doi":"10.1162/99608f92.abf14c9d","title":"Challenges and Opportunities for Statistics in the Era of Data Science","year":2025,"lang":"en","type":"article","venue":"Harvard Data Science Review","topic":"Big Data and Business Intelligence","field":"Business, Management and Accounting","cited_by":3,"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 Alberta","funders":"","keywords":"Data science; Statistics; Computer science; Mathematics","score_opus":0.47368453054411114,"score_gpt":0.41809491985268776,"score_spread":0.055589610691423375,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4409640743","genre_codex":"commentary","genre_gemma":"empirical","domain_codex":null,"domain_gemma":"methods","model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":"methods","domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0010793784,0.21702646,0.044163533,0.7221779,0.007605814,0.000035856276,0.00022385523,0.00015130517,0.0075358977],"genre_scores_gemma":[0.11542,0.46029297,0.10963847,0.20264079,0.10522365,0.0005534158,0.00053367036,0.00067540724,0.005021674],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9058151,0.063900396,0.003883298,0.0051575815,0.019443257,0.0018002465],"domain_scores_gemma":[0.58309454,0.3701651,0.005110968,0.014512947,0.021509085,0.005607412],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.1518123,0.0014954247,0.004271997,0.008697428,0.0050451388,0.02015693,0.004269716,0.017157223,0.005668201],"category_scores_gemma":[0.18255344,0.0013493205,0.0025443146,0.008862064,0.048484977,0.055268932,0.010828593,0.03634132,0.0027593998],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000043762266,0.00003582269,0.00033102353,0.0006649467,0.00004664982,0.0000693115,0.0007263682,0.00077377784,0.000108383436,0.9318642,0.03000138,0.035334367],"study_design_scores_gemma":[0.0000165873,0.000027625221,0.0002584491,0.00083559245,0.000010272029,0.00007943179,0.00058870215,0.0009767711,0.00005913998,0.89671314,0.10039032,0.000044034226],"about_ca_topic_score_codex":0.0034543986,"about_ca_topic_score_gemma":0.0018805213,"teacher_disagreement_score":0.8481877,"about_ca_system_score_codex":0.008853793,"about_ca_system_score_gemma":0.01819619,"threshold_uncertainty_score":0.8028698},"labels":[],"label_agreement":null},{"id":"W4411542049","doi":"10.1162/99608f92.dc58eb7a","title":"Navigating the Privacy Landscape: Harmonizing Legislative and Public Sector Approaches in the Canadian Context","year":2025,"lang":"en","type":"article","venue":"Harvard Data Science Review","topic":"Criminal Law and Evidence","field":"Social Sciences","cited_by":2,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"Statistics Canada","funders":"","keywords":"Legislature; Context (archaeology); Public sector; Political science; Public relations; Internet privacy; Public administration; Geography; Computer science; Law; Archaeology","score_opus":0.35686093694706433,"score_gpt":0.4034727452268728,"score_spread":0.046611808279808487,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4411542049","genre_codex":"commentary","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.09525631,0.025651786,0.014306633,0.62022316,0.0011000218,0.0004118767,0.0007353351,0.00017381915,0.24214098],"genre_scores_gemma":[0.9282249,0.013633707,0.011385784,0.032373164,0.00020635204,0.00018661511,0.00024212075,0.0001255542,0.013621795],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.9234424,0.027209716,0.0025351446,0.0049359156,0.02523106,0.01664585],"domain_scores_gemma":[0.90353364,0.0407812,0.0032925066,0.0034941386,0.03897553,0.009922929],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.062363848,0.00069761573,0.0011894209,0.011701681,0.06430032,0.04348149,0.007844977,0.0075821048,0.0040682536],"category_scores_gemma":[0.07501945,0.0011436876,0.001036062,0.019967975,0.059576016,0.01079135,0.01427923,0.011517581,0.00027182372],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_system_candidate":true,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00004903227,0.000044571094,0.0073970323,0.0004990572,0.000060978895,0.0006324354,0.17446668,0.0021463905,0.0005928638,0.731578,0.037827324,0.04470574],"study_design_scores_gemma":[0.000033255594,0.000037601887,0.020969642,0.0020090665,0.00013367833,0.00022085742,0.31861824,0.0022092545,0.00086479814,0.098812796,0.5557391,0.00035179686],"about_ca_topic_score_codex":0.99643886,"about_ca_topic_score_gemma":0.9979838,"teacher_disagreement_score":0.4934841,"about_ca_system_score_codex":0.4934841,"about_ca_system_score_gemma":0.7156559,"threshold_uncertainty_score":0.5874866},"labels":[],"label_agreement":null},{"id":"W4415667311","doi":"10.1162/99608f92.28e9f828","title":"An Undergraduate Course in Causality","year":2025,"lang":"en","type":"article","venue":"Harvard Data Science Review","topic":"Statistics Education and Methodologies","field":"Mathematics","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":"Course (navigation); Set (abstract data type); Causality (physics); Quarter (Canadian coin); Graduate students; Advice (programming)","score_opus":0.383188568202988,"score_gpt":0.5747813044805024,"score_spread":0.19159273627751433,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4415667311","genre_codex":"other","genre_gemma":"other","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"other","genre_consensus":"other","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.022734968,0.03095343,0.09227399,0.1561235,0.04660886,0.0016589215,0.024519088,0.013917691,0.6112096],"genre_scores_gemma":[0.11061087,0.022477094,0.07088618,0.06427172,0.019694364,0.0019730248,0.026892845,0.0032110526,0.67998284],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99887425,0.0002177528,0.00009275222,0.0001839609,0.00041743138,0.00021384351],"domain_scores_gemma":[0.9942749,0.001506237,0.00026201794,0.00041477015,0.0015138066,0.0020282331],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0039191493,0.0013268061,0.001269529,0.0018315311,0.0012454232,0.0024475942,0.0010587606,0.0014969982,0.27495375],"category_scores_gemma":[0.010405201,0.0004923027,0.0010347625,0.001616483,0.0006963753,0.0031850303,0.002829715,0.0026553646,0.07717764],"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":[0.00005231241,0.00029414528,0.0009523887,0.00026558482,0.000011534552,0.00011169414,0.00017394686,0.00025404783,0.00041894073,0.010292412,0.8556942,0.13147876],"study_design_scores_gemma":[0.000056377285,0.00014042326,0.0036041725,0.00042617484,0.000014533875,0.00025432752,0.00022530943,0.0006180539,0.00030104627,0.03309239,0.96125346,0.000013674984],"about_ca_topic_score_codex":0.0017792382,"about_ca_topic_score_gemma":0.004085799,"teacher_disagreement_score":0.27495375,"about_ca_system_score_codex":0.0024943312,"about_ca_system_score_gemma":0.0050914516,"threshold_uncertainty_score":0.91981196},"labels":[],"label_agreement":null},{"id":"W4415741689","doi":"10.1162/99608f92.dbc60044","title":"AI and Writing in the Straddle Generation","year":2025,"lang":"en","type":"article","venue":"Harvard Data Science Review","topic":"Cybernetics and Technology in Society","field":"Arts and Humanities","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"University of Toronto","funders":"","keywords":"Column (typography); Straddle; Face (sociological concept); Technical writing; Affect (linguistics)","score_opus":0.1138791503751696,"score_gpt":0.3484377760184191,"score_spread":0.23455862564324947,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4415741689","genre_codex":"review","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.004577162,0.765931,0.00065004017,0.17325406,0.0058014444,0.000011040216,0.000054610755,0.000030109351,0.049690537],"genre_scores_gemma":[0.14064655,0.7471969,0.0008497859,0.056512125,0.010855337,0.00004740408,0.00013676683,0.00010840972,0.043646716],"study_design_codex":"not_applicable","study_design_gemma":"qualitative","domain_scores_codex":[0.9961654,0.0019648008,0.00027181735,0.00029133752,0.0009795723,0.0003271188],"domain_scores_gemma":[0.98653257,0.008641291,0.0009769467,0.0005102285,0.0021594905,0.0011795418],"candidate_categories":["sts"],"consensus_categories":[],"category_scores_codex":[0.00526403,0.0002059695,0.00042537213,0.002285286,0.0020854184,0.006230391,0.00065418764,0.002517562,0.005205438],"category_scores_gemma":[0.011182545,0.0002003255,0.0003146485,0.0017599737,0.008165288,0.00602417,0.002046612,0.0051032114,0.0017335651],"study_design_candidate":"qualitative","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0000714467,0.00008105011,0.0016261986,0.001499691,0.00005898129,0.00052741444,0.018578555,0.00021066792,0.0003797936,0.27657047,0.35753337,0.3428624],"study_design_scores_gemma":[0.000008935134,0.000025048525,0.001560674,0.00085567107,0.000008641221,0.0003859875,0.002657985,0.000028098346,0.00010737863,0.017379712,0.9769687,0.0000130925155],"about_ca_topic_score_codex":0.010317421,"about_ca_topic_score_gemma":0.0123173315,"teacher_disagreement_score":0.99791455,"about_ca_system_score_codex":0.0028848753,"about_ca_system_score_gemma":0.0042775827,"threshold_uncertainty_score":0.027839184},"labels":[],"label_agreement":null}]}