{"meta":{"page":1,"per_page":50,"max_per_page":100,"total":26,"total_is_capped":false,"direct_labels_cover":0,"predictions_cover":26,"direct_label_status":"direct model label, unvalidated","prediction_status":"machine_predicted_unvalidated (Codex and Gemma teacher distillation)","score_status":"score_only:v0-immature-baseline (scores rank; they never assert a category)","snapshot":{"source":"OpenAlex, pinned release, all 482 partitions","release":"2026-06-24","frame_built":"2026-07-12"},"query_hash":"ce48bc7d526f","filters":{"venue":"Linguistics Vanguard"}},"results":[{"id":"W2892039495","doi":"10.1515/lingvan-2017-0027","title":"The role of predictability in shaping phonological patterns","year":2018,"lang":"en","type":"article","venue":"Linguistics Vanguard","topic":"Phonetics and Phonology Research","field":"Psychology","cited_by":108,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of British Columbia","funders":"","keywords":"Predictability; Context (archaeology); Computer science; Set (abstract data type); Comprehension; Reduction (mathematics); Cognitive psychology; Linguistics; Natural language processing; Psychology; Mathematics; Statistics; History","retraction":null,"screen_n_in":null,"score":{"opus":0.04052652632450414,"gpt":0.3620769572044678,"spread":0.3215504308799637,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006478362,0.0001018903,0.0001785149,0.00006167284,0.0001016738,0.00001446632,0.0004248204,0.0001647234,0.0005340938],"category_scores_gemma":[0.002489907,0.00007483588,0.0000505432,0.000158266,0.0005267668,0.000003925022,0.000156163,0.0003508033,0.0001226859],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002544287,"about_ca_system_score_gemma":0.00003995542,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000731262,"about_ca_topic_score_gemma":0.0007721783,"domain_scores_codex":[0.9986441,0.0001766599,0.0003397404,0.0002661021,0.0001492131,0.0004242305],"domain_scores_gemma":[0.9985314,0.0005543086,0.00007805302,0.0005260561,0.0002519801,0.00005817848],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.00033528,0.000322118,0.9408848,0.00001458124,0.00007159435,0.00002881318,0.01170945,0.000001858673,0.001239542,0.02621062,0.0005214901,0.01865985],"study_design_scores_gemma":[0.0005272108,0.0004890959,0.9227637,0.00002404845,0.00001104093,0.000004867123,0.001354896,0.0004901481,0.002567464,0.03723066,0.03439258,0.0001442639],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9494593,0.0007172826,0.00008061403,0.0000629803,0.001514144,0.0001730137,0.0000378732,0.00002820882,0.04792657],"genre_scores_gemma":[0.9988282,0.00002938172,0.0001284622,0.00004393829,0.00074412,0.00002082889,0.000004279545,0.00001145888,0.0001893788],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.04936884,"threshold_uncertainty_score":0.5847952,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2885411231","doi":"10.1515/lingvan-2017-0018","title":"Practice makes perfect: the consequences of lexical proficiency for articulation","year":2018,"lang":"en","type":"article","venue":"Linguistics Vanguard","topic":"Phonetics and Phonology Research","field":"Psychology","cited_by":64,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"Deutsche Forschungsgemeinschaft","keywords":"Coarticulation; Articulation (sociology); Speech recognition; Linguistics; Psychology; Computer science; Speech production; Cognitive psychology; Constraint (computer-aided design); Mathematics; Vowel","retraction":null,"screen_n_in":null,"score":{"opus":0.06326567726277735,"gpt":0.4367760378886497,"spread":0.3735103606258723,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.000729889,0.00008199106,0.0001257421,0.00004531744,0.0001703391,0.00002005282,0.0002088969,0.00009770432,0.0002112427],"category_scores_gemma":[0.009058444,0.00005736864,0.0000515027,0.0001470248,0.0008969777,0.000008826358,0.00003376569,0.0001454294,0.00009802198],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001090577,"about_ca_system_score_gemma":0.00008294907,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001519932,"about_ca_topic_score_gemma":0.00002484216,"domain_scores_codex":[0.9989881,0.0001465088,0.0002370548,0.0002007861,0.0001591303,0.0002684328],"domain_scores_gemma":[0.9973739,0.001256008,0.0001241716,0.0003149718,0.0008907295,0.00004023673],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.001433558,0.0007937686,0.009155502,0.00009041991,0.000312492,0.00003389416,0.04272612,0.00002222657,0.02479868,0.8954041,0.01752679,0.007702479],"study_design_scores_gemma":[0.003227343,0.008199631,0.1021982,0.000106436,0.0006087114,0.0001809285,0.01056531,0.004956033,0.08708089,0.1789439,0.6030039,0.0009286919],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.7574847,0.001330615,0.01185448,0.004363406,0.007717516,0.00174817,0.000112595,0.00009190885,0.2152966],"genre_scores_gemma":[0.9953833,0.000005605237,0.00240374,0.0002593168,0.001105438,0.00004779985,0.000006253027,0.00001087831,0.0007776528],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7164602,"threshold_uncertainty_score":0.9992887,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2800761643","doi":"10.1515/lingvan-2016-0081","title":"The syntax-prosody interface: current theoretical approaches and outstanding questions","year":2018,"lang":"en","type":"article","venue":"Linguistics Vanguard","topic":"Phonetics and Phonology Research","field":"Psychology","cited_by":50,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"York University","funders":"","keywords":"Prosody; Syntax; Interface (matter); Linguistics; Computer science; Situated; Field (mathematics); Natural language processing; Recursion (computer science); Current (fluid); Artificial intelligence; Programming language; Physics; Mathematics; Speech recognition; Philosophy","retraction":null,"screen_n_in":null,"score":{"opus":0.06771065085330953,"gpt":0.4011014568515419,"spread":0.3333908059982323,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006201073,0.0001490582,0.0001521657,0.00005888725,0.0004864408,0.0001106526,0.0002898899,0.0001155291,0.0002326335],"category_scores_gemma":[0.002303725,0.0001046397,0.00004140141,0.0001178228,0.001863785,0.000006036028,0.0002027887,0.0004770393,0.0002499834],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000274618,"about_ca_system_score_gemma":0.00003525438,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002458564,"about_ca_topic_score_gemma":0.00003113921,"domain_scores_codex":[0.9986475,0.0001841836,0.000227586,0.0003273008,0.0001531353,0.000460317],"domain_scores_gemma":[0.9985744,0.0006005972,0.00005376539,0.0004506165,0.0001898621,0.0001307949],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0001078585,0.00008474397,0.002208216,0.00001211959,0.00007327605,0.00001095479,0.006543234,5.036981e-7,0.00006145829,0.9758859,0.003013389,0.01199839],"study_design_scores_gemma":[0.001527747,0.00140142,0.02737514,0.000193372,0.0002196295,0.0001271848,0.004780917,0.004842154,0.001499418,0.5580691,0.3990067,0.0009571689],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5508738,0.02433426,0.02440832,0.002694475,0.02899742,0.001402464,0.0001999001,0.0003995777,0.3666898],"genre_scores_gemma":[0.9971669,0.0001072916,0.0003729199,0.00002549537,0.001642831,0.00003488968,0.000004169131,0.00002356352,0.0006219433],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4462931,"threshold_uncertainty_score":0.6867194,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2609372572","doi":"10.1515/lingvan-2015-0032","title":"Is <i>like</i> like <i>like</i>?: Evaluating the same variant across multiple variables","year":2017,"lang":"en","type":"article","venue":"Linguistics Vanguard","topic":"Linguistic Variation and Morphology","field":"Social Sciences","cited_by":35,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"","keywords":"Salient; Meaning (existential); Variation (astronomy); Perception; Linguistics; Psychology; Variable (mathematics); Verb; Perspective (graphical); Transitive relation; Cognitive psychology; Social psychology; Mathematics; Computer science","retraction":null,"screen_n_in":null,"score":{"opus":0.07200595367626476,"gpt":0.3984130001953998,"spread":0.326407046519135,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","metaepi_narrow","sts"],"consensus_categories":[],"category_scores_codex":[0.003966228,0.0003401949,0.0004380493,0.00004505835,0.006296459,0.0009911651,0.001734982,0.0003591336,0.0004829149],"category_scores_gemma":[0.05470127,0.0002847021,0.0002094417,0.0001676184,0.0009028098,0.00007348118,0.0005014567,0.0005232553,0.000333963],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001462582,"about_ca_system_score_gemma":0.0006331211,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01555432,"about_ca_topic_score_gemma":0.004447624,"domain_scores_codex":[0.996197,0.000448696,0.0006307496,0.0006974865,0.0009446282,0.001081407],"domain_scores_gemma":[0.9945827,0.001483155,0.0007324493,0.001594745,0.001314209,0.0002927493],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0003922825,0.0007167386,0.04324954,0.0001515887,0.0006790035,0.0006467864,0.3150577,0.00040333,0.0008536417,0.4616953,0.1620257,0.01412846],"study_design_scores_gemma":[0.001246784,0.0001046652,0.002241599,0.00005257501,0.0001586032,0.000009839844,0.004011905,0.003310067,0.00005327383,0.01359855,0.9747027,0.0005094408],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"other","genre_gemma":"empirical","genre_scores_codex":[0.05498469,0.001374031,0.0201131,0.01930149,0.1460776,0.004189054,0.001402467,0.001574691,0.7509829],"genre_scores_gemma":[0.9501131,0.0001271821,0.0131259,0.01057978,0.008888119,0.00005794592,0.00003471881,0.00006792948,0.01700528],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8951284,"threshold_uncertainty_score":0.9999605,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2030992269","doi":"10.1515/lingvan-2014-1008","title":"Language structure and social agency: Confirming polar questions in conversation","year":2015,"lang":"en","type":"article","venue":"Linguistics Vanguard","topic":"Language, Discourse, Communication Strategies","field":"Arts and Humanities","cited_by":25,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"","keywords":"Semiotics; Linguistics; Conversation; Agency (philosophy); Repetition (rhetorical device); Function (biology); Conversation analysis; Selection (genetic algorithm); Computer science; Epistemology; Psychology; Artificial intelligence; Philosophy","retraction":null,"screen_n_in":null,"score":{"opus":0.04871211242979754,"gpt":0.3079362462959112,"spread":0.2592241338661136,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002096908,0.0001212622,0.0001563616,0.0001041942,0.0001655182,0.0001894659,0.0001408003,0.00006361209,0.0002014391],"category_scores_gemma":[0.0006681449,0.0001227904,0.00002508995,0.00003702843,0.0002252603,0.000067318,0.00005553029,0.0002034906,0.00001855647],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004942196,"about_ca_system_score_gemma":0.0000975889,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001693373,"about_ca_topic_score_gemma":0.006245084,"domain_scores_codex":[0.9992064,0.0001094434,0.0002104416,0.0001474954,0.0001571034,0.0001691397],"domain_scores_gemma":[0.9993386,0.00008331185,0.00009498847,0.0001940867,0.0002374868,0.00005157522],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"qualitative","study_design_scores_codex":[0.00001252819,0.00002329284,0.0009487022,0.00002619209,0.00001625275,0.00002154817,0.4642265,0.00000725485,0.00006092272,0.5333145,0.0005940315,0.0007482219],"study_design_scores_gemma":[0.002706569,0.0001402334,0.01037364,0.0001343396,0.0001884277,0.00001387817,0.5839049,0.0009497092,0.0003029241,0.07776256,0.3224002,0.001122689],"study_design_candidate":"qualitative","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8740878,0.002052102,0.00008473593,0.0008446592,0.001622329,0.0003228534,0.0003662526,0.0002217768,0.1203975],"genre_scores_gemma":[0.9973555,0.000007583486,0.0005064412,0.0001817267,0.001163722,0.000005087605,0.0001379629,0.00001800117,0.0006240272],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.455552,"threshold_uncertainty_score":0.5007246,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2335647126","doi":"10.1515/lingvan-2015-0010","title":"An evaluation of noise on LPC-based vowel formant estimates: Implications for sociolinguistic data collection","year":2016,"lang":"en","type":"article","venue":"Linguistics Vanguard","topic":"Linguistic Variation and Morphology","field":"Social Sciences","cited_by":21,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Memorial University of Newfoundland","funders":"","keywords":"Formant; Spurious relationship; Noise (video); Vowel; Computer science; Speech recognition; Data collection; SIGNAL (programming language); Interference (communication); Acoustics; Telecommunications; Artificial intelligence; Mathematics; Statistics; Physics; Machine learning","retraction":null,"screen_n_in":null,"score":{"opus":0.1564448628560941,"gpt":0.4492947400870724,"spread":0.2928498772309783,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.002754555,0.0001449976,0.0002211431,0.0001554913,0.0005295292,0.00003560562,0.0005189349,0.0001660176,0.0001196828],"category_scores_gemma":[0.07845806,0.0001227372,0.00006328375,0.0001892919,0.0002148713,0.00003448502,0.00003217714,0.00006594905,0.00002002998],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002756346,"about_ca_system_score_gemma":0.001225138,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0003775142,"about_ca_topic_score_gemma":0.0006692701,"domain_scores_codex":[0.9981651,0.0002051543,0.0004406635,0.0004193546,0.0004444587,0.0003253169],"domain_scores_gemma":[0.994322,0.001383069,0.0003415131,0.0008113388,0.002999767,0.0001423259],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"theoretical_or_conceptual","study_design_scores_codex":[0.0004151838,0.001233289,0.009786829,0.0001152739,0.0001516199,0.000002431529,0.01059919,0.001279998,0.002394876,0.9217427,0.01600485,0.0362737],"study_design_scores_gemma":[0.01212978,0.001901008,0.03600545,0.0004532261,0.002426763,0.000002435535,0.001345385,0.2515105,0.003312286,0.4490386,0.2402009,0.001673638],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02391722,0.0002006644,0.8575555,0.005517544,0.01555969,0.005975783,0.004135894,0.0007873171,0.08635037],"genre_scores_gemma":[0.9803358,0.0000125153,0.01690877,0.0002467692,0.001822954,0.0001083062,0.0003935725,0.00002583172,0.0001455278],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9564185,"threshold_uncertainty_score":0.9293045,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2565228806","doi":"10.1515/lingvan-2015-0012","title":"Extending ELAN into variationist sociolinguistics","year":2015,"lang":"en","type":"article","venue":"Linguistics Vanguard","topic":"Linguistic Variation and Morphology","field":"Social Sciences","cited_by":16,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"","keywords":"Computer science; Coding (social sciences); Transcription (linguistics); Linguistics; Natural language processing; Statistics; Mathematics","retraction":null,"screen_n_in":null,"score":{"opus":0.05606557689429376,"gpt":0.3702220372922769,"spread":0.3141564603979831,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.001736075,0.0001895083,0.0002675616,0.0001359213,0.0008280791,0.0001904394,0.0003875233,0.0002540984,0.0002095313],"category_scores_gemma":[0.08496456,0.0002120179,0.00008526232,0.0002895538,0.0003227751,0.000022266,0.00009038387,0.0002724227,0.0004567648],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000412946,"about_ca_system_score_gemma":0.000983961,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004404895,"about_ca_topic_score_gemma":0.001095521,"domain_scores_codex":[0.9977829,0.0002723573,0.0004392652,0.0003561649,0.0006194236,0.0005298223],"domain_scores_gemma":[0.9968484,0.0003769467,0.0002243497,0.0003256416,0.001730785,0.0004938958],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00001442833,0.00008198884,0.002481443,0.000008685107,0.00002645208,0.0000704267,0.03050897,0.00005212337,0.000006316073,0.957077,0.009016959,0.0006552306],"study_design_scores_gemma":[0.0004864643,0.00005649521,0.0002214931,0.00001234442,0.00005740389,0.00000213419,0.003566882,0.0003992124,0.00001316692,0.1611912,0.833716,0.0002771687],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"other","genre_gemma":"empirical","genre_scores_codex":[0.001455675,0.0002656966,0.03400223,0.001078213,0.02195857,0.0003546261,0.00002517966,0.0005975165,0.9402623],"genre_scores_gemma":[0.9442065,0.00003654828,0.03691076,0.0008244087,0.01318197,0.00001327411,0.00005948528,0.00003707188,0.004729935],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9427509,"threshold_uncertainty_score":0.9227432,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4306249762","doi":"10.1515/lingvan-2021-0122","title":"Phonetic change over the career: a case study","year":2022,"lang":"en","type":"article","venue":"Linguistics Vanguard","topic":"Linguistic Variation and Morphology","field":"Social Sciences","cited_by":14,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"European Commission","keywords":"Flexibility (engineering); Perspective (graphical); Psychology; Linguistics; Longitudinal study; Feature (linguistics); Computer science; Mathematics","retraction":null,"screen_n_in":null,"score":{"opus":0.07789135288875891,"gpt":0.3521037197557527,"spread":0.2742123668669938,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["sts","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.001174585,0.0001090692,0.000144282,0.00007038042,0.001930848,0.00006361878,0.0003609475,0.00004001481,0.001638638],"category_scores_gemma":[0.004797682,0.00009522582,0.00006248751,0.0003372344,0.0001262533,0.000007894088,0.0002352242,0.0002973334,0.00003636017],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001460402,"about_ca_system_score_gemma":0.0001708738,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.04066017,"about_ca_topic_score_gemma":0.01042874,"domain_scores_codex":[0.9980891,0.0005866437,0.0002142022,0.0002482675,0.0005228744,0.0003389161],"domain_scores_gemma":[0.9988839,0.0003338671,0.00010974,0.0003442342,0.0002290792,0.00009919305],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00002990213,0.0007090044,0.03522528,0.000008751606,0.0001022125,0.01588521,0.8066575,0.00006155956,0.000002881986,0.1245781,0.01379267,0.002946971],"study_design_scores_gemma":[0.0005066087,0.0001713149,0.001990304,0.000001344562,0.00009194743,0.0001109959,0.1385687,0.0001348156,5.070017e-7,0.001145238,0.8571028,0.0001754393],"study_design_candidate":"qualitative","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8362563,0.0006480202,0.0001530079,0.00164115,0.03010899,0.003015409,0.000164727,0.0004309623,0.1275815],"genre_scores_gemma":[0.9921755,0.000006309687,0.0001138335,0.001639415,0.004113343,0.0002134956,0.00000647005,0.00001819479,0.001713447],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.8433102,"threshold_uncertainty_score":0.9993685,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W2590027471","doi":"10.1515/lingvan-2016-0025","title":"Individual differences in second language speech perception across tasks and contrasts: The case of English vowel contrasts by Korean learners","year":2017,"lang":"en","type":"article","venue":"Linguistics Vanguard","topic":"Phonetics and Phonology Research","field":"Psychology","cited_by":12,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"McGill University","funders":"Santé, Sciences Biologiques et Chimie du Vivant","keywords":"Psychology; Perception; Vowel; Formant; Weighting; Task (project management); Contrast (vision); Duration (music); Cognitive psychology; Speech perception; Natural (archaeology); Two-alternative forced choice; American English; Linguistics; Speech recognition; Computer science; Artificial intelligence","retraction":null,"screen_n_in":null,"score":{"opus":0.02822879181188652,"gpt":0.3582191485080133,"spread":0.3299903566961268,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0007831457,0.0001828764,0.0003326683,0.00005926545,0.0002991825,0.0001525919,0.0004711918,0.0002368908,0.0003820506],"category_scores_gemma":[0.00318542,0.0001447708,0.00004734545,0.0000515942,0.0008660239,0.00001974492,0.0001775562,0.0006020233,0.00001237392],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001864137,"about_ca_system_score_gemma":0.00002852481,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005565835,"about_ca_topic_score_gemma":0.009653666,"domain_scores_codex":[0.9985169,0.0001935718,0.0003126823,0.0003477122,0.0001481915,0.0004809655],"domain_scores_gemma":[0.9983889,0.0004671225,0.0002056224,0.0006022353,0.0002337148,0.0001023722],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"observational","study_design_scores_codex":[0.000454141,0.0005103272,0.2825139,0.0001124055,0.0004289052,0.00378771,0.6636983,7.352448e-7,0.004012932,0.002120613,0.006164393,0.0361957],"study_design_scores_gemma":[0.004269274,0.0005105149,0.8796031,0.00006690637,0.00008213358,0.0002118007,0.1103129,0.0001564841,0.0008390786,0.001666488,0.001788823,0.000492533],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9880775,0.0005912608,0.000005295432,0.00005139386,0.001228164,0.0002518953,0.0009150328,0.00001720354,0.008862226],"genre_scores_gemma":[0.998409,0.00004229469,0.0000550698,0.0000752234,0.0005846138,0.00001635303,0.00004123579,0.00002074822,0.0007554974],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.5970892,"threshold_uncertainty_score":0.8413914,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4214853268","doi":"10.1515/lingvan-2021-0057","title":"Disruptions due to COVID-19: using mixed methods to identify factors influencing language maintenance and shift","year":2022,"lang":"en","type":"article","venue":"Linguistics Vanguard","topic":"Multilingual Education and Policy","field":"Social Sciences","cited_by":5,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"","keywords":"Context (archaeology); Social distance; Recreation; Language shift; Perception; Psychology; Distancing; Affect (linguistics); Coronavirus disease 2019 (COVID-19); Identity (music); Heritage language; Social psychology; Sociology; Pedagogy; Linguistics; Geography; Political science; Medicine; Communication","retraction":null,"screen_n_in":null,"score":{"opus":0.1032113126007449,"gpt":0.5455870256524398,"spread":0.4423757130516949,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch","sts"],"consensus_categories":[],"category_scores_codex":[0.001389524,0.0001478563,0.0002131379,0.0002603147,0.001513195,0.0001531593,0.0003082004,0.00005423035,0.0004547508],"category_scores_gemma":[0.03474924,0.0001649727,0.00006184293,0.0006690709,0.00009391699,0.00002711027,0.0002584173,0.0002308699,0.00001409815],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0005138802,"about_ca_system_score_gemma":0.0006882994,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.02866468,"about_ca_topic_score_gemma":0.003688114,"domain_scores_codex":[0.9980452,0.00054636,0.000285787,0.0003401501,0.0003383489,0.0004441588],"domain_scores_gemma":[0.998056,0.0006237666,0.0001030681,0.0002589915,0.0001353818,0.0008227903],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00001995847,0.00006642647,0.005055289,0.00004307658,0.00002531085,0.00004904556,0.9441041,0.002123094,0.001194299,0.0403217,0.001896103,0.005101577],"study_design_scores_gemma":[0.0001836843,0.00003898494,0.01210767,0.00001938909,0.00004404971,0.000004274748,0.2120316,0.0001991557,0.0001816603,0.001917849,0.772859,0.0004127059],"study_design_candidate":"qualitative","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9734365,0.0001046943,0.01569947,0.00136591,0.003485306,0.0004687337,0.0001919195,0.0001483833,0.005099077],"genre_scores_gemma":[0.9559951,0.000002921638,0.03724935,0.004744507,0.0008846924,0.00003020418,0.00001298669,0.00002392939,0.001056242],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7709629,"threshold_uncertainty_score":0.9997867,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4312200521","doi":"10.1515/lingvan-2022-0017","title":"The Red Hen Anonymizer and the Red Hen Protocol for de-identifying audiovisual recordings","year":2022,"lang":"en","type":"article","venue":"Linguistics Vanguard","topic":"Hearing Impairment and Communication","field":"Psychology","cited_by":4,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of British Columbia","funders":"Fundación Séneca; Google; Alexander von Humboldt-Stiftung","keywords":"JSON; Computer science; Software; Protocol (science); Identification (biology); Face (sociological concept); Multimedia; Human–computer interaction; Speech recognition; World Wide Web; Programming language; Linguistics","retraction":null,"screen_n_in":null,"score":{"opus":0.06178125760216429,"gpt":0.4070607818337643,"spread":0.3452795242316,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["sts"],"consensus_categories":[],"category_scores_codex":[0.002369405,0.0001309077,0.0001748994,0.00004293756,0.001385985,0.0001409702,0.0005149454,0.00005755413,0.0001521612],"category_scores_gemma":[0.002245691,0.00009363369,0.00009739187,0.0001250019,0.000133334,0.00001278822,0.000336372,0.0003567381,0.000008914038],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00008516163,"about_ca_system_score_gemma":0.00005905051,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001704237,"about_ca_topic_score_gemma":0.00001720208,"domain_scores_codex":[0.9983411,0.0005294798,0.000344111,0.0002383657,0.0001876657,0.0003593039],"domain_scores_gemma":[0.9978731,0.001073215,0.0001951249,0.0006785789,0.0001272869,0.00005265286],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.01684669,0.0008855751,0.0436768,0.0003462574,0.00110309,0.00005495935,0.07790195,0.00007520764,0.0006502143,0.2515112,0.5143948,0.09255324],"study_design_scores_gemma":[0.004225361,0.0002365014,0.003713012,0.00002276069,0.00006311396,0.00001477368,0.002878117,0.000866512,0.00002451224,0.02065813,0.9671224,0.0001747506],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"protocol","genre_gemma":"empirical","genre_scores_codex":[0.07064747,0.0009676875,0.01316404,0.03093107,0.01295587,0.5364692,0.0001412069,0.001449628,0.3332738],"genre_scores_gemma":[0.5990519,0.00001040134,0.004361557,0.0009749964,0.001594319,0.3731782,0.00003183052,0.00009358528,0.02070323],"genre_candidate":"protocol","genre_consensus":null,"teacher_disagreement_score":0.5284045,"threshold_uncertainty_score":0.9999141,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4210262403","doi":"10.1515/lingvan-2020-0013","title":"A socially anchored approach to spatial language in Kalaallisut","year":2022,"lang":"en","type":"article","venue":"Linguistics Vanguard","topic":"Categorization, perception, and language","field":"Psychology","cited_by":3,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"","keywords":"Sociocultural evolution; Affect (linguistics); Encoding (memory); Space (punctuation); Arctic; Geography; Sociology; Frame (networking); Linguistics; Psychology; Computer science; Communication; Ecology; Cognitive psychology; Anthropology; Philosophy","retraction":null,"screen_n_in":null,"score":{"opus":0.01596708737772196,"gpt":0.3064200608114498,"spread":0.2904529734337278,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0004002118,0.0001955207,0.0002649415,0.000275815,0.0002120847,0.00004056221,0.000351448,0.00009432044,0.003526985],"category_scores_gemma":[0.0005434396,0.0002308374,0.00009138254,0.000533508,0.00003718103,0.00001150189,0.0001402005,0.0003499314,0.0001877208],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001911937,"about_ca_system_score_gemma":0.0001134015,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004173617,"about_ca_topic_score_gemma":0.001499187,"domain_scores_codex":[0.9981329,0.000207363,0.0003681352,0.000500656,0.0003470881,0.0004438246],"domain_scores_gemma":[0.9991899,0.00006051773,0.00009038088,0.0004405303,0.00009859149,0.0001200578],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0006477095,0.002853896,0.04203585,0.00009261102,0.0001641987,0.0006779336,0.6891073,0.003439514,0.001842352,0.1835666,0.0546731,0.02089891],"study_design_scores_gemma":[0.006005803,0.0007836807,0.2526287,0.00002552924,0.0001723861,0.00007216985,0.1201841,0.001598785,0.00007091754,0.003934887,0.6123113,0.002211708],"study_design_candidate":"qualitative","study_design_consensus":null,"genre_codex":"other","genre_gemma":"empirical","genre_scores_codex":[0.4551732,0.0005811626,0.01365761,0.0001751008,0.007231135,0.00128021,0.0004924398,0.0004106332,0.5209985],"genre_scores_gemma":[0.9856815,0.000002550769,0.001803951,0.0008529087,0.002649513,0.0002109157,0.0005612986,0.0000601585,0.008177241],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.5689232,"threshold_uncertainty_score":0.997384,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4285586648","doi":"10.1515/lingvan-2021-0048","title":"Cross-dialectal synchronic variation of a diachronic conditioned merger in Tlingit","year":2022,"lang":"en","type":"article","venue":"Linguistics Vanguard","topic":"Phonetics and Phonology Research","field":"Psychology","cited_by":2,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"McGill University; University of British Columbia","funders":"","keywords":"Variation (astronomy); Sound change; Perception; Linguistics; Indigenous; Set (abstract data type); History; Computer science; Psychology; Biology","retraction":null,"screen_n_in":null,"score":{"opus":0.02094465324947253,"gpt":0.3604735833782272,"spread":0.3395289301287547,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.0005818018,0.0001246799,0.0002492107,0.0002722772,0.0001541472,0.00001561991,0.0002912941,0.0001035596,0.008088144],"category_scores_gemma":[0.0009811647,0.0001493927,0.00008476734,0.0004349921,0.000136911,0.000009087147,0.0001654877,0.0005289699,0.0001026303],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001909116,"about_ca_system_score_gemma":0.0002236634,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0006255769,"about_ca_topic_score_gemma":0.00008236059,"domain_scores_codex":[0.9983071,0.0002508154,0.0004221812,0.0003270593,0.0002755141,0.0004173065],"domain_scores_gemma":[0.9989324,0.0002988208,0.0001623301,0.000383806,0.0001669504,0.00005574102],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.003947408,0.00726965,0.345552,0.0005631509,0.002019008,0.002310273,0.226648,0.03245692,0.06329413,0.2992784,0.01133716,0.005323901],"study_design_scores_gemma":[0.004492316,0.001195025,0.9567093,0.00002372205,0.00007105155,0.00003108281,0.0005617509,0.004177642,0.0007003571,0.01361966,0.0178597,0.0005584147],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9837407,0.0008889957,0.0001680357,0.00003466164,0.00330972,0.0002642343,0.0002328813,0.00003661974,0.0113241],"genre_scores_gemma":[0.9980419,0.000009425576,0.0001241673,0.00005437377,0.0004034917,0.0001199317,0.0001096778,0.00002859179,0.001108402],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.6111573,"threshold_uncertainty_score":0.9928186,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4210791957","doi":"10.1515/lingvan-2020-0016","title":"Conflation of spatial reference frames in deaf community sign languages","year":2022,"lang":"en","type":"article","venue":"Linguistics Vanguard","topic":"Hearing Impairment and Communication","field":"Psychology","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of British Columbia","funders":"","keywords":"Modality (human–computer interaction); Conflation; Affordance; Sign language; Modalities; Sign (mathematics); American Sign Language; Frame of reference; Frame (networking); Computer science; Linguistics; Manually coded language; Feature (linguistics); Sociolinguistics of sign languages; Psychology; Artificial intelligence; Sociology; Human–computer interaction; Mathematics","retraction":null,"screen_n_in":null,"score":{"opus":0.0589559648813121,"gpt":0.3698895156348347,"spread":0.3109335507535226,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006548563,0.00007687213,0.0001453568,0.0001081414,0.0001464382,0.000008197885,0.0003279322,0.00005651806,0.0007500916],"category_scores_gemma":[0.0008163465,0.00009028758,0.00002696365,0.0001317414,0.00004528767,0.000007496276,0.000180551,0.000608095,0.00002095151],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00005155575,"about_ca_system_score_gemma":0.00003357975,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.009371114,"about_ca_topic_score_gemma":0.0005461115,"domain_scores_codex":[0.9986112,0.000718658,0.000270789,0.00009799404,0.0001543893,0.0001470378],"domain_scores_gemma":[0.9988126,0.0003862836,0.0001276942,0.0005659357,0.00008184509,0.00002568738],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0006690178,0.002797573,0.5927911,0.0001410667,0.0001389782,0.00006960111,0.09903287,0.00119866,0.002230643,0.2286529,0.003482508,0.06879513],"study_design_scores_gemma":[0.001814111,0.001043864,0.8698359,0.00005865721,0.00005097602,0.000006277596,0.02376723,0.001008494,0.0007125019,0.00792852,0.09333003,0.0004434975],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8785906,0.0001927358,0.0003856163,0.00005696954,0.0005948053,0.000185522,0.00003696039,0.00005783618,0.119899],"genre_scores_gemma":[0.9984816,0.000004567541,0.0005620446,0.00009266484,0.00008509319,0.00003416324,0.000107488,0.00001210627,0.0006202697],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.2770448,"threshold_uncertainty_score":0.9972256,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4294663728","doi":"10.1515/lingvan-2021-0109","title":"Exploring variation and change in a small-scale Indigenous society: the case of (s) in Pirahã","year":2022,"lang":"en","type":"article","venue":"Linguistics Vanguard","topic":"Linguistic Variation and Morphology","field":"Social Sciences","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of British Columbia","funders":"","keywords":"Variation (astronomy); Sound change; Indigenous; Context (archaeology); Scale (ratio); Language change; Social change; Psychology; Demography; Geography; Social psychology; Sociology; Linguistics; Political science; Cartography; Ecology; Biology","retraction":null,"screen_n_in":null,"score":{"opus":0.1217047399356564,"gpt":0.3238299199749468,"spread":0.2021251800392904,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001416626,0.00006872105,0.000140154,0.00009324172,0.0003673709,0.00001784216,0.000130098,0.00004424258,0.00005138182],"category_scores_gemma":[0.001908455,0.00006845085,0.00003359323,0.000378065,0.00008421523,0.00001641823,0.0001119127,0.0002214123,9.4311e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001389139,"about_ca_system_score_gemma":0.000125316,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_topic_score_codex":0.07392507,"about_ca_topic_score_gemma":0.04688304,"domain_scores_codex":[0.9989864,0.0002555675,0.0002447105,0.0001662932,0.0001197451,0.0002272325],"domain_scores_gemma":[0.9993369,0.0002957503,0.0001164613,0.0001233626,0.0000898956,0.00003761816],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"qualitative","study_design_scores_codex":[0.00001321353,0.00009443733,0.0197467,0.00001845056,0.000008666831,0.0003630562,0.9278809,0.00007629187,0.00001747655,0.04829109,0.00001331457,0.003476384],"study_design_scores_gemma":[0.005216312,0.0006386887,0.3217284,0.0001091172,0.0002100056,0.0002270648,0.4797274,0.008554881,0.00009903773,0.04729557,0.134867,0.001326494],"study_design_candidate":"qualitative","study_design_consensus":"qualitative","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.995123,0.0001901056,0.0001059748,0.0003154304,0.001395449,0.0004101396,0.00002330376,0.0000230374,0.002413531],"genre_scores_gemma":[0.9979302,0.00009343631,0.0008378426,0.0003252612,0.000605705,0.0001386512,0.000004670121,0.000009005332,0.00005518328],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4481535,"threshold_uncertainty_score":0.9705089,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4402422143","doi":"10.1515/lingvan-2023-0102","title":"Bibliographic bias and information-density sampling","year":2024,"lang":"en","type":"article","venue":"Linguistics Vanguard","topic":"Advanced Text Analysis Techniques","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"Marcus och Amalia Wallenbergs minnesfond","keywords":"Sampling bias; Sampling (signal processing); Statistics; Information retrieval; Computer science; Geography; Mathematics; Sample size determination; Telecommunications","retraction":null,"screen_n_in":null,"score":{"opus":0.02514485621604285,"gpt":0.2974952076023429,"spread":0.2723503513863001,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002958767,0.0001277588,0.0001418073,0.001683202,0.0001068023,0.0006350323,0.0003219405,0.00006018353,0.000003844234],"category_scores_gemma":[0.001131674,0.0001201307,0.00007298486,0.002736359,0.00004987914,0.0004701679,0.0002339889,0.0001789488,0.00004269451],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001679643,"about_ca_system_score_gemma":0.00003301121,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003091949,"about_ca_topic_score_gemma":0.000008130378,"domain_scores_codex":[0.9990843,0.00001887395,0.0002474861,0.0002350824,0.0002240596,0.0001902181],"domain_scores_gemma":[0.9989987,0.0002275576,0.00005338088,0.0003934102,0.0002487478,0.00007820459],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.000001415808,0.000009976211,0.002087915,0.00008953161,0.00004973013,0.00003598423,0.0006902107,0.00004904174,0.00007479623,0.9393233,0.001636933,0.05595113],"study_design_scores_gemma":[0.0001211368,0.00007187185,0.003570243,0.0001780864,0.00007043198,0.00003193379,0.0000445423,0.09879767,0.001559801,0.2851722,0.6098071,0.0005749663],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.003108893,0.0007748639,0.9896422,0.0001230073,0.0006127658,0.000077845,0.000004662995,0.001117081,0.004538704],"genre_scores_gemma":[0.7037522,0.0003451489,0.2953639,0.0002694214,0.0002193179,0.000004930253,0.000006486187,0.00000769359,0.00003093257],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.7006433,"threshold_uncertainty_score":0.6123632,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W3092479361","doi":"10.1515/lingvan-2018-0068","title":"<i>I feel like</i> and <i>it feels like</i>: Two paths to the emergence of epistemic markers","year":2020,"lang":"en","type":"article","venue":"Linguistics Vanguard","topic":"Linguistic Variation and Morphology","field":"Social Sciences","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"University of Toronto","funders":"","keywords":"Complementizer; Preference; Variation (astronomy); Collocation (remote sensing); Vernacular; Persistence (discontinuity); Value (mathematics); Envelope (radar); Psychology; Epistemology; Linguistics; Philosophy; Computer science; Mathematics; Statistics; Syntax; Physics","retraction":null,"screen_n_in":null,"score":{"opus":0.02804845608191886,"gpt":0.3099939949485337,"spread":0.2819455388666148,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0007638871,0.0001610795,0.000269659,0.00003969014,0.0003429656,0.00004796504,0.0004508207,0.00009922077,0.0002645998],"category_scores_gemma":[0.01244406,0.000139218,0.00007709639,0.0003894497,0.000257112,0.00001423164,0.0001586956,0.0001955448,0.00009789645],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00002156411,"about_ca_system_score_gemma":0.0002199598,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00159103,"about_ca_topic_score_gemma":0.0009081907,"domain_scores_codex":[0.9982213,0.0002263022,0.0004279719,0.0003574238,0.0003988433,0.0003681387],"domain_scores_gemma":[0.9983094,0.0004091297,0.0001900448,0.0002732282,0.0004852276,0.0003329293],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0002916538,0.0001407308,0.01161803,0.0001833839,0.0001795836,0.0001325679,0.1884433,0.0003380516,0.001218847,0.3880974,0.4028535,0.006502924],"study_design_scores_gemma":[0.0003878697,0.000117142,0.0003894764,0.00003178684,0.0000758169,0.000001856586,0.007076262,0.0002603747,0.00007496194,0.002310644,0.9890319,0.0002419196],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"other","genre_gemma":"empirical","genre_scores_codex":[0.03137252,0.002300518,0.02302842,0.1595785,0.03298092,0.003655621,0.0006060594,0.0006510681,0.7458263],"genre_scores_gemma":[0.9762957,0.0002074687,0.003948093,0.01597594,0.002357602,0.00001509687,0.000008837073,0.00002132288,0.001169955],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9449232,"threshold_uncertainty_score":0.9958745,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4408918469","doi":"10.1515/lingvan-2024-0247","title":"Asymmetry in French speech-in-noise perception: the effects of native dialect and cross-dialectal exposure","year":2025,"lang":"en","type":"article","venue":"Linguistics Vanguard","topic":"Phonetics and Phonology Research","field":"Psychology","cited_by":1,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true},"ca_institutions":"","funders":"","keywords":"Asymmetry; Speech perception; Perception; Linguistics; Noise (video); Psychology; Speech recognition; Acoustics; Audiology; Computer science; Physics; Artificial intelligence; Philosophy; Medicine","retraction":null,"screen_n_in":null,"score":{"opus":0.009096022296645455,"gpt":0.3488086382358057,"spread":0.3397126159391602,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004396696,0.0001446297,0.0002755279,0.0003046545,0.00006516255,0.00002651913,0.0002509145,0.0002125927,0.0001049809],"category_scores_gemma":[0.004277814,0.0001217766,0.00005139774,0.0005435413,0.000365095,0.000007570172,0.0001191029,0.0005656026,0.00001612134],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004739103,"about_ca_system_score_gemma":0.00008354442,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001037759,"about_ca_topic_score_gemma":0.0008221577,"domain_scores_codex":[0.9986416,0.000272538,0.0003019073,0.0003132284,0.0001410955,0.0003296264],"domain_scores_gemma":[0.9979097,0.00149904,0.00006052223,0.0003293188,0.0001617102,0.00003977273],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.000277572,0.0004941331,0.9382898,0.0003207528,0.0001555577,0.00053003,0.02866623,0.00001174734,0.004446249,0.00815368,0.002331551,0.01632269],"study_design_scores_gemma":[0.001342691,0.0002657104,0.98935,0.0000893138,0.00001739466,0.000005848236,0.0002371616,0.0000618954,0.001055652,0.00669199,0.0007669915,0.0001153207],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9539638,0.002498159,0.00003604122,0.00006689267,0.002099889,0.0003516363,0.0000175719,0.00001272832,0.04095324],"genre_scores_gemma":[0.9975464,0.00009746466,0.0001416931,0.0001053258,0.0002788032,0.00003853462,0.000006094407,0.00001197989,0.001773714],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.05106023,"threshold_uncertainty_score":0.5121252,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4391967989","doi":"10.1515/lingvan-2023-0038","title":"Plains Cree Order as alternation","year":2024,"lang":"en","type":"article","venue":"Linguistics Vanguard","topic":"Lexicography and Language Studies","field":"Arts and Humanities","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"","keywords":"Alternation (linguistics); Phenomenon; Lexeme; Linguistics; Computer science; Mathematics; Epistemology; Philosophy","retraction":null,"screen_n_in":null,"score":{"opus":0.01925015524288478,"gpt":0.2696289642022423,"spread":0.2503788089593575,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.00008933042,0.0001346991,0.0001101511,0.0001331224,0.0002437765,0.0003370536,0.0001082332,0.0000327344,0.002142049],"category_scores_gemma":[0.0006585389,0.0001107911,0.0000898749,0.00006653988,0.0001319052,0.00003353253,0.0000377972,0.0001507209,0.0005650498],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0000173265,"about_ca_system_score_gemma":0.00002939318,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0004566753,"about_ca_topic_score_gemma":0.000889986,"domain_scores_codex":[0.999245,0.0000152208,0.000159204,0.0002021923,0.0001832266,0.0001951864],"domain_scores_gemma":[0.9994445,0.0001035857,0.00002559631,0.0001429014,0.0002429319,0.00004042612],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.000006446752,0.00002440416,0.00004760454,0.00007402873,0.0001192083,0.0001640162,0.05597092,0.000005011182,0.000004595963,0.9239141,0.01634548,0.003324175],"study_design_scores_gemma":[0.00008969878,0.00005627204,0.00002540705,0.00007582787,0.00005484371,0.000002261836,0.002833108,0.0002184848,0.00004485577,0.01635142,0.9800824,0.0001654287],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"other","genre_gemma":"empirical","genre_scores_codex":[0.009732145,0.003642394,0.0002396769,0.0002281562,0.007707779,0.0001163076,0.00008482033,0.0004645002,0.9777842],"genre_scores_gemma":[0.964896,0.00008258088,0.0002860661,0.0005232405,0.009631821,0.00001314048,0.00003906848,0.0000297694,0.02449827],"genre_candidate":"other","genre_consensus":null,"teacher_disagreement_score":0.9637369,"threshold_uncertainty_score":0.9987701,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4399982581","doi":"10.1515/lingvan-2023-0148","title":"The syntax of African American English borrowings in the Louisiana Creole tense-mood-aspect system","year":2024,"lang":"en","type":"article","venue":"Linguistics Vanguard","topic":"Linguistic Variation and Morphology","field":"Social Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"","funders":"Arts and Humanities Research Council; University of Oxford; Georgetown University; Louisiana State University; McGill University; University of Sussex; University of Pennsylvania","keywords":"Creole language; Linguistics; Syntax; English-based creole languages; History; Mood; American English; Psychology; Philosophy; Language assessment","retraction":null,"screen_n_in":null,"score":{"opus":0.01134392482593893,"gpt":0.2920378295530314,"spread":0.2806939047270925,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.002266879,0.0001431068,0.0002551332,0.0001113602,0.0004063964,0.0001907862,0.0006212485,0.00007982221,0.00002247803],"category_scores_gemma":[0.02232214,0.00009404821,0.0001195988,0.0008287581,0.0006162282,0.00001791745,0.00005170481,0.0002850856,0.00002529705],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001685111,"about_ca_system_score_gemma":0.0003476304,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.008080106,"about_ca_topic_score_gemma":0.004151959,"domain_scores_codex":[0.9978747,0.0004750171,0.000432877,0.0002787756,0.0005065915,0.0004320802],"domain_scores_gemma":[0.9968728,0.002049256,0.0001595494,0.0003797007,0.0004593423,0.00007928914],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00001239247,0.00002729264,0.0007426064,0.00004033275,0.00003488473,0.0001163373,0.06954073,0.000009778832,0.00000822583,0.9248581,0.003789199,0.000820104],"study_design_scores_gemma":[0.0001258233,0.00007491672,0.0009144733,0.0001002693,0.00008101869,0.000003106979,0.1029189,0.0004247799,0.00002630442,0.002544633,0.8926145,0.0001712543],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"other","genre_gemma":"empirical","genre_scores_codex":[0.01559403,0.001347279,0.0007931207,0.004649816,0.01427082,0.0008289873,0.0001072488,0.0004851526,0.9619235],"genre_scores_gemma":[0.9952581,0.00005582561,0.0003798831,0.000243178,0.003389483,0.00002225442,0.000004487781,0.00002052077,0.0006262484],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9796641,"threshold_uncertainty_score":0.9985252,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4415076690","doi":"10.1515/lingvan-2025-0160","title":"Linguistic approaches to fake news research are growing and maturing: commentary on a special issue","year":2025,"lang":"en","type":"article","venue":"Linguistics Vanguard","topic":"Misinformation and Its Impacts","field":"Social Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"Simon Fraser University","funders":"","keywords":"Interpretation (philosophy); Sophistication; Context (archaeology); Fake news; Flourishing; Corpus linguistics; Social media; Natural (archaeology)","retraction":null,"screen_n_in":null,"score":{"opus":0.1733501797692354,"gpt":0.3979844354813963,"spread":0.2246342557121608,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.001124986,0.0001417358,0.0002034211,0.0003699096,0.0008994539,0.0003384449,0.0002822066,0.000116965,0.0001408545],"category_scores_gemma":[0.01762085,0.0001446581,0.00004113287,0.0005261741,0.0001709623,0.00004165611,0.0001510526,0.0003857475,0.0001297398],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001822795,"about_ca_system_score_gemma":0.0001312496,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0008938341,"about_ca_topic_score_gemma":0.001028936,"domain_scores_codex":[0.9982036,0.0001905382,0.0002759234,0.0002542699,0.0005711431,0.0005045187],"domain_scores_gemma":[0.9986715,0.0004944525,0.00005950107,0.0002537745,0.0002404172,0.0002803561],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.000120675,0.0001208894,0.001226687,0.0001391367,0.00003880336,0.00003472727,0.1352185,0.00007235836,0.000003599893,0.309214,0.5311844,0.02262636],"study_design_scores_gemma":[0.000227173,0.00005355745,0.0008448716,0.0001893704,0.00001234606,1.91452e-7,0.02576681,0.0000770524,0.00008004814,0.004333822,0.9682645,0.0001502701],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"other","genre_gemma":"empirical","genre_scores_codex":[0.01545509,0.00008377829,0.000120756,0.01546881,0.003230825,0.0005749637,0.00002344752,0.0001028078,0.9649395],"genre_scores_gemma":[0.9413716,0.00007828943,0.002670331,0.01596939,0.02777234,0.00001019,0.0000186795,0.0000246433,0.01208459],"genre_candidate":"other","genre_consensus":null,"teacher_disagreement_score":0.9528549,"threshold_uncertainty_score":0.9906542,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4393936388","doi":"10.1515/lingvan-2023-0087","title":"A diachronic consequence of intransitivity: structural underspecification and processing biases in Old French","year":2024,"lang":"en","type":"article","venue":"Linguistics Vanguard","topic":"Categorization, perception, and language","field":"Psychology","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"","keywords":"Underspecification; Linguistics; Verb; Grammar; Representation (politics); Object (grammar); Parsing; Corollary; Optimality theory; Computer science; History; Philosophy; Artificial intelligence; Phonology; Mathematics; Political science","retraction":null,"screen_n_in":null,"score":{"opus":0.03199731922509109,"gpt":0.3445328847187323,"spread":0.3125355654936412,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0001520218,0.000107237,0.0001478895,0.0001636683,0.00003948822,0.00004976011,0.00005856972,0.00008278988,0.0002706501],"category_scores_gemma":[0.0004090274,0.0001064565,0.00002591936,0.0002782196,0.0001666656,0.00003083569,0.00001017032,0.0001339094,0.000009358833],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00006269169,"about_ca_system_score_gemma":0.0000924655,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001148291,"about_ca_topic_score_gemma":0.0008297333,"domain_scores_codex":[0.9991595,0.00005469101,0.0002410682,0.0002730598,0.0001004635,0.0001712673],"domain_scores_gemma":[0.9994883,0.0001704977,0.0000486875,0.000150777,0.0001046954,0.0000370741],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"observational","study_design_scores_codex":[0.0001361041,0.0003212058,0.05997428,0.002448688,0.0001952242,0.001256579,0.313114,0.0005198246,0.0185733,0.4578576,0.0009877612,0.1446154],"study_design_scores_gemma":[0.002701849,0.0005099822,0.8692546,0.002152876,0.0004375334,0.0005033861,0.02867957,0.03303038,0.00211906,0.04997373,0.008919956,0.001717008],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.985525,0.00595664,0.003158437,0.00006921715,0.0008657805,0.0001709924,0.00005289195,0.00008737145,0.004113702],"genre_scores_gemma":[0.9987323,0.00008003675,0.0003252906,0.00003053706,0.0004281252,0.000008247602,0.00006196219,0.00001696968,0.0003165102],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.8092804,"threshold_uncertainty_score":0.4341171,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4377224620","doi":"10.1515/lingvan-2021-0143","title":"How did COVID-19 impact the use of Japanese complex words with <i>masuku</i> ‘mask’ in 2020?","year":2023,"lang":"en","type":"article","venue":"Linguistics Vanguard","topic":"Language, Metaphor, and Cognition","field":"Psychology","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"York University","funders":"","keywords":"Newspaper; Compounding; Coronavirus disease 2019 (COVID-19); Sentence; Computer science; Advertising; History; Artificial intelligence; Sociology; Media studies; Business","retraction":null,"screen_n_in":null,"score":{"opus":0.08374675119605893,"gpt":0.3525271364932683,"spread":0.2687803852972094,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004240892,0.0002637243,0.0003897778,0.0001921322,0.00009984563,0.0000966887,0.0003096885,0.0001336565,0.0005314802],"category_scores_gemma":[0.003707802,0.0001703049,0.0001431631,0.00104968,0.000217979,0.0000312021,0.00006988362,0.0003150542,0.0001081775],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00006775927,"about_ca_system_score_gemma":0.0001016858,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.002650514,"about_ca_topic_score_gemma":0.001237444,"domain_scores_codex":[0.9981573,0.0002733544,0.0003286381,0.0003885968,0.0003591212,0.0004929906],"domain_scores_gemma":[0.9977918,0.000923968,0.0002065793,0.0007049465,0.000197583,0.0001751747],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"qualitative","study_design_gemma":"not_applicable","study_design_scores_codex":[0.005700879,0.001890371,0.1306472,0.0008534758,0.002095821,0.008843275,0.423337,0.004210298,0.002649274,0.01876104,0.3853714,0.01563997],"study_design_scores_gemma":[0.01078985,0.00208119,0.1346315,0.0002912442,0.0009120122,0.0002685778,0.1108335,0.002423952,0.0002672429,0.005456034,0.7297949,0.002249863],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9715118,0.0004870901,0.0008815865,0.001654945,0.002215734,0.001093952,0.0007637817,0.0003871828,0.02100394],"genre_scores_gemma":[0.9912972,0.00003199095,0.0003413936,0.00153323,0.0007466527,0.00005022069,0.000260776,0.00005361988,0.005684843],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3444236,"threshold_uncertainty_score":0.6944832,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4413796994","doi":"10.1515/lingvan-2024-0201","title":"Instance memory models as a general computational framework for exploring language processing: bringing the lexicon to life","year":2025,"lang":"en","type":"article","venue":"Linguistics Vanguard","topic":"Topic Modeling","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Manitoba; McGill University","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Lexicon; Computer science; Cognitive science; Natural language processing; Linguistics; Artificial intelligence; Psychology; Cognitive psychology; Philosophy","retraction":null,"screen_n_in":null,"score":{"opus":0.0570391012999994,"gpt":0.3250468555101352,"spread":0.2680077542101358,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003188622,0.0001843908,0.0002034304,0.000140087,0.0003373503,0.0003174124,0.0009266119,0.00006768251,0.000002023943],"category_scores_gemma":[0.002704585,0.0001681083,0.00007824817,0.0004017673,0.00002838017,0.0001106808,0.0003171467,0.0002391138,0.000007619726],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00008102937,"about_ca_system_score_gemma":0.0003859309,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0000334105,"about_ca_topic_score_gemma":0.000008530047,"domain_scores_codex":[0.9984484,0.0000302773,0.0003380709,0.0004948474,0.0003045675,0.0003838048],"domain_scores_gemma":[0.9985638,0.000363892,0.00009311986,0.0005319502,0.00033649,0.0001107157],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000009468894,0.00001474879,0.0000155182,0.00005661573,0.00001486251,0.000006726599,0.006241598,0.3567054,0.00001638979,0.6132261,0.0001834432,0.0235092],"study_design_scores_gemma":[0.0001683129,0.00001614248,0.00001813674,0.0001861416,0.00001076702,0.000001300059,0.0003468059,0.7930908,0.0002707127,0.2033974,0.002329005,0.0001644365],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01300169,0.0007404099,0.9774246,0.00137056,0.001632374,0.0003840114,0.0000047004,0.0002519868,0.005189619],"genre_scores_gemma":[0.5199715,0.000002206049,0.4757912,0.003096508,0.0007191008,0.00008999118,0.000002072244,0.00001334622,0.0003140593],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.5069697,"threshold_uncertainty_score":0.6855258,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4403221822","doi":"10.1515/lingvan-2024-0091","title":"Using constructed languages to introduce and teach linguistics","year":2024,"lang":"en","type":"article","venue":"Linguistics Vanguard","topic":"EFL/ESL Teaching and Learning","field":"Arts and Humanities","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Alberta","funders":"","keywords":"Linguistics; Applied linguistics; Computer science; Quantitative linguistics; Media linguistics; Programming language; Sociology; Philosophy","retraction":null,"screen_n_in":null,"score":{"opus":0.03780478594028577,"gpt":0.3182800522976544,"spread":0.2804752663573687,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.0003335598,0.000218401,0.0002355524,0.0002123214,0.000386712,0.0006459425,0.0001279174,0.00006549455,0.0003479569],"category_scores_gemma":[0.008518349,0.0002046898,0.00005368876,0.00007100891,0.0001706948,0.00001535298,0.00009176387,0.000505003,0.0001036932],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004656304,"about_ca_system_score_gemma":0.00006395435,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0004775954,"about_ca_topic_score_gemma":0.0001152552,"domain_scores_codex":[0.998753,0.00006878886,0.0002706344,0.0003881778,0.0001918469,0.0003275518],"domain_scores_gemma":[0.999052,0.0002309742,0.00004399645,0.0002301116,0.0003045125,0.0001383966],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","study_design_scores_codex":[0.00001315175,0.00002618079,0.000544761,0.0002473507,0.00009661172,0.000193425,0.06168822,0.0001414636,0.0002295978,0.9194338,0.01023174,0.007153722],"study_design_scores_gemma":[0.0001018868,0.00006545037,0.00002874467,0.0002634303,0.0001089726,0.00001315559,0.006138844,0.003955021,0.00007689555,0.001394881,0.9875261,0.0003266148],"study_design_candidate":"not_applicable","study_design_consensus":null,"genre_codex":"other","genre_gemma":"empirical","genre_scores_codex":[0.2887845,0.008141188,0.002970899,0.0003828463,0.02424932,0.0005571583,0.0002812912,0.002234078,0.6723987],"genre_scores_gemma":[0.9658515,0.00001146131,0.01182063,0.0002901301,0.01673294,0.000003029898,0.00002395342,0.00006373489,0.005202632],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9772944,"threshold_uncertainty_score":0.9998333,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null},{"id":"W4392656786","doi":"10.1515/lingvan-2023-0051","title":"The role of syntax in hashtag popularity","year":2024,"lang":"en","type":"article","venue":"Linguistics Vanguard","topic":"Sentiment Analysis and Opinion Mining","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"routes":{"ca_aff":true,"ca_fund":false,"ca_venue":false,"about_ca":false},"ca_institutions":"University of Toronto","funders":"","keywords":"Popularity; Computer science; Syntax; Natural language processing; Semantics (computer science); CLARITY; Artificial intelligence; Linguistics; Information retrieval; Psychology; Programming language; Social psychology","retraction":null,"screen_n_in":null,"score":{"opus":0.008434858255253173,"gpt":0.2652054880703701,"spread":0.2567706298151169,"validation_status":"score_only:v0-immature-baseline"},"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0003867327,0.00007013111,0.0001123433,0.00009329064,0.00006107939,0.0001695168,0.0004564379,0.00003257855,0.00000754437],"category_scores_gemma":[0.0005611698,0.00005148026,0.00007898675,0.0003858279,0.00002703525,0.00002472894,0.000133961,0.0001189901,0.00002165276],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00001917464,"about_ca_system_score_gemma":0.00003927709,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001014086,"about_ca_topic_score_gemma":0.0000596373,"domain_scores_codex":[0.9991722,0.0000340539,0.0002446618,0.0001901461,0.000201598,0.0001573864],"domain_scores_gemma":[0.9993324,0.0001953415,0.00004232527,0.000330155,0.00007146242,0.00002832924],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000002267612,0.00002879739,0.01215472,0.0000187027,0.00004032979,0.00003380275,0.001106389,0.00007760841,0.0001732317,0.9390206,0.0004703933,0.04687311],"study_design_scores_gemma":[0.0001126035,0.0000407606,0.002396614,0.0001577564,0.00002789021,0.000002452476,0.0003417051,0.5399159,0.003188165,0.06999846,0.3836102,0.0002075644],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"other","genre_gemma":"empirical","genre_scores_codex":[0.1647999,0.130706,0.3210594,0.003694265,0.03694276,0.001235596,0.00005558529,0.001444555,0.3400619],"genre_scores_gemma":[0.9918298,0.00005153603,0.007541629,0.00002373233,0.0002505164,0.000002422539,0.000002242727,0.000005311922,0.0002927817],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8690222,"threshold_uncertainty_score":0.2099304,"prediction_status":"machine_predicted_unvalidated"},"labels":[],"label_agreement":null}]}