{"meta":{"query_hash":"802234436123","filters":{"venue":"Linguistics"},"cohort_total":17,"direct_labels_cover":0,"predictions_cover":17,"exported":17,"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/802234436123","api":"https://metacan.xera.ac/api/v1/cohort?venue=Linguistics"},"results":[{"id":"W2508414753","doi":"10.1515/ling-2016-0019","title":"The lexicon in Functional Discourse Grammar: Theory, typology, description","year":2016,"lang":"en","type":"article","venue":"Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":8,"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 Lethbridge","funders":"","keywords":"Lexeme; Lexicon; Linguistics; Computer science; Morpheme; Grammar; Natural language processing; Artificial intelligence; Lexical item; Frame (networking); Philosophy","score_opus":0.01899950790440063,"score_gpt":0.28098725840908806,"score_spread":0.2619877505046874,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2508414753","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.0007145592,0.0009002402,0.9921208,0.0018546077,0.0017669902,0.00008329202,0.0000021598285,0.00028064076,0.0022767333],"genre_scores_gemma":[0.9147161,0.000032410862,0.0836448,0.00016925969,0.00038215946,0.000010795275,0.000001192008,0.0000069782313,0.0010363059],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99924284,0.00007972652,0.00014974944,0.00018926794,0.00013184016,0.00020657938],"domain_scores_gemma":[0.99909323,0.00036301577,0.000064628766,0.00032265307,0.00012616577,0.000030307981],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00051269215,0.00008269951,0.000067105444,0.000052712414,0.0001276618,0.000095276,0.00050247635,0.00006411251,0.0000053120007],"category_scores_gemma":[0.0027413734,0.000042504224,0.000025054078,0.00013287188,0.0001674912,0.00006770153,0.0001445066,0.0001267999,0.000020486132],"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":[0.000010881009,0.000011700801,0.0005469236,0.0000017909783,0.0000022592815,0.000007897315,0.00005124379,3.8680554e-7,0.0004635916,0.950628,0.0007089887,0.047566384],"study_design_scores_gemma":[0.00012682608,0.00002911262,0.00064385665,0.000032690594,0.0000028933014,0.000006750748,0.000019707357,0.0003697901,0.0018188877,0.986796,0.010057913,0.00009554361],"about_ca_topic_score_codex":0.0000147006485,"about_ca_topic_score_gemma":0.000061491126,"teacher_disagreement_score":0.9140015,"about_ca_system_score_codex":0.000059468373,"about_ca_system_score_gemma":0.00006505896,"threshold_uncertainty_score":0.3281878},"labels":[],"label_agreement":null},{"id":"W2510591540","doi":"10.1515/ling-2016-0026","title":"Blackfoot causative formation between lexicon and grammar","year":2016,"lang":"en","type":"article","venue":"Linguistics","topic":"Syntax, Semantics, Linguistic Variation","field":"Arts and Humanities","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":"University of Lethbridge","funders":"","keywords":"Lexeme; Linguistics; Lexicon; Morpheme; Grammar; Lexical functional grammar; Lexical item; Verb; Lexical grammar; Causative; Computer science; Emergent grammar; Focus (optics); Argument (complex analysis); Psychology; Relational grammar; Philosophy","score_opus":0.042915326649801425,"score_gpt":0.2532778118619995,"score_spread":0.21036248521219808,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2510591540","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":[0.43444407,0.0002530855,0.0968378,0.033738133,0.020490834,0.0019443977,0.0014702819,0.0013778062,0.4094436],"genre_scores_gemma":[0.99051934,0.000017402155,0.0010227574,0.00007943013,0.006572896,0.0000055203463,0.0000335987,0.000023184824,0.0017258585],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99905974,0.00003958521,0.00031975124,0.00018516423,0.00017879637,0.0002169852],"domain_scores_gemma":[0.9987096,0.00039600563,0.00017716918,0.00017221538,0.00047248905,0.00007250774],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.000206267,0.00014854735,0.00018281951,0.00008374423,0.0002310073,0.0001284945,0.0000850829,0.00006096452,0.00016838142],"category_scores_gemma":[0.004433574,0.00011025164,0.000032152264,0.000019936819,0.00019690716,0.000059218644,0.000045163368,0.00007406246,0.00013644525],"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":[0.0000074872305,0.000012822066,0.0019602193,0.000034438148,0.000029025994,0.0000030068522,0.015995672,3.0370458e-7,0.000029403132,0.97747874,0.00029114904,0.004157751],"study_design_scores_gemma":[0.0007628372,0.0001500319,0.002702636,0.00016094849,0.00016564583,0.0000016591265,0.001884943,0.0002859332,0.001095581,0.82910407,0.16328183,0.00040385986],"about_ca_topic_score_codex":0.00030105963,"about_ca_topic_score_gemma":0.0005767979,"teacher_disagreement_score":0.5560753,"about_ca_system_score_codex":0.000056302655,"about_ca_system_score_gemma":0.000027996914,"threshold_uncertainty_score":0.53077227},"labels":[],"label_agreement":null},{"id":"W2517767936","doi":"10.1515/ling-2015-0017","title":"How multiple past tenses divide the labor: The case of South Baffin Inuktitut","year":2015,"lang":"en","type":"article","venue":"Linguistics","topic":"Syntax, Semantics, Linguistic Variation","field":"Arts and Humanities","cited_by":5,"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":"Indeterminacy (philosophy); Icelandic; Linguistics; History; Variety (cybernetics); Simple past; Computer science; Epistemology; Philosophy; Grammar; Artificial intelligence","score_opus":0.054288159743357337,"score_gpt":0.24624292770710746,"score_spread":0.1919547679637501,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2517767936","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":[0.8053532,0.001283855,0.00517024,0.063676186,0.045323163,0.0024320201,0.0026296417,0.0007008602,0.07343082],"genre_scores_gemma":[0.98743224,0.0000034035386,0.0004995244,0.00029190394,0.010158049,0.000011492279,0.000019513614,0.000030719577,0.0015531519],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.9989022,0.00008926385,0.00032188083,0.00018628115,0.00025395144,0.00024640744],"domain_scores_gemma":[0.9969462,0.0007162322,0.00029495076,0.0005445503,0.0014262983,0.000071737944],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0004886647,0.00018210312,0.0002152135,0.000052130377,0.0004190198,0.0003505743,0.00030766387,0.00005531801,0.00003123918],"category_scores_gemma":[0.023840943,0.000105164756,0.00007317843,0.000059868282,0.0004376694,0.000024157684,0.00011666699,0.00019785723,0.00003298119],"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":[0.000023239194,0.00006843128,0.0047937026,0.000040719788,0.000077717974,0.00015889015,0.13670738,0.00009397345,0.00000463173,0.8539853,0.0035350905,0.000510913],"study_design_scores_gemma":[0.0013865057,0.00020076644,0.0008702885,0.00011690938,0.0005754389,0.00008869412,0.13576753,0.007256349,0.00042020774,0.16021435,0.6924746,0.00062834285],"about_ca_topic_score_codex":0.0026613094,"about_ca_topic_score_gemma":0.007399713,"teacher_disagreement_score":0.69377095,"about_ca_system_score_codex":0.000030387759,"about_ca_system_score_gemma":0.000093407674,"threshold_uncertainty_score":0.9843817},"labels":[],"label_agreement":null},{"id":"W2546420900","doi":"10.1515/ling-2016-0003","title":"Against Isomorphism and the Maxim of Charity in child language acquisition: Implications for the validity of the TVJT methodology","year":2016,"lang":"en","type":"article","venue":"Linguistics","topic":"Language Development and Disorders","field":"Psychology","cited_by":1,"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":"McGill University","keywords":"Maxim; Linguistics; Interpretation (philosophy); Negation; Isomorphism (crystallography); Psychology; Morpheme; Computer science; Natural language processing; Philosophy; Epistemology; Chemistry","score_opus":0.07108231059101608,"score_gpt":0.353755263213052,"score_spread":0.28267295262203596,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2546420900","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":[0.8973665,0.0029139218,0.019775419,0.026148587,0.0034276901,0.0024225179,0.00064053066,0.000048222617,0.047256563],"genre_scores_gemma":[0.9980793,0.00007753136,0.00079471426,0.00060204486,0.00016316262,0.000056337052,0.000006931913,0.0000073837637,0.00021262214],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"observational","domain_scores_codex":[0.9992468,0.00022016092,0.0002307319,0.00012260808,0.00005429074,0.00012545877],"domain_scores_gemma":[0.99720585,0.002150331,0.00015074582,0.00038203245,0.00009765212,0.00001341475],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0008268232,0.00007210873,0.00015585155,0.000024893867,0.000096909505,0.0000037147859,0.0002535548,0.00006916423,0.00005927278],"category_scores_gemma":[0.0024036348,0.00002999719,0.000055871464,0.000101094745,0.00037488044,0.0000048886945,0.00007021404,0.00007375362,0.0000015050449],"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":[0.00034824078,0.00021355369,0.12845618,0.00008095908,0.00017979465,0.0000020416917,0.026369462,0.0000029430014,0.00086351426,0.81644726,0.0031006457,0.023935428],"study_design_scores_gemma":[0.0044823033,0.000054307755,0.9366469,0.000064511216,0.00013546784,0.000010239358,0.0027352045,0.000028809653,0.001510642,0.045193426,0.008961903,0.00017632367],"about_ca_topic_score_codex":0.000093249524,"about_ca_topic_score_gemma":0.00007078228,"teacher_disagreement_score":0.8081907,"about_ca_system_score_codex":0.000007919829,"about_ca_system_score_gemma":0.000021160538,"threshold_uncertainty_score":0.2877549},"labels":[],"label_agreement":null},{"id":"W2547732919","doi":"10.1515/ling-2015-0044","title":"Uniqueness and grammatical relations in Upper Necaxa Totonac","year":2016,"lang":"en","type":"article","venue":"Linguistics","topic":"Syntax, Semantics, Linguistic Variation","field":"Arts and Humanities","cited_by":5,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"Social Sciences and Humanities Research Council of Canada","keywords":"Uniqueness; Axiom; Linguistics; Object (grammar); Universal grammar; Grammar; Relation (database); Agreement; Mathematics; Computer science; Philosophy","score_opus":0.023817369128572078,"score_gpt":0.24477048544828303,"score_spread":0.22095311631971096,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2547732919","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":[0.5049006,0.00021423862,0.023948956,0.042050354,0.014699761,0.0012791685,0.00026904492,0.00065984816,0.41197807],"genre_scores_gemma":[0.99274296,0.000011196128,0.0025904041,0.00010748371,0.0019034579,0.000011004092,0.000008174767,0.000022522008,0.002602788],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99909174,0.000043465694,0.0003274599,0.00018784613,0.0001432006,0.00020625858],"domain_scores_gemma":[0.99888515,0.00051918515,0.00007673538,0.00018577353,0.00027059403,0.000062533734],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002062236,0.00012116644,0.00016194116,0.00010762578,0.00013619897,0.00008304834,0.00008191471,0.0000680229,0.0004196356],"category_scores_gemma":[0.007763824,0.00009169648,0.000026404257,0.000030798717,0.00017805959,0.000029630388,0.00004250398,0.00009292198,0.000102289625],"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":[0.0000060270268,0.000032513017,0.008618118,0.000016880345,0.000008176938,0.000008741291,0.005464229,0.0000017771526,0.000010953337,0.9849997,0.000104872895,0.00072801515],"study_design_scores_gemma":[0.0005171896,0.000039233248,0.007250985,0.00017032477,0.000041600302,0.0000023067957,0.0007573849,0.0011218109,0.0000607938,0.9127288,0.07704891,0.00026066438],"about_ca_topic_score_codex":0.00045889238,"about_ca_topic_score_gemma":0.0012684254,"teacher_disagreement_score":0.4878424,"about_ca_system_score_codex":0.00005423613,"about_ca_system_score_gemma":0.000043963286,"threshold_uncertainty_score":0.9294583},"labels":[],"label_agreement":null},{"id":"W2591714696","doi":"10.1515/ling-2020-0184","title":"Culminating and non-culminating accomplishments in Malagasy","year":2020,"lang":"en","type":"article","venue":"Linguistics","topic":"Syntax, Semantics, Linguistic Variation","field":"Arts and Humanities","cited_by":25,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Western University","funders":"Nederlandse Organisatie voor Wetenschappelijk Onderzoek","keywords":"Circumstantial evidence; Causation; Prefix; Linguistics; History; Epistemology; Philosophy","score_opus":0.034838147866294675,"score_gpt":0.25477925165475795,"score_spread":0.21994110378846327,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2591714696","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":[0.6574781,0.00014888354,0.0027956548,0.011796365,0.0072997655,0.0009274656,0.00011489788,0.00039354124,0.31904534],"genre_scores_gemma":[0.9913048,0.0000052519204,0.0037055947,0.000555268,0.0041804826,0.0000068654917,0.000022552957,0.000032526284,0.00018666795],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.99878055,0.00002900712,0.0004144531,0.00029551398,0.00020950417,0.0002709586],"domain_scores_gemma":[0.99913716,0.00020462494,0.00016868724,0.00012432694,0.00027058096,0.000094646035],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00019512356,0.00017379827,0.00023793585,0.000079151316,0.00020226403,0.0002436443,0.00014487928,0.000052584022,0.00013725994],"category_scores_gemma":[0.0081028845,0.00018752663,0.0000311885,0.00005752749,0.000088318164,0.000039650276,0.00012383085,0.00020996721,0.000047484296],"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":[0.000024057656,0.00011922974,0.02663457,0.00046347658,0.0000453181,0.00012860529,0.23528181,0.000101734695,0.000119725984,0.7309782,0.00038376902,0.0057195043],"study_design_scores_gemma":[0.006394964,0.0007918017,0.042364325,0.0018918262,0.00046312515,0.0000151672175,0.10326847,0.48315313,0.0014873276,0.24267621,0.113933265,0.0035603913],"about_ca_topic_score_codex":0.0003929023,"about_ca_topic_score_gemma":0.00022454152,"teacher_disagreement_score":0.488302,"about_ca_system_score_codex":0.00003763539,"about_ca_system_score_gemma":0.000023535264,"threshold_uncertainty_score":0.9700495},"labels":[],"label_agreement":null},{"id":"W2760562865","doi":"10.1515/ling-2017-0019","title":"The development of gender-specific patterns in the production of voiceless sibilant fricatives in Mandarin Chinese","year":2017,"lang":"en","type":"article","venue":"Linguistics","topic":"Phonetics and Phonology Research","field":"Psychology","cited_by":18,"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 Lethbridge","funders":"","keywords":"Mandarin Chinese; Variation (astronomy); Articulation (sociology); Psychology; Speech production; Linguistics; Repetition (rhetorical device); Phonological development; Phonology","score_opus":0.09231072968453163,"score_gpt":0.39954172846112157,"score_spread":0.30723099877658994,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2760562865","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":[0.991116,0.00024995647,0.00005793974,0.00012872131,0.00092103035,0.00020031916,0.000008498397,0.0000022193756,0.0073153116],"genre_scores_gemma":[0.9993247,0.00008634975,0.00020495398,0.0000041174244,0.00016111799,0.000026039435,0.0000036246618,0.0000062523754,0.00018284773],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.9990733,0.00012888329,0.00031994114,0.00015963144,0.00014232575,0.00017591451],"domain_scores_gemma":[0.99878234,0.00025361418,0.00018780277,0.0006449996,0.00011793482,0.000013296842],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00093489623,0.00007253736,0.00013355156,0.000058575475,0.0001581831,0.000018674356,0.0005759326,0.000054583506,0.000013905215],"category_scores_gemma":[0.0010975666,0.00004162268,0.00001850434,0.00007377471,0.00019581478,0.0000045405222,0.000091156435,0.00023448,0.000006300666],"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":[0.00015343248,0.00043281232,0.86375153,0.00005378439,0.000039176317,0.000014563773,0.12078203,0.000007696218,0.0010948724,0.006353607,0.00021597746,0.007100513],"study_design_scores_gemma":[0.00019377476,0.000024620105,0.9913353,0.000020069236,0.0000017932583,0.0000011297202,0.001383682,0.000014662648,0.00075575116,0.0027014976,0.003520494,0.0000472248],"about_ca_topic_score_codex":0.00016673861,"about_ca_topic_score_gemma":0.0016279471,"teacher_disagreement_score":0.12758377,"about_ca_system_score_codex":0.000015483134,"about_ca_system_score_gemma":0.000046905734,"threshold_uncertainty_score":0.16973236},"labels":[],"label_agreement":null},{"id":"W2782848721","doi":"10.1515/ling-2017-0032","title":"Reproducible research in linguistics: A position statement on data citation and attribution in our field","year":2017,"lang":"en","type":"article","venue":"Linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":172,"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","funders":"National Science Foundation","keywords":"Attribution; Linguistics; Statement (logic); Citation; Field (mathematics); Applied linguistics; Quantitative linguistics; Authorship attribution; Position statement; Clinical linguistics; Psychology; Computer science; Sociology; Philosophy; Social psychology; Library science; Medicine","score_opus":0.18682001548247734,"score_gpt":0.4697376448389886,"score_spread":0.28291762935651127,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2782848721","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.072596945,0.005318496,0.85105115,0.034781776,0.010439589,0.003919933,0.00022804995,0.0012835254,0.020380555],"genre_scores_gemma":[0.8526615,0.000049463528,0.14671613,0.00008664814,0.00040254334,0.000008036784,0.000033960165,0.000005531258,0.00003621926],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99849534,0.0000985521,0.0002567304,0.0005889689,0.0003185056,0.00024187357],"domain_scores_gemma":[0.9975211,0.00016312851,0.00012824498,0.0016716693,0.0004765799,0.00003928895],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0026794674,0.000085676314,0.00011107987,0.0002698767,0.00021776523,0.00041821162,0.0009966257,0.000082657105,3.951258e-7],"category_scores_gemma":[0.031020427,0.00008634339,0.000006569964,0.00018703472,0.000026060534,0.00009979183,0.00075107237,0.00040690374,0.0000043607374],"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":[0.00012687113,0.00035075934,0.016155306,0.00030906877,0.000009191141,0.0005429016,0.0015497718,0.000034168257,0.0005807601,0.91050375,0.0074639725,0.062373463],"study_design_scores_gemma":[0.0013203454,0.0009474818,0.014111425,0.002161491,0.000011549716,0.000007650941,0.0001779332,0.15675136,0.029967288,0.773992,0.019847866,0.00070360396],"about_ca_topic_score_codex":0.00059513556,"about_ca_topic_score_gemma":0.000227366,"teacher_disagreement_score":0.7800645,"about_ca_system_score_codex":0.00012423216,"about_ca_system_score_gemma":0.00008197606,"threshold_uncertainty_score":0.9771417},"labels":[],"label_agreement":null},{"id":"W2909345388","doi":"10.1515/ling-2018-0034","title":"The use of<i>any</i>with factive predicates","year":2019,"lang":"en","type":"article","venue":"Linguistics","topic":"Syntax, Semantics, Linguistic Variation","field":"Arts and Humanities","cited_by":2,"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 Laval","funders":"","keywords":"Negation; Emotive; Regret; License; Counterfactual thinking; Linguistics; Grammar; Polarity (international relations); Meaning (existential); Computer science; Psychology; Epistemology; Social psychology; Philosophy","score_opus":0.03779951894372826,"score_gpt":0.2280744792269931,"score_spread":0.19027496028326482,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2909345388","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":[0.5267379,0.00036310017,0.0058732256,0.008604778,0.0419576,0.0033337625,0.0010399704,0.0007956749,0.41129398],"genre_scores_gemma":[0.9924896,0.000013253624,0.0012805936,0.00007775071,0.0017323523,0.0000054484517,0.000019182784,0.00002588078,0.0043559535],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.99909496,0.000032148055,0.00025880802,0.00015869863,0.000258578,0.00019681075],"domain_scores_gemma":[0.99769145,0.00093293766,0.0002272914,0.0003380599,0.00077280204,0.000037460028],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001264887,0.00012500564,0.00016559105,0.000037737595,0.00020984915,0.00017703214,0.0001638669,0.000037478272,0.00016630244],"category_scores_gemma":[0.0038653314,0.000078765275,0.000042155505,0.000027809687,0.0002256059,0.0000266877,0.000040686587,0.0001265477,0.000108350694],"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":[0.000046031873,0.00003840299,0.0055011,0.00003633794,0.000088134766,0.0000023180166,0.011969824,0.000103984,0.000009111751,0.9815194,0.00045187704,0.0002334823],"study_design_scores_gemma":[0.00052940845,0.00030453884,0.0025436555,0.0001415662,0.00018694552,0.000001573524,0.0039205025,0.005231377,0.0005659333,0.06460366,0.9216372,0.00033364454],"about_ca_topic_score_codex":0.00066985947,"about_ca_topic_score_gemma":0.00047750605,"teacher_disagreement_score":0.9211853,"about_ca_system_score_codex":0.000020952484,"about_ca_system_score_gemma":0.00005530164,"threshold_uncertainty_score":0.4627442},"labels":[],"label_agreement":null},{"id":"W3004777944","doi":"10.1515/ling-2019-0039","title":"Emphatic consonants beyond Arabic: The emergence and proliferation of uvular-pharyngeal emphasis in Kumzari","year":2020,"lang":"en","type":"article","venue":"Linguistics","topic":"Phonetics and Phonology Research","field":"Psychology","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":"Carleton University","funders":"Uppsala Universitet; Universiteit Leiden; Sultan Qaboos University; Alexander von Humboldt-Stiftung","keywords":"Semitic languages; Emphasis (telecommunications); Linguistics; Arabic; Language contact; Phenomenon; History; Vocabulary; Phonology; Computer science; Philosophy","score_opus":0.053120270806025535,"score_gpt":0.3509634372871384,"score_spread":0.29784316648111286,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3004777944","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":[0.98836297,0.0021629792,0.0005155786,0.0020553854,0.0009124759,0.00046003357,0.00005958213,0.00001972345,0.005451278],"genre_scores_gemma":[0.99856484,0.00012022444,0.00040124994,0.0004968823,0.00020626576,0.000019248559,0.0000084295125,0.000011117711,0.0001717656],"study_design_codex":"observational","study_design_gemma":"observational","domain_scores_codex":[0.9990108,0.00014262141,0.00028871215,0.00021583788,0.00012765285,0.00021438986],"domain_scores_gemma":[0.9993483,0.00018338421,0.00007373294,0.00018685644,0.00013208631,0.00007563841],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002615255,0.000093762304,0.00018165697,0.00004336313,0.000059690108,0.000013392914,0.00018155815,0.0000915456,0.00020403876],"category_scores_gemma":[0.0013597691,0.000074479416,0.00002226316,0.00021232184,0.00019720731,0.000007789301,0.000067876375,0.00021286141,0.000031198357],"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":[0.0007050352,0.0008212471,0.807961,0.00032381705,0.00030243953,0.00028111628,0.10366704,0.0001061969,0.005366393,0.05226244,0.01318489,0.015018375],"study_design_scores_gemma":[0.004274201,0.0012283635,0.9212405,0.000069961316,0.00013836418,0.000034912067,0.002151276,0.02672805,0.004529012,0.016090237,0.022906104,0.0006090235],"about_ca_topic_score_codex":0.00010283564,"about_ca_topic_score_gemma":0.000103727754,"teacher_disagreement_score":0.11327949,"about_ca_system_score_codex":0.0000054380057,"about_ca_system_score_gemma":0.000055761455,"threshold_uncertainty_score":0.30371824},"labels":[],"label_agreement":null},{"id":"W3135677970","doi":"10.1515/ling-2021-0021","title":"Expectations in language processing and production: an introduction to the special issue","year":2021,"lang":"en","type":"article","venue":"Linguistics","topic":"Neurobiology of Language and Bilingualism","field":"Neuroscience","cited_by":17,"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":"Norges Forskningsråd; Eberhard Karls Universität Tübingen","keywords":"Linguistics; Production (economics); Computer science; Philosophy; Economics","score_opus":0.02026753532018514,"score_gpt":0.3109884979845905,"score_spread":0.29072096266440534,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3135677970","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":[0.9771716,0.00021625895,0.000034940924,0.015349098,0.004271309,0.00021486836,0.000007987479,0.00006905826,0.00266488],"genre_scores_gemma":[0.9448576,0.000010354193,0.001032192,0.0012232693,0.051233456,0.000008677157,0.000007969168,0.000010794607,0.0016157072],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9991854,0.00010030839,0.00012993469,0.00036698085,0.00008294635,0.00013446597],"domain_scores_gemma":[0.9995949,0.000048985432,0.000031652115,0.00020432744,0.000083328276,0.000036774203],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.000097768585,0.00007111359,0.00007682315,0.00004825367,0.00016206053,0.00005995409,0.00008339935,0.00003091072,0.000050613384],"category_scores_gemma":[0.0084419055,0.000056657158,0.000009909077,0.00030748048,0.00006501749,0.000029992963,0.000042361185,0.00014295337,0.000018197252],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00010205465,0.0005427606,0.0005938861,0.00010026183,0.000005300625,0.00322936,0.1768372,0.0005093675,0.6657806,0.00464223,0.022117041,0.12553996],"study_design_scores_gemma":[0.00026128645,0.000121074954,0.000407804,0.000025853931,0.000020538044,0.00077850616,0.017343799,0.00026420248,0.5020776,0.00080173195,0.47758764,0.00030998755],"about_ca_topic_score_codex":0.000006552442,"about_ca_topic_score_gemma":0.00023548094,"teacher_disagreement_score":0.4554706,"about_ca_system_score_codex":0.000011196095,"about_ca_system_score_gemma":0.000039614166,"threshold_uncertainty_score":0.9999104},"labels":[],"label_agreement":null},{"id":"W3153295875","doi":"10.1515/ling-2020-0110","title":"Spanish embedded question island effects revisited: an experimental study","year":2021,"lang":"en","type":"article","venue":"Linguistics","topic":"Neurobiology of Language and Bilingualism","field":"Neuroscience","cited_by":28,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"The Scarborough Hospital; University of Toronto","funders":"","keywords":"Presupposition; Verb; Focus (optics); Ask price; Constraint (computer-aided design); Bounding overwatch; Linguistics; Contrast (vision); Mathematics; Computer science; Artificial intelligence; Philosophy; Physics; Economics","score_opus":0.024465818251006115,"score_gpt":0.34027008011750454,"score_spread":0.3158042618664984,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3153295875","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":[0.9940211,0.00017091112,0.000032788434,0.00001154106,0.001629414,0.0002515899,0.000015007238,0.00014334728,0.0037242898],"genre_scores_gemma":[0.99674743,0.000009236966,0.00038813316,0.0011586678,0.0011793573,0.000008016103,0.000022249946,0.000022351904,0.00046454565],"study_design_codex":"bench_or_experimental","study_design_gemma":"bench_or_experimental","domain_scores_codex":[0.9983848,0.00046198556,0.00020699362,0.00053351594,0.0001778043,0.00023493828],"domain_scores_gemma":[0.9990567,0.00026366452,0.000070552815,0.0004102035,0.000107304346,0.00009156276],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00014795152,0.00016041024,0.00020392475,0.00003881011,0.00016309433,0.00008435322,0.00017022203,0.000074068266,0.000073329094],"category_scores_gemma":[0.006570629,0.0001511498,0.00004285913,0.00015465119,0.00006193657,0.000028778742,0.00009534759,0.0001948729,0.000050235954],"study_design_candidate":"bench_or_experimental","study_design_consensus":"bench_or_experimental","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000047828347,0.0022355532,0.0019037222,0.00004078123,0.000009009253,0.017073011,0.0041271425,0.000011788298,0.97142595,0.0023358746,0.00048857054,0.00030078783],"study_design_scores_gemma":[0.0007773626,0.0010208489,0.00021473426,0.00003442541,0.000040818035,0.0002513906,0.00068532664,0.0001490931,0.9930758,0.00039624923,0.0030901323,0.00026381543],"about_ca_topic_score_codex":0.000011259149,"about_ca_topic_score_gemma":0.000020848802,"teacher_disagreement_score":0.021649873,"about_ca_system_score_codex":0.000017892318,"about_ca_system_score_gemma":0.000038926362,"threshold_uncertainty_score":0.7866131},"labels":[],"label_agreement":null},{"id":"W4244767861","doi":"10.1515/ling-2013-0039","title":"“Sometimes I'll start a sentence in Spanish Y TERMINO EN ESPAÑOL”: Toward a typology of code-switching","year":2013,"lang":"en","type":"article","venue":"Linguistics","topic":"Spanish Linguistics and Language Studies","field":"Arts and Humanities","cited_by":41,"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 Ottawa","funders":"","keywords":"Typology; Code-switching; Linguistics; Sentence; Linguistic typology; Syntax; Code (set theory); Philosophy; Psychology; Humanities; History; Sociology; Computer science; Programming language; Anthropology","score_opus":0.02611073656453649,"score_gpt":0.2504161320522807,"score_spread":0.2243053954877442,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4244767861","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":[0.49753642,0.0016110542,0.00029702045,0.0004581168,0.0059341304,0.0008246271,0.0002357241,0.00020591752,0.492897],"genre_scores_gemma":[0.9940837,0.00007747397,0.0018012441,0.00026415574,0.0018415687,0.000021746951,0.000011560707,0.000029885576,0.0018686794],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.9986118,0.000052319807,0.0004977072,0.00026625558,0.00019151621,0.00038038802],"domain_scores_gemma":[0.9986321,0.00033477362,0.00017470903,0.00024878932,0.0005455873,0.00006404497],"candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.00019565751,0.0002128652,0.00042810538,0.00015765193,0.00011744037,0.00009534486,0.0002502689,0.000070724105,0.0010616444],"category_scores_gemma":[0.0039098044,0.000190142,0.0000770499,0.000056896388,0.00023254631,0.000025412914,0.00015178934,0.00023631887,0.000083444735],"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":[0.000039397994,0.00035467593,0.02204881,0.00037800003,0.00017087831,0.00017860136,0.16688089,0.000027477518,0.00030565122,0.8013371,0.005573051,0.002705417],"study_design_scores_gemma":[0.0022629045,0.0009698991,0.015342069,0.0007326024,0.00031982726,0.00001725237,0.13210174,0.0049598482,0.0010456302,0.06742758,0.77297175,0.0018488754],"about_ca_topic_score_codex":0.004548146,"about_ca_topic_score_gemma":0.0026188383,"teacher_disagreement_score":0.7673987,"about_ca_system_score_codex":0.000032560016,"about_ca_system_score_gemma":0.000052128038,"threshold_uncertainty_score":0.9998515},"labels":[],"label_agreement":null},{"id":"W4291001811","doi":"10.1515/ling-2020-0211","title":"St’át’imcets frustratives as not-at-issue modals","year":2022,"lang":"en","type":"article","venue":"Linguistics","topic":"Syntax, Semantics, Linguistic Variation","field":"Arts and Humanities","cited_by":2,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia","funders":"Leibniz-Gemeinschaft; Social Sciences and Humanities Research Council of Canada; University of British Columbia","keywords":"Linguistics; Proposition; Grammaticality; Philosophy; Causation; Evidentiality; Coreference; Grammar; Epistemology; Computer science; Artificial intelligence","score_opus":0.03677723792240093,"score_gpt":0.268292201424065,"score_spread":0.2315149635016641,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4291001811","genre_codex":"other","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.15135625,0.00023477749,0.00029813603,0.010566261,0.034832954,0.0010029009,0.001957148,0.0007738248,0.79897773],"genre_scores_gemma":[0.95785195,0.000007502912,0.00084386114,0.00091199315,0.009740094,0.000048416707,0.000189203,0.00005975837,0.030347245],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.9980461,0.00009550617,0.0004859953,0.00039500266,0.0005870948,0.00039029834],"domain_scores_gemma":[0.99837905,0.00035801952,0.00028267087,0.0004064121,0.00046279008,0.00011102683],"candidate_categories":["metaepi_narrow","sts","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.00035324108,0.00025604805,0.00030608548,0.00013005557,0.0013133371,0.00018016674,0.00034113292,0.00004480509,0.016709805],"category_scores_gemma":[0.005009024,0.00027918362,0.0000855382,0.000057632275,0.00017164114,0.00002474866,0.0002812347,0.00033292783,0.0006229419],"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":[0.000059610094,0.0001568237,0.00017820574,0.0000343135,0.00006457601,0.00006887585,0.033172134,0.00022472777,0.000055950564,0.95989895,0.0059360694,0.0001497787],"study_design_scores_gemma":[0.0004674663,0.00038363013,0.00013238925,0.00001864514,0.000104886596,0.000007895738,0.0070244865,0.0014719033,0.00048554593,0.255473,0.7339879,0.00044225223],"about_ca_topic_score_codex":0.0012183656,"about_ca_topic_score_gemma":0.00051393034,"teacher_disagreement_score":0.80649567,"about_ca_system_score_codex":0.00018125714,"about_ca_system_score_gemma":0.000115229115,"threshold_uncertainty_score":0.9999868},"labels":[],"label_agreement":null},{"id":"W4308014841","doi":"10.1515/ling-2020-0035","title":"Non-syntactic factors and accessibility to relativization: evidence from Armenian","year":2022,"lang":"en","type":"article","venue":"Linguistics","topic":"Syntax, Semantics, Linguistic Variation","field":"Arts and Humanities","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":"Genitive case; Linguistics; Hierarchy; Armenian; Dative case; Demonstrative; Oblique case; Noun phrase; Universal grammar; Grammar; Predicate (mathematical logic); Noun; Mathematics; Philosophy; Computer science; Political science","score_opus":0.056478637433292296,"score_gpt":0.282028371915199,"score_spread":0.22554973448190674,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4308014841","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":[0.97728634,0.000074154996,0.0021067562,0.002290897,0.008159646,0.00044006133,0.00025988396,0.00014522689,0.009237028],"genre_scores_gemma":[0.9949859,0.0000048783686,0.0014987497,0.00022638931,0.00253066,0.000023844776,0.000063115884,0.000028741526,0.00063775544],"study_design_codex":"observational","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99858934,0.000077318015,0.0003637877,0.00040345427,0.00035754227,0.00020856429],"domain_scores_gemma":[0.9982912,0.00074445,0.00018340278,0.0003557766,0.00031154946,0.00011364453],"candidate_categories":["metaresearch","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.00028706813,0.00017892252,0.00022084221,0.00009576668,0.00081189646,0.0002614103,0.00022578206,0.00003328238,0.0018273623],"category_scores_gemma":[0.009269554,0.00017953444,0.000038578608,0.00007800711,0.00007269557,0.000061678824,0.00026366982,0.0002296975,0.000035221103],"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":[0.000109102286,0.00022764246,0.4263323,0.00010304474,0.00012336277,0.00003694826,0.31351218,0.00036366825,0.000101499834,0.25750226,0.0010103565,0.0005776476],"study_design_scores_gemma":[0.0011562844,0.0008194025,0.36157373,0.00040418672,0.0007321271,0.0000035500045,0.04022027,0.017848464,0.000819671,0.43495268,0.13942268,0.0020469606],"about_ca_topic_score_codex":0.0032526292,"about_ca_topic_score_gemma":0.0010695481,"teacher_disagreement_score":0.2732919,"about_ca_system_score_codex":0.00014130973,"about_ca_system_score_gemma":0.00007133697,"threshold_uncertainty_score":0.9990851},"labels":[],"label_agreement":null},{"id":"W4388233341","doi":"10.1515/ling-2022-0017","title":"Beyond alienability: factors determining possessive classes in Piaroa","year":2023,"lang":"en","type":"article","venue":"Linguistics","topic":"Syntax, Semantics, Linguistic Variation","field":"Arts and Humanities","cited_by":2,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":true,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Alberta","funders":"University of British Columbia; Killam Trusts; Endangered Languages Documentation Programme","keywords":"Possessive; Possession (linguistics); Linguistics; Noun; Noun phrase; Genitive case; History; Philosophy","score_opus":0.0435401410852999,"score_gpt":0.28045475933366903,"score_spread":0.23691461824836912,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4388233341","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":[0.8962482,0.000017472297,0.000096805175,0.0009455825,0.011804106,0.00030627308,0.00012602778,0.0005019648,0.08995353],"genre_scores_gemma":[0.9938752,0.0000078441,0.00040671928,0.00012520306,0.0036722063,0.0000130547805,0.00012284,0.00004720986,0.0017296976],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99832785,0.00006228232,0.0005171795,0.00035553882,0.00029658843,0.0004405867],"domain_scores_gemma":[0.9981163,0.0008468381,0.00018551177,0.00032478478,0.00044692747,0.00007959948],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.00032006265,0.00022736333,0.00030590873,0.00030065796,0.00025505672,0.0002458184,0.00021765855,0.00009424853,0.0002372624],"category_scores_gemma":[0.01632513,0.00022292978,0.00007381975,0.00015187923,0.00017548919,0.00004261334,0.00010171977,0.00022223365,0.00018027231],"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":[0.000010365782,0.000087562345,0.05486498,0.00011133761,0.000035907302,0.00011912016,0.10114241,0.00014150092,0.000021552676,0.8422745,0.00063598517,0.00055481825],"study_design_scores_gemma":[0.0008509205,0.00017954106,0.033975888,0.00022968337,0.00014360811,0.0000023786456,0.054435793,0.011320287,0.00040431842,0.8262334,0.07120236,0.0010218122],"about_ca_topic_score_codex":0.0007528546,"about_ca_topic_score_gemma":0.0027340504,"teacher_disagreement_score":0.097626984,"about_ca_system_score_codex":0.000093994364,"about_ca_system_score_gemma":0.00009638411,"threshold_uncertainty_score":0.99196076},"labels":[],"label_agreement":null},{"id":"W4408226726","doi":"10.1515/ling-2025-0016","title":"“I haven’t seen any results yet”: on ethical collaborative research in linguistics and the need for a standard protocol","year":2025,"lang":"en","type":"article","venue":"Linguistics","topic":"Language, Discourse, Communication Strategies","field":"Arts and Humanities","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":"Leverhulme Trust; Trent University; Nottingham Trent University","keywords":"Linguistics; Applied linguistics; Protocol (science); Haven; Clinical linguistics; Psychology; Philosophy; Medicine","score_opus":0.10790292738503555,"score_gpt":0.454158373337925,"score_spread":0.3462554459528895,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4408226726","genre_codex":"other","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.00033788194,0.00012318756,0.000057850346,0.008681544,0.00081353425,0.03801413,0.00086648477,0.00008493936,0.9510204],"genre_scores_gemma":[0.9621122,0.00002608798,0.0017416314,0.0010130244,0.0017370342,0.024875248,0.000055615295,0.000034963716,0.008404192],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.9980769,0.0005341907,0.0004712899,0.00026111104,0.00034579285,0.0003107215],"domain_scores_gemma":[0.99175084,0.0051091886,0.00012100789,0.00058081036,0.0023914624,0.000046723446],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0026493834,0.00016055822,0.00028913008,0.00018291893,0.0006877292,0.00052156457,0.0003690148,0.000149744,0.000015821424],"category_scores_gemma":[0.035525396,0.00010950387,0.000038327613,0.00018258646,0.0014320214,0.000009253327,0.00016253091,0.000998582,0.000005265209],"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":[0.002448564,0.000058043093,0.00002484793,0.00011438898,0.000027643999,0.0000051250545,0.027046796,0.000019195126,5.6458236e-7,0.9518347,0.018259155,0.00016098925],"study_design_scores_gemma":[0.0038407207,0.00021404868,0.000023828487,0.00030239913,0.000019418732,8.875903e-8,0.023291713,0.0005580108,0.00004523045,0.16745752,0.8041296,0.00011743121],"about_ca_topic_score_codex":0.00024846,"about_ca_topic_score_gemma":0.0030163368,"teacher_disagreement_score":0.96177435,"about_ca_system_score_codex":0.00009528511,"about_ca_system_score_gemma":0.0005965844,"threshold_uncertainty_score":0.9725988},"labels":[],"label_agreement":null}]}