{"meta":{"query_hash":"a87db38be0a9","filters":{"venue":"Corpus Linguistics and Linguistic Theory"},"cohort_total":15,"direct_labels_cover":0,"predictions_cover":15,"exported":15,"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/a87db38be0a9","api":"https://metacan.xera.ac/api/v1/cohort?venue=Corpus+Linguistics+and+Linguistic+Theory"},"results":[{"id":"W2068619837","doi":"10.1515/cllt-2014-0011","title":"A multivariate analysis of the Old English ACC+DAT double object alternation","year":2014,"lang":"en","type":"article","venue":"Corpus Linguistics and Linguistic Theory","topic":"Syntax, Semantics, Linguistic Variation","field":"Arts and Humanities","cited_by":27,"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":"Alternation (linguistics); Linguistics; Object (grammar); Variation (astronomy); Verb; Dative case; Semantics (computer science); Psychology; Computer science; Mathematics; Philosophy","score_opus":0.017902671986126415,"score_gpt":0.24031874023093547,"score_spread":0.22241606824480906,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2068619837","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":[0.36144456,0.0006031181,0.019809581,0.0005543529,0.06986589,0.0017585111,0.00078355364,0.0006810895,0.54449934],"genre_scores_gemma":[0.9897997,0.000017108627,0.0004484601,0.00017473935,0.0074877287,0.000013381082,0.000065066364,0.000048062964,0.0019457841],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99779993,0.00022906602,0.00075192726,0.00047007736,0.0004061087,0.00034287304],"domain_scores_gemma":[0.9950235,0.0016000742,0.00068389793,0.00073869346,0.0018435429,0.00011031253],"candidate_categories":["metaresearch","metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0014658697,0.00031720704,0.0005859711,0.0003407486,0.00049156096,0.0002482207,0.0004459921,0.00010861543,0.00021855443],"category_scores_gemma":[0.03689745,0.00024860722,0.00024694344,0.00022136782,0.00043237887,0.000017643095,0.00019303296,0.00026447573,0.000009170578],"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":[0.00009563101,0.00008514178,0.0021508199,0.00006717261,0.0005304982,0.000002132821,0.033918783,0.00020325287,0.000025701802,0.96229285,0.000016567688,0.00061146467],"study_design_scores_gemma":[0.0024394968,0.00018403017,0.008580056,0.00028292031,0.00717499,0.0000017021877,0.0022483056,0.05641645,0.0006418206,0.81721085,0.1038828,0.0009365882],"about_ca_topic_score_codex":0.0019055568,"about_ca_topic_score_gemma":0.0010198214,"teacher_disagreement_score":0.6283551,"about_ca_system_score_codex":0.00005424589,"about_ca_system_score_gemma":0.00007075665,"threshold_uncertainty_score":0.9999966},"labels":[],"label_agreement":null},{"id":"W2315347275","doi":"10.1515/cllt-2014-0020","title":"The meaning of intonation in yes-no questions in American English: A corpus study","year":2014,"lang":"en","type":"article","venue":"Corpus Linguistics and Linguistic Theory","topic":"Language, Discourse, Communication Strategies","field":"Arts and Humanities","cited_by":58,"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":"Intonation (linguistics); Linguistics; Pitch accent; Utterance; Meaning (existential); Interrogative; Context (archaeology); Sentence; Pitch contour; American English; Stress (linguistics); Phrase; Word order; Psychology; Prosody; History; Philosophy","score_opus":0.017252652359156854,"score_gpt":0.2736270630687182,"score_spread":0.25637441070956135,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2315347275","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":[0.31527647,0.0014540424,0.00020995463,0.000037630536,0.0031142808,0.0006801352,0.00003455713,0.00012617785,0.6790667],"genre_scores_gemma":[0.9979078,0.000099293575,0.00019428054,0.00003572127,0.0011059066,0.000034334702,0.000010478682,0.000024775754,0.0005873757],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9983208,0.0004892095,0.0005564191,0.00022127481,0.00018722199,0.00022506139],"domain_scores_gemma":[0.99576,0.0025646994,0.0003250327,0.00048151365,0.00081826345,0.0000504855],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0017297588,0.00017414163,0.00030354966,0.00017525448,0.0002899662,0.00017873313,0.00033256895,0.000030254398,0.00002360357],"category_scores_gemma":[0.031193145,0.00013600709,0.000035211488,0.00010989929,0.0008386102,0.000016747548,0.000114868264,0.00030295656,0.000004946495],"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":[0.000046830526,0.00018745758,0.003504359,0.0000187547,0.000016283646,0.00000457029,0.07082974,0.00003112683,0.0000036972506,0.92097974,0.000011477013,0.004365951],"study_design_scores_gemma":[0.0021003245,0.00087591336,0.011045865,0.0004380518,0.00014877018,0.0000019571714,0.12012383,0.0042058015,0.000019539966,0.6420239,0.21835692,0.00065910275],"about_ca_topic_score_codex":0.0016383094,"about_ca_topic_score_gemma":0.009936289,"teacher_disagreement_score":0.6826314,"about_ca_system_score_codex":0.0000436775,"about_ca_system_score_gemma":0.000085449785,"threshold_uncertainty_score":0.9769675},"labels":[],"label_agreement":null},{"id":"W2327588828","doi":"10.1515/cllt.2011.008","title":"Safe harbour: Ethics and accessibility in sociolinguistic corpus building","year":2011,"lang":"en","type":"article","venue":"Corpus Linguistics and Linguistic Theory","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":15,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Memorial University of Newfoundland","funders":"","keywords":"Globe; Corpus linguistics; Linguistics; Computer science; Applied linguistics; Computational linguistics; Sociology; Data sharing; Natural language processing; Psychology; Philosophy","score_opus":0.03852763765365219,"score_gpt":0.3088898463052056,"score_spread":0.2703622086515534,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2327588828","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":[0.044021595,0.044655357,0.8581573,0.00012463683,0.0073999986,0.0010667159,0.0000475862,0.0018421097,0.042684726],"genre_scores_gemma":[0.7360911,0.00021852143,0.26275355,0.00033256423,0.00048399402,0.0000125957085,0.0000024148208,0.000030781557,0.000074511474],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9970224,0.00040964893,0.00067969377,0.0009617337,0.0003228485,0.00060370704],"domain_scores_gemma":[0.99551165,0.0024004264,0.0003424094,0.00078276417,0.00068333326,0.00027944383],"candidate_categories":["metaresearch","metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0037229352,0.00041460517,0.0005079851,0.00028707724,0.00044012244,0.0002769927,0.0010025648,0.00037609204,0.000013133147],"category_scores_gemma":[0.06953516,0.00039233264,0.00006177234,0.00027595932,0.0005526629,0.000052002593,0.0008830013,0.0015447594,0.000003557308],"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":[0.000052369047,0.000073200725,0.0039339103,0.00020318506,0.0000146085395,0.00025859344,0.007835351,3.8387535e-7,0.00011730629,0.9769413,0.000006962819,0.0105628595],"study_design_scores_gemma":[0.00040298045,0.00010597853,0.0026986045,0.00032774423,0.000042455555,0.000038470997,0.000055966546,0.003581907,0.0011588564,0.9887118,0.0023308161,0.00054441753],"about_ca_topic_score_codex":0.00021004448,"about_ca_topic_score_gemma":0.0000356249,"teacher_disagreement_score":0.6920695,"about_ca_system_score_codex":0.00009580346,"about_ca_system_score_gemma":0.00029971614,"threshold_uncertainty_score":0.99985284},"labels":[],"label_agreement":null},{"id":"W2332964166","doi":"10.1515/cllt-2013-0014","title":"Topic marking in a Shanghainese corpus: from observation to prediction","year":2013,"lang":"en","type":"article","venue":"Corpus Linguistics and Linguistic Theory","topic":"Linguistic Variation and Morphology","field":"Social Sciences","cited_by":14,"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":"Polytomous Rasch model; Computer science; Natural language processing; Logistic regression; Artificial intelligence; Multivariate statistics; Selection (genetic algorithm); Computational linguistics; Machine learning; Statistics; Item response theory; Mathematics; Psychometrics","score_opus":0.022892869246276686,"score_gpt":0.2744754931635357,"score_spread":0.251582623917259,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2332964166","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":[0.73986334,0.0012889964,0.017223805,0.0010066546,0.022919947,0.0022026186,0.00016745334,0.0005156207,0.21481155],"genre_scores_gemma":[0.9879003,0.00008918421,0.0042617978,0.0011125853,0.004843429,0.00007000991,0.000052372125,0.000025288397,0.001645015],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9981308,0.0002864179,0.00047817442,0.00043041274,0.00026032917,0.0004138913],"domain_scores_gemma":[0.99746686,0.0012127499,0.00014961707,0.0002571269,0.00063869864,0.00027496522],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.001063074,0.0001877822,0.0002728918,0.00019079776,0.00032482244,0.00015874124,0.00022160776,0.00019086966,0.00056588685],"category_scores_gemma":[0.058119196,0.00019688753,0.000041013616,0.00029049162,0.00015605496,0.000020505699,0.00009209584,0.00021615546,0.00008095481],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000045206674,0.00006711382,0.061433833,0.000016481887,0.000017681252,0.000031389438,0.015682552,0.00001597955,0.0001352327,0.9117686,0.00039308288,0.010392859],"study_design_scores_gemma":[0.00075034885,0.000086533335,0.12679298,0.00014487786,0.000054050095,0.0000012699254,0.0016722231,0.0015152284,0.000019847961,0.6081796,0.26040676,0.0003762263],"about_ca_topic_score_codex":0.013068329,"about_ca_topic_score_gemma":0.0021404857,"teacher_disagreement_score":0.30358896,"about_ca_system_score_codex":0.00014747841,"about_ca_system_score_gemma":0.00018602805,"threshold_uncertainty_score":0.99350375},"labels":[],"label_agreement":null},{"id":"W2563400189","doi":"10.1515/cllt-2015-0047","title":"Mass counts in World Englishes: A corpus linguistic study of noun countability in non-native varieties of English","year":2016,"lang":"en","type":"article","venue":"Corpus Linguistics and Linguistic Theory","topic":"Linguistic Variation and Morphology","field":"Social Sciences","cited_by":12,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"McMaster University","funders":"York St John University","keywords":"Linguistics; Noun; Syntax; Nominalization; World Englishes; Corpus linguistics; Applied linguistics; Varieties of English; Natural language processing; Computer science; Philosophy","score_opus":0.016003089102713267,"score_gpt":0.2859726892322059,"score_spread":0.26996960012949267,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2563400189","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":[0.24613297,0.0006722658,0.001478543,0.000021667505,0.01871729,0.0016843955,0.00022572704,0.00011855404,0.73094857],"genre_scores_gemma":[0.9967846,0.000102435246,0.0003097388,0.000033754346,0.00160269,0.000025870131,0.0000046728655,0.000026572852,0.0011096674],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99726987,0.0005095105,0.0009360815,0.00045389356,0.00042450888,0.00040613467],"domain_scores_gemma":[0.99231815,0.0041123815,0.00046678563,0.0003812392,0.0025856155,0.000135803],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0030821713,0.00024012056,0.000641846,0.00040704507,0.00011398487,0.000028066404,0.0003301833,0.000168461,0.00018890822],"category_scores_gemma":[0.17711668,0.00020915241,0.00005707408,0.00048670283,0.0008133995,0.000013063322,0.00008837087,0.00026644548,0.0000050180947],"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":[0.0002191852,0.00054333435,0.07765221,0.00008059808,0.00003459252,0.00004764464,0.08179921,0.000006060511,0.000016967813,0.8389749,0.00003341221,0.00059189583],"study_design_scores_gemma":[0.0067717815,0.0008263008,0.03380688,0.0007942454,0.0002617061,9.637841e-7,0.020491535,0.00021693126,0.00009361434,0.78134596,0.15435724,0.0010328208],"about_ca_topic_score_codex":0.0019336744,"about_ca_topic_score_gemma":0.0037230195,"teacher_disagreement_score":0.7506516,"about_ca_system_score_codex":0.00018657873,"about_ca_system_score_gemma":0.0004940558,"threshold_uncertainty_score":0.8528987},"labels":[],"label_agreement":null},{"id":"W2756059139","doi":"10.1515/cllt-2016-0052","title":"Common ground across globalized English varieties: A multivariate exploration of mental predicates in World Englishes","year":2017,"lang":"en","type":"article","venue":"Corpus Linguistics and Linguistic Theory","topic":"Linguistic Variation and Morphology","field":"Social Sciences","cited_by":28,"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":"World Englishes; Irish; Focus (optics); Linguistics; Common ground; Varieties of English; Multivariate statistics; American English; Sociology; Computer science; Philosophy","score_opus":0.036931674976311626,"score_gpt":0.34753770573814985,"score_spread":0.31060603076183824,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2756059139","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":[0.23114976,0.0022519052,0.0053255144,0.0006012532,0.089035474,0.00393127,0.009329479,0.00078730227,0.657588],"genre_scores_gemma":[0.995694,0.00016832429,0.000706234,0.00007881143,0.0024025876,0.000025266982,0.00019514823,0.000020734267,0.0007089355],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9977284,0.00040234678,0.00066055555,0.00040310368,0.0003333194,0.00047229268],"domain_scores_gemma":[0.9965411,0.0012141896,0.00066542206,0.0005131936,0.00087373843,0.00019238156],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.002800921,0.00023100224,0.00047989946,0.00011848166,0.0012024912,0.00032383055,0.0005566476,0.00019228293,0.00007718678],"category_scores_gemma":[0.113900304,0.0002390541,0.00007352573,0.0001337421,0.0013031227,0.000064814776,0.00024567894,0.00026692476,0.0000032024848],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00024264514,0.00016183221,0.010143484,0.00004473336,0.000037351798,0.0000315287,0.085278034,0.000006505955,0.000013631404,0.9029628,0.0002436196,0.000833829],"study_design_scores_gemma":[0.002924255,0.00011717601,0.010378658,0.00021140042,0.000111146066,0.0000010712806,0.009270851,0.00035764676,0.000079103476,0.5092325,0.466806,0.000510196],"about_ca_topic_score_codex":0.008438894,"about_ca_topic_score_gemma":0.0074272407,"teacher_disagreement_score":0.7645442,"about_ca_system_score_codex":0.00010501662,"about_ca_system_score_gemma":0.0002241139,"threshold_uncertainty_score":0.998164},"labels":[],"label_agreement":null},{"id":"W2803450908","doi":"10.1515/cllt-2016-0078","title":"Unifying dimensions in coherence relations: How various annotation frameworks are related","year":2018,"lang":"en","type":"article","venue":"Corpus Linguistics and Linguistic Theory","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":71,"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":"Universiteit Utrecht; Schweizerischer Nationalfonds zur Förderung der Wissenschaftlichen Forschung; Deutsche Forschungsgemeinschaft","keywords":"Treebank; Computer science; Annotation; Coherence (philosophical gambling strategy); Dimension (graph theory); Set (abstract data type); Representation (politics); Rhetorical question; Natural language processing; Artificial intelligence; Interface (matter); Linguistics; Programming language; Mathematics","score_opus":0.010668414240520572,"score_gpt":0.2592416203831324,"score_spread":0.24857320614261186,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2803450908","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":[0.008714328,0.011159933,0.965491,0.00035751314,0.003750036,0.00047983942,0.000012494816,0.001118884,0.008915956],"genre_scores_gemma":[0.80431324,0.000027005028,0.19469133,0.00020882722,0.00035545471,0.0000100100615,0.000006150519,0.00001920059,0.00036881035],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99842966,0.00016346796,0.00035144718,0.0005205051,0.00020678657,0.00032812788],"domain_scores_gemma":[0.99719214,0.0010275244,0.00031114375,0.00052975654,0.00082588464,0.000113571565],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.000812283,0.0002211942,0.00024574605,0.00025465252,0.00035034344,0.00027426108,0.00048698197,0.00034365896,0.000008424689],"category_scores_gemma":[0.03281824,0.00021242186,0.000040288996,0.00061382423,0.00028346764,0.000051300645,0.00026385143,0.00084088615,0.000012289236],"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":[0.000013767528,0.00003648168,0.0014936046,0.000024746703,0.000011525315,0.00011616967,0.0023329242,0.000009624879,0.00008570567,0.98956466,0.00009742273,0.0062133675],"study_design_scores_gemma":[0.00023647935,0.00009742706,0.0008267665,0.00053763285,0.000026856997,0.000022371469,0.0000741176,0.023583265,0.00025008593,0.97097385,0.003044203,0.00032692682],"about_ca_topic_score_codex":0.00003296509,"about_ca_topic_score_gemma":0.000027315911,"teacher_disagreement_score":0.79559886,"about_ca_system_score_codex":0.00007476048,"about_ca_system_score_gemma":0.000121925754,"threshold_uncertainty_score":0.97532874},"labels":[],"label_agreement":null},{"id":"W2807690984","doi":"10.1515/cllt-2017-0031","title":"Dependency profiles in the large-scale analysis of discourse connectives","year":2018,"lang":"en","type":"article","venue":"Corpus Linguistics and Linguistic Theory","topic":"Natural Language Processing Techniques","field":"Computer Science","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":"McMaster University; Brock University","funders":"Turun Yliopisto","keywords":"Dependency (UML); Computer science; Syntax; Natural language processing; Scope (computer science); Artificial intelligence; Linguistics; Cluster analysis; Focus (optics); Annotation","score_opus":0.00909766402107549,"score_gpt":0.2919299757573142,"score_spread":0.28283231173623874,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2807690984","genre_codex":"methods","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":[0.05592057,0.0074711586,0.91581273,0.00009160334,0.0013735266,0.00046048258,0.00008027069,0.00027143676,0.018518224],"genre_scores_gemma":[0.94203883,0.000024967088,0.057288222,0.00013370119,0.0004336173,0.000010438237,0.000006283097,0.000009197206,0.000054724806],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9985777,0.00019646551,0.00033875683,0.00037711902,0.00023804564,0.00027192463],"domain_scores_gemma":[0.9977282,0.000931224,0.00020608341,0.0005922937,0.00049165485,0.000050532715],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.001492473,0.00016606635,0.00031752398,0.00033050528,0.00016083957,0.00009942574,0.00089115,0.00007872504,0.000010913736],"category_scores_gemma":[0.009501935,0.00011391188,0.000075803226,0.00090347923,0.0003694422,0.000021242113,0.00023729136,0.00021392247,0.0000015290158],"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":[0.000019316856,0.000085678,0.0030781813,0.000026434282,0.000059548674,0.000026805998,0.009592428,0.0000014618496,0.00006619215,0.9840117,0.00001252895,0.0030197124],"study_design_scores_gemma":[0.0002634989,0.0002042318,0.0031949037,0.00011288526,0.00039010204,0.0000076438,0.000977191,0.02266317,0.0020303065,0.9684435,0.0013996599,0.00031287613],"about_ca_topic_score_codex":0.000046827117,"about_ca_topic_score_gemma":0.00013264139,"teacher_disagreement_score":0.8861183,"about_ca_system_score_codex":0.000016669388,"about_ca_system_score_gemma":0.0000791128,"threshold_uncertainty_score":0.99884146},"labels":[],"label_agreement":null},{"id":"W2898954967","doi":"10.1515/cllt-2018-0033","title":"An information-theoretic view on language complexity and register variation: Compressing naturalistic corpus data","year":2018,"lang":"en","type":"article","venue":"Corpus Linguistics and Linguistic Theory","topic":"Natural Language Processing Techniques","field":"Computer Science","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":"Simon Fraser University","funders":"","keywords":"Computer science; Variation (astronomy); Register (sociolinguistics); Linguistics; Formality; Context (archaeology); Conversation; Sentence; Natural language processing; Artificial intelligence","score_opus":0.030986459640575396,"score_gpt":0.31101361548343714,"score_spread":0.28002715584286175,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2898954967","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":[0.0034265597,0.008459058,0.9663231,0.00019256474,0.0040946584,0.00060302217,0.00026084523,0.0012969581,0.015343255],"genre_scores_gemma":[0.8274653,0.00004899852,0.16976048,0.0012655823,0.0012520627,0.000004819646,0.00013145032,0.000021349815,0.000049949787],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.99793,0.00023714936,0.0004901536,0.00064820034,0.00033484065,0.00035966194],"domain_scores_gemma":[0.99623084,0.00078688323,0.0003598877,0.0017148865,0.0006857817,0.00022173587],"candidate_categories":["metaresearch","metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0012161508,0.00033119592,0.0003516678,0.00017665401,0.0005610811,0.0009320101,0.0014505874,0.00014470726,0.00001722256],"category_scores_gemma":[0.010997278,0.0002868888,0.000025157582,0.00022892658,0.0008169664,0.00021376743,0.00075995596,0.0003985438,0.000015252129],"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":[0.000051023195,0.00003556044,0.000032303033,0.00013251041,0.000014553596,0.0000342098,0.002961298,6.5006066e-7,0.000056281468,0.9707473,0.00010333656,0.025830986],"study_design_scores_gemma":[0.00055254105,0.00041509882,0.00043737615,0.0005123899,0.00009403037,0.000106591506,0.00007372833,0.09118986,0.00033194083,0.87885875,0.026729297,0.0006984137],"about_ca_topic_score_codex":0.000077935285,"about_ca_topic_score_gemma":0.000013129709,"teacher_disagreement_score":0.82403874,"about_ca_system_score_codex":0.00004191221,"about_ca_system_score_gemma":0.00010314846,"threshold_uncertainty_score":0.99995834},"labels":[],"label_agreement":null},{"id":"W3161046106","doi":"10.1515/cllt-2020-0028","title":"Switch-reference and its role in referential choice in Mbyá Guaraní narratives","year":2021,"lang":"en","type":"article","venue":"Corpus Linguistics and Linguistic Theory","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 Toronto","funders":"Social Sciences and Humanities Research Council of Canada; Connaught Fund","keywords":"Ambiguity; Self-reference; Reference model; Computer science; Function (biology); Narrative; Linguistics; Philosophy","score_opus":0.027123376247697555,"score_gpt":0.2525062085082269,"score_spread":0.22538283226052935,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3161046106","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":[0.8095641,0.0076768044,0.00022706205,0.00054608233,0.007634541,0.0005267218,0.00018096594,0.0001529652,0.17349075],"genre_scores_gemma":[0.99399245,0.00021570174,0.00033093547,0.00013622858,0.0034788747,0.000019232892,0.00006539088,0.000047960268,0.0017132083],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9976746,0.00022084468,0.00071962056,0.0006627531,0.00026359988,0.0004585981],"domain_scores_gemma":[0.99704146,0.0014003626,0.00022960249,0.00031502717,0.0008743744,0.00013914604],"candidate_categories":["metaresearch","metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0007624318,0.0003523191,0.0005210596,0.00025020863,0.0002712749,0.00029198846,0.00018509777,0.00016740871,0.00033932112],"category_scores_gemma":[0.030489882,0.00036490365,0.000044293018,0.000105405335,0.00023367611,0.000026791606,0.0001754258,0.0005357156,0.000020111444],"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":[0.00008988033,0.00013520752,0.0022867378,0.00015109986,0.000028382281,0.00013222423,0.039458904,0.000007633697,0.00015039077,0.95681816,0.000005271932,0.0007361104],"study_design_scores_gemma":[0.0013297678,0.00010896485,0.0045764544,0.00050777494,0.000079826816,0.000011232771,0.008520356,0.0014956659,0.0002152327,0.94940466,0.033144742,0.0006052977],"about_ca_topic_score_codex":0.00087566697,"about_ca_topic_score_gemma":0.012296161,"teacher_disagreement_score":0.18442835,"about_ca_system_score_codex":0.00006323929,"about_ca_system_score_gemma":0.00021200805,"threshold_uncertainty_score":0.9998803},"labels":[],"label_agreement":null},{"id":"W4241042034","doi":"10.1515/cllt-2012-masthead1","title":"Masthead","year":2012,"lang":"en","type":"paratext","venue":"Corpus Linguistics and Linguistic Theory","topic":"Marriage and Sexual Relationships","field":"Social Sciences","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 Alberta","funders":"","keywords":"Computer science","score_opus":0.04361840048096364,"score_gpt":0.33183670909712976,"score_spread":0.2882183086161661,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4241042034","genre_codex":"other","genre_gemma":"other","domain_codex":null,"domain_gemma":null,"model_version":"codex-gemma-dda1882f352a","genre_candidate":"other","genre_consensus":"other","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.000049921666,0.019269206,0.0010959401,0.00003225725,0.026837286,0.000313911,0.00017034414,0.00007559519,0.95215553],"genre_scores_gemma":[0.32258376,0.0029977115,0.00048124022,0.00024418804,0.036048792,0.000023901477,0.00017461691,0.000087903616,0.6373579],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.99720174,0.00064342533,0.00049702,0.00045717077,0.0004521292,0.00074853847],"domain_scores_gemma":[0.9949797,0.0031095566,0.0003350515,0.00043101274,0.0006182246,0.00052646233],"candidate_categories":["metaresearch","metaepi_narrow","insufficient_payload"],"consensus_categories":["insufficient_payload"],"category_scores_codex":[0.0023686518,0.00039422212,0.00053674064,0.00018678023,0.0011733292,0.00022432914,0.0003929035,0.0007217611,0.006348023],"category_scores_gemma":[0.05743038,0.00039638707,0.00011383575,0.00017628526,0.000715295,0.000010580087,0.00013685417,0.0008585393,0.0034568035],"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":[0.000031629505,0.000047292026,0.000056809975,0.00011257783,0.000051254458,0.000017590684,0.009722944,0.0000014149261,0.0000010139029,0.9467673,0.039016366,0.004173809],"study_design_scores_gemma":[0.000111882466,0.000031743446,0.000016242519,0.0001748504,0.00018221299,0.000001621038,0.00065460324,0.0000070605834,0.000001651749,0.07283348,0.9255341,0.00045055282],"about_ca_topic_score_codex":0.0005124739,"about_ca_topic_score_gemma":0.00016378085,"teacher_disagreement_score":0.88651776,"about_ca_system_score_codex":0.00012674941,"about_ca_system_score_gemma":0.0007262166,"threshold_uncertainty_score":0.9998488},"labels":[],"label_agreement":null},{"id":"W4388046533","doi":"10.1515/cllt-2021-0058","title":"Truth be told: a corpus-based study of the cross-linguistic colexification of representational and (inter)subjective meanings","year":2023,"lang":"en","type":"article","venue":"Corpus Linguistics and Linguistic Theory","topic":"Language, Metaphor, and Cognition","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":"York University; University of Toronto","funders":"Jackman Humanities Institute, University of Toronto; Canadian Network for Research and Innovation in Machining Technology, Natural Sciences and Engineering Research Council of Canada","keywords":"Linguistics; Focus (optics); Meaning (existential); Corpus linguistics; Variation (astronomy); Computer science; Epistemology; Philosophy","score_opus":0.03143056356154125,"score_gpt":0.3366006567825191,"score_spread":0.3051700932209779,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4388046533","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":[0.9756017,0.00049823173,0.00095121557,0.000011151358,0.004121173,0.00075963285,0.00024019189,0.00009420639,0.01772252],"genre_scores_gemma":[0.99805564,0.000012642693,0.0000990002,0.00008097409,0.0006923586,0.00006501598,0.000055663775,0.000044164935,0.00089454296],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"observational","domain_scores_codex":[0.9976723,0.0003998211,0.0007243867,0.0005604957,0.00036292203,0.00028006925],"domain_scores_gemma":[0.9948603,0.0027006816,0.0006176049,0.0005717909,0.0011514087,0.00009822356],"candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0011643257,0.0002561544,0.00042713401,0.00022528255,0.00025159854,0.0000503413,0.00026397215,0.00012669845,0.00008574088],"category_scores_gemma":[0.027343202,0.00021147038,0.000099816854,0.00045154034,0.0005283419,0.000007271792,0.00012396248,0.00025298801,0.0000061924256],"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":[0.0016137332,0.00093139615,0.062336855,0.00027811193,0.00036406767,0.00008661116,0.046149194,0.00006170715,0.0004516643,0.88503647,0.00008124034,0.00260895],"study_design_scores_gemma":[0.009945262,0.0023950436,0.5203324,0.00059012236,0.0018766515,0.000037239137,0.024900299,0.0038476626,0.0027585991,0.42842528,0.003757844,0.001133606],"about_ca_topic_score_codex":0.0003029087,"about_ca_topic_score_gemma":0.00004617879,"teacher_disagreement_score":0.45799553,"about_ca_system_score_codex":0.000028221823,"about_ca_system_score_gemma":0.0001110375,"threshold_uncertainty_score":0.9808499},"labels":[],"label_agreement":null},{"id":"W4394842830","doi":"10.1515/cllt-2023-0028","title":"The distributional properties of long nominal compounds in scientific articles: an investigation based on the uniform information density hypothesis","year":2024,"lang":"en","type":"article","venue":"Corpus Linguistics and Linguistic Theory","topic":"Advanced Text Analysis Techniques","field":"Computer Science","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 Alberta","funders":"","keywords":"Computer science; Constant (computer programming); Scientific literature; Information transmission; Econometrics; Data science; Information retrieval; Mathematics; Biology","score_opus":0.021486894088511916,"score_gpt":0.23154092932942288,"score_spread":0.21005403524091096,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4394842830","genre_codex":"methods","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":[0.48873302,0.00043243347,0.5076927,0.0003275823,0.0011365707,0.00041405193,0.00002729543,0.00023556364,0.0010007999],"genre_scores_gemma":[0.9964358,0.00000678408,0.003307343,0.000079806116,0.00011338922,0.000014947865,0.000011655321,0.000006215166,0.00002405128],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.9987395,0.00015087082,0.0003918876,0.00022333224,0.00030863256,0.00018577767],"domain_scores_gemma":[0.99747676,0.0013843717,0.00013530884,0.00047671338,0.00046703356,0.000059808197],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0021588833,0.00013372072,0.00013222093,0.0001701995,0.0004678295,0.00062748365,0.0004570065,0.000043611966,0.0000011914287],"category_scores_gemma":[0.008099127,0.0000808144,0.00004421234,0.00047563235,0.0006852428,0.000087896355,0.00010135602,0.00019088075,0.0000038470357],"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":[0.000030082323,0.000026604794,0.00077652436,0.000048937167,0.000008640644,0.0000055624814,0.0010005693,0.00021957199,0.00029434092,0.9877283,0.000015830832,0.009845027],"study_design_scores_gemma":[0.00008966772,0.000076813296,0.0030941688,0.0002685451,0.000027561491,0.0000033870003,0.00011085457,0.56250453,0.010444951,0.4221849,0.0010459118,0.00014869515],"about_ca_topic_score_codex":0.000032182994,"about_ca_topic_score_gemma":0.000058731304,"teacher_disagreement_score":0.56554335,"about_ca_system_score_codex":0.000086801876,"about_ca_system_score_gemma":0.00022827921,"threshold_uncertainty_score":0.9695996},"labels":[],"label_agreement":null},{"id":"W4394855929","doi":"10.1515/cllt-2023-0104","title":"Corpus-based discourse analysis: from meta-reflection to accountability","year":2024,"lang":"en","type":"article","venue":"Corpus Linguistics and Linguistic Theory","topic":"Discourse Analysis in Language Studies","field":"Arts and Humanities","cited_by":17,"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":"Australian Research Data Commons; Simon Fraser University; University of Sydney","keywords":"Accountability; Transparency (behavior); Objectivity (philosophy); Reflexivity; Computer science; Corpus linguistics; Consistency (knowledge bases); Subjectivity; Analytics; Data science; Reflection (computer programming); Discourse analysis; Phenomenon; Epistemology; Linguistics; Sociology; Natural language processing; Political science; Artificial intelligence; Social science; Programming language; Philosophy","score_opus":0.04261411326796034,"score_gpt":0.3263412858121069,"score_spread":0.2837271725441466,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4394855929","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":[0.120396875,0.12298944,0.07037655,0.0008560695,0.039039347,0.0018619496,0.006397379,0.0026796383,0.63540274],"genre_scores_gemma":[0.9889687,0.000056242203,0.0005275541,0.00049745955,0.006094821,0.000069195834,0.00015724206,0.000056904937,0.0035719255],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.9976368,0.0001697561,0.0005781447,0.00083518715,0.0003840354,0.0003961009],"domain_scores_gemma":[0.9969619,0.0014813951,0.00012556775,0.0006338585,0.0005980307,0.00019922838],"candidate_categories":["metaepi_narrow","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0009894955,0.00044519283,0.0008636234,0.0005294365,0.0005206843,0.00085248245,0.00026885088,0.00007437427,0.002536039],"category_scores_gemma":[0.005617981,0.00033882566,0.0005654879,0.0003556853,0.0006458539,0.000026734328,0.00012620245,0.00029810885,0.00010863512],"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":[0.00010761938,0.00013505905,0.00033326427,0.00008982736,0.011285699,0.00009580814,0.022252856,0.0003752208,0.000015202389,0.9609221,0.00063963427,0.0037476707],"study_design_scores_gemma":[0.0002729688,0.00018356658,0.00022272141,0.00014885027,0.070445575,8.926394e-7,0.006098666,0.0067101633,0.00010512799,0.19971892,0.71509546,0.000997097],"about_ca_topic_score_codex":0.0016833775,"about_ca_topic_score_gemma":0.0045286966,"teacher_disagreement_score":0.86857176,"about_ca_system_score_codex":0.0000787789,"about_ca_system_score_gemma":0.00011616766,"threshold_uncertainty_score":0.99990636},"labels":[],"label_agreement":null},{"id":"W4405006658","doi":"10.1515/cllt-2024-0015","title":"A multivariate analysis of canonical and non-canonical uses of switch-reference markers in Mbyá narratives","year":2024,"lang":"en","type":"article","venue":"Corpus Linguistics and Linguistic Theory","topic":"Syntax, Semantics, Linguistic Variation","field":"Arts and Humanities","cited_by":0,"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 Toronto","funders":"Social Sciences and Humanities Research Council of Canada","keywords":"Canonical form; Computer science; Linguistics; Non canonical; Narrative; Epiphenomenon; Argument (complex analysis); Mathematics; Pure mathematics; Philosophy; Medicine","score_opus":0.024903084471919315,"score_gpt":0.27109074584601395,"score_spread":0.24618766137409465,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4405006658","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":[0.90373605,0.0040627834,0.0042765513,0.00028747437,0.006256714,0.00069670926,0.00088945037,0.00014128107,0.07965297],"genre_scores_gemma":[0.99743444,0.0001244637,0.0011407756,0.000027184216,0.00087504403,0.000010851092,0.00004005275,0.00003679108,0.0003104098],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","domain_scores_codex":[0.9977602,0.00017856907,0.00096030545,0.00052321,0.00026770085,0.00031000873],"domain_scores_gemma":[0.9958156,0.002827311,0.00029352936,0.00032021658,0.00061768485,0.0001256904],"candidate_categories":["metaresearch","metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0011283872,0.00030401768,0.00083235593,0.00064620556,0.00012288764,0.00013799481,0.0001883236,0.00013439608,0.00020884196],"category_scores_gemma":[0.014841398,0.00027248918,0.00012844642,0.0002572711,0.0008477524,0.00001744408,0.00013323437,0.0003169888,0.0000025840618],"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":[0.00019364968,0.00010737305,0.0031384358,0.00038582407,0.00068768277,0.000050175535,0.043195527,0.0000376499,0.00014725514,0.9513435,0.000006494733,0.00070645835],"study_design_scores_gemma":[0.0016462328,0.0006537895,0.03240704,0.002120022,0.0047660605,0.0000073337187,0.013546571,0.05572878,0.0002261156,0.8743546,0.013312573,0.0012308734],"about_ca_topic_score_codex":0.0030919772,"about_ca_topic_score_gemma":0.0030975447,"teacher_disagreement_score":0.09369837,"about_ca_system_score_codex":0.00004773345,"about_ca_system_score_gemma":0.00030759102,"threshold_uncertainty_score":0.9999727},"labels":[],"label_agreement":null}]}