{"meta":{"query_hash":"28b6cc5f6b26","filters":{"venue":"Studies in corpus linguistics"},"cohort_total":12,"direct_labels_cover":0,"predictions_cover":12,"exported":12,"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/28b6cc5f6b26","api":"https://metacan.xera.ac/api/v1/cohort?venue=Studies+in+corpus+linguistics"},"results":[{"id":"W2476135460","doi":"10.1075/scl.67.03dar","title":"At the crossroads of change","year":2015,"lang":"en","type":"book-chapter","venue":"Studies in corpus linguistics","topic":"Linguistic Variation and Morphology","field":"Social Sciences","cited_by":7,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"University of Victoria","funders":"","keywords":"Possession (linguistics); Newspaper; Negation; History; Language change; Geography; Linguistics; Media studies; Sociology","score_opus":0.3719527862548491,"score_gpt":0.4396158681615569,"score_spread":0.06766308190670778,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2476135460","genre_codex":"other","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.1683137,0.010089276,0.0023155308,0.01406235,0.00030733412,0.000032260486,0.0001555488,0.00004358916,0.80468047],"genre_scores_gemma":[0.96492183,0.0019859762,0.0005080819,0.0006409599,0.000048097394,0.000012007733,0.00005444829,0.00004956751,0.031779],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"qualitative","domain_scores_codex":[0.99853134,0.00040509916,0.000036408645,0.00026465874,0.00039060347,0.00037186433],"domain_scores_gemma":[0.9992656,0.0002826691,0.000052592142,0.00010198112,0.00020131118,0.0000959142],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0011301113,0.00018283234,0.00026393094,0.0009190198,0.007574129,0.005123753,0.00073592796,0.00079742644,0.0061782515],"category_scores_gemma":[0.0023257323,0.00018668249,0.00013535909,0.0017242464,0.0202711,0.004347155,0.002590972,0.0021268264,0.00033169735],"study_design_candidate":"qualitative","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000031839914,0.0000093660165,0.0024962781,0.00007522395,0.0000070536103,0.00034064302,0.19234572,0.00009748711,0.00095379935,0.7517376,0.0062771006,0.045627862],"study_design_scores_gemma":[0.0000072189537,0.000020097048,0.02786783,0.0003270533,0.00001474305,0.00033352376,0.1438862,0.00019724046,0.0004941771,0.09538342,0.73143846,0.000030081272],"about_ca_topic_score_codex":0.4657509,"about_ca_topic_score_gemma":0.5786963,"teacher_disagreement_score":0.4657509,"about_ca_system_score_codex":0.020644162,"about_ca_system_score_gemma":0.009768841,"threshold_uncertainty_score":0.92607987},"labels":[],"label_agreement":null},{"id":"W2485877726","doi":"10.1075/scl.67.07mey","title":"Passives of so-called ‘ditransitives’ in nineteenth century and present-day Canadian English","year":2015,"lang":"en","type":"book-chapter","venue":"Studies in corpus linguistics","topic":"Linguistic Variation and Morphology","field":"Social Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"","keywords":"Complement (music); Dominance (genetics); Linguistics; Matching (statistics); Literature; History; Art; Philosophy; Mathematics; Chemistry","score_opus":0.0688148870694358,"score_gpt":0.3428270208740393,"score_spread":0.27401213380460354,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2485877726","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.7528069,0.0036843235,0.0069807013,0.00083456666,0.0001396061,0.0001312266,0.0020436994,0.00018953664,0.2331894],"genre_scores_gemma":[0.9754219,0.0010617567,0.0021577955,0.00009383724,0.000022476135,0.00003653031,0.00089552376,0.00013703146,0.020173196],"study_design_codex":"qualitative","study_design_gemma":"observational","domain_scores_codex":[0.9989201,0.0001635683,0.000065401895,0.00023418796,0.00036953748,0.00024729164],"domain_scores_gemma":[0.99806696,0.00075096043,0.00021251233,0.00018784971,0.00066819007,0.00011353669],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0013572864,0.0005914583,0.0003372779,0.003264055,0.005112742,0.0031894192,0.001035586,0.000568107,0.0067861574],"category_scores_gemma":[0.0027049605,0.00039138214,0.00029298122,0.003943217,0.0066662314,0.0025832155,0.0014360887,0.0012581929,0.0005972147],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0002759134,0.00005297547,0.018844208,0.00070318475,0.000027681015,0.0011371869,0.5001956,0.00027383384,0.021672962,0.3799916,0.0062497826,0.07057504],"study_design_scores_gemma":[0.0000464411,0.00008696436,0.16955666,0.0005771291,0.00013759885,0.0028555624,0.13346995,0.0011931964,0.015308955,0.014292606,0.6622251,0.00024976145],"about_ca_topic_score_codex":0.7847671,"about_ca_topic_score_gemma":0.87858886,"teacher_disagreement_score":0.21523291,"about_ca_system_score_codex":0.019525107,"about_ca_system_score_gemma":0.016194962,"threshold_uncertainty_score":0.43300086},"labels":[],"label_agreement":null},{"id":"W2498427696","doi":"10.1075/scl.13.11bri","title":"Subject clitics in English","year":2004,"lang":"en","type":"book-chapter","venue":"Studies in corpus linguistics","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":"University of British Columbia","funders":"","keywords":"Subject (documents); History; Linguistics; Computer science; Philosophy; World Wide Web","score_opus":0.08111191884132003,"score_gpt":0.36748845532521995,"score_spread":0.28637653648389993,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2498427696","genre_codex":"other","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.12028294,0.01666976,0.0193772,0.0028725436,0.0006732281,0.00007600276,0.00031628643,0.0003067899,0.8394252],"genre_scores_gemma":[0.885699,0.0071470705,0.005086274,0.00056068663,0.00043493632,0.00006594106,0.00029171462,0.0002377266,0.10047663],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"observational","domain_scores_codex":[0.9995964,0.00011303223,0.000035976987,0.000092754504,0.000109724504,0.00005209998],"domain_scores_gemma":[0.999569,0.00017827179,0.000040823998,0.000059414888,0.00012305324,0.000029360766],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00060954224,0.0004428656,0.0002955753,0.0011919144,0.002025686,0.0026994299,0.0004129158,0.00058108114,0.0042344173],"category_scores_gemma":[0.0009414805,0.00023784972,0.0002011096,0.0012012922,0.0036645632,0.0033605497,0.0013414117,0.0010682915,0.0009961866],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000023920687,0.00002583667,0.0013013704,0.00014923776,0.0000050160143,0.00020390878,0.016914567,0.00010945106,0.001295559,0.93759197,0.0041992506,0.038179986],"study_design_scores_gemma":[0.000027555929,0.00007241516,0.0029717535,0.00024007529,0.000019635556,0.00081421237,0.005859093,0.00052594923,0.002129796,0.21793647,0.7693749,0.000028125642],"about_ca_topic_score_codex":0.006235752,"about_ca_topic_score_gemma":0.0074025304,"teacher_disagreement_score":0.006235752,"about_ca_system_score_codex":0.0019064632,"about_ca_system_score_gemma":0.0014217334,"threshold_uncertainty_score":0.014165461},"labels":[],"label_agreement":null},{"id":"W2505386209","doi":"10.1075/scl.13.04tag","title":"Have to, gotta, must?","year":2004,"lang":"en","type":"book-chapter","venue":"Studies in corpus linguistics","topic":"Linguistic Variation and Morphology","field":"Social Sciences","cited_by":37,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of Toronto","funders":"","keywords":"History","score_opus":0.1029219713045221,"score_gpt":0.39003760022818473,"score_spread":0.2871156289236626,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W2505386209","genre_codex":"other","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.24907283,0.008206421,0.014382006,0.0052206307,0.0003831003,0.000034777062,0.00045534456,0.00011587933,0.722129],"genre_scores_gemma":[0.8913478,0.0017751984,0.004909741,0.00044996353,0.000072899136,0.000017503335,0.0003219646,0.000082180755,0.101022825],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.9996469,0.00014475777,0.000021244872,0.000048836017,0.000106020634,0.000032247193],"domain_scores_gemma":[0.9995826,0.00020077884,0.00006912702,0.000035788005,0.00006904453,0.000042753793],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00046874117,0.0001984809,0.00016069759,0.00059404894,0.0009770003,0.0015130795,0.00020636107,0.00029091898,0.005709479],"category_scores_gemma":[0.0009469832,0.00012410803,0.00011646423,0.0008007055,0.0027366115,0.0017699972,0.0004260344,0.0006949022,0.00076366804],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00013361718,0.00003369632,0.024167404,0.00044814066,0.000028798802,0.00048338465,0.08900049,0.0001657889,0.0070258514,0.6133179,0.049592346,0.21560265],"study_design_scores_gemma":[0.000011361143,0.000102952356,0.13163754,0.000372961,0.000039545368,0.0036081492,0.050377775,0.0009024533,0.002252239,0.11662942,0.6940152,0.000050366543],"about_ca_topic_score_codex":0.0060932487,"about_ca_topic_score_gemma":0.027154014,"teacher_disagreement_score":0.0060932487,"about_ca_system_score_codex":0.0009091573,"about_ca_system_score_gemma":0.00060023996,"threshold_uncertainty_score":0.01910007},"labels":[],"label_agreement":null},{"id":"W3011686062","doi":"10.1075/scl.96.10bri","title":"The development and pragmatic function of a non-inference marker","year":2020,"lang":"en","type":"book-chapter","venue":"Studies in corpus linguistics","topic":"Language, Metaphor, and Cognition","field":"Psychology","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"University of British Columbia","funders":"","keywords":"Grammaticalization; Inference; Linguistics; Divergence (linguistics); Function (biology); Reading (process); Context (archaeology); Drama; History; Literature; Philosophy; Art; Epistemology","score_opus":0.05131447241279757,"score_gpt":0.32952148696499123,"score_spread":0.27820701455219365,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W3011686062","genre_codex":"other","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.27017286,0.0015041508,0.1330205,0.0035336881,0.00048360377,0.00013546385,0.0002804115,0.0007329335,0.5901364],"genre_scores_gemma":[0.9583995,0.00025449664,0.021294544,0.0002444592,0.00008992652,0.000042313965,0.00022313987,0.00063560816,0.018816028],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"qualitative","domain_scores_codex":[0.9979431,0.00089250214,0.00008108748,0.00050620694,0.0004475789,0.00012956036],"domain_scores_gemma":[0.9946089,0.0031995452,0.00030669954,0.00088174816,0.00077632925,0.00022676113],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002618549,0.0003512549,0.0004257957,0.0011079658,0.001531159,0.0043108854,0.0011797419,0.0013669847,0.011581991],"category_scores_gemma":[0.0069604036,0.00057750853,0.00027033684,0.00060992053,0.0049586557,0.00840972,0.0031743618,0.0029388417,0.0026223399],"study_design_candidate":"qualitative","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00009291238,0.00006110964,0.0027622033,0.00017467586,0.00001152753,0.0010227528,0.024730613,0.00017171098,0.023993405,0.9065619,0.0021477612,0.038269393],"study_design_scores_gemma":[0.000070238646,0.0003944981,0.026331887,0.0006942905,0.00012949332,0.0072133373,0.02412099,0.013847941,0.0568127,0.4463113,0.42390642,0.00016688887],"about_ca_topic_score_codex":0.0016428299,"about_ca_topic_score_gemma":0.0018862105,"teacher_disagreement_score":0.011581991,"about_ca_system_score_codex":0.001911995,"about_ca_system_score_gemma":0.001304219,"threshold_uncertainty_score":0.038745582},"labels":[],"label_agreement":null},{"id":"W4379415672","doi":"10.1075/scl.109.09gei","title":"DocuScope Write &amp; Audit as an early feedback machine ingenre-based writing","year":2023,"lang":"en","type":"book-chapter","venue":"Studies in corpus linguistics","topic":"Discourse Analysis in Language Studies","field":"Arts and Humanities","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Simon Fraser University","funders":"","keywords":"Audit; Context (archaeology); Computer science; Focus (optics); Report writing; Management; History; Library science","score_opus":0.09666553958224774,"score_gpt":0.34309553982103885,"score_spread":0.2464300002387911,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4379415672","genre_codex":"methods","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.22855201,0.0014142839,0.47826776,0.014620636,0.0026699998,0.0018949861,0.0015067915,0.03433745,0.23673613],"genre_scores_gemma":[0.41793206,0.000542852,0.44689262,0.001202483,0.0003416448,0.0010034242,0.0009888563,0.002768425,0.12832767],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.9891222,0.00592663,0.00045398792,0.0005940519,0.0036100033,0.00029313902],"domain_scores_gemma":[0.94208294,0.036519855,0.0020395198,0.008182109,0.008409856,0.0027656746],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.012830124,0.00053890556,0.00047115303,0.0025683467,0.0014512686,0.005258859,0.0014465593,0.00076389115,0.010506874],"category_scores_gemma":[0.041545916,0.00039310817,0.00019836171,0.0013792628,0.0021196543,0.0030698269,0.0036206394,0.0018086099,0.0038666003],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0003435447,0.00028308996,0.00452586,0.0003797455,0.000006787634,0.00036440897,0.020312956,0.0008401977,0.012082199,0.020039916,0.075069875,0.86575127],"study_design_scores_gemma":[0.000183922,0.0011815473,0.013906313,0.0012173412,0.000032687967,0.0018523562,0.0165822,0.018644283,0.049853463,0.025798578,0.8704964,0.00025091495],"about_ca_topic_score_codex":0.0008312684,"about_ca_topic_score_gemma":0.0022292815,"teacher_disagreement_score":0.012830124,"about_ca_system_score_codex":0.0012961279,"about_ca_system_score_gemma":0.0029967413,"threshold_uncertainty_score":0.067852974},"labels":[],"label_agreement":null},{"id":"W4385228274","doi":"10.1075/scl.110.03cot","title":"The <i>Phonologie du français contemporain</i> projectin Quebec","year":2023,"lang":"en","type":"book-chapter","venue":"Studies in corpus linguistics","topic":"Linguistic Variation and Morphology","field":"Social Sciences","cited_by":3,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"Université de Sherbrooke","funders":"","keywords":"Pronunciation; Variety (cybernetics); Representation (politics); Linguistics; French; Subject (documents); Relevance (law); Variation (astronomy); History; Computer science; Artificial intelligence; Philosophy; Political science; Library science","score_opus":0.11830890903078506,"score_gpt":0.36347782692427305,"score_spread":0.245168917893488,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4385228274","genre_codex":"other","genre_gemma":"other","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"other","genre_consensus":"other","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.33252457,0.060653903,0.014309579,0.008183979,0.0011575411,0.00025956894,0.0786933,0.00091618154,0.5033013],"genre_scores_gemma":[0.8371694,0.010447859,0.0085687665,0.00074226013,0.00017604283,0.00019651774,0.01930153,0.00043925102,0.12295839],"study_design_codex":"design_other","study_design_gemma":"not_applicable","domain_scores_codex":[0.9995627,0.00009734306,0.000015088747,0.00010859786,0.00014001473,0.00007623942],"domain_scores_gemma":[0.99949074,0.000138722,0.00003316041,0.00004374921,0.00025312696,0.00004053544],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00055283686,0.0005129011,0.00027437802,0.002605048,0.0031423923,0.0026034291,0.00048260656,0.00044745833,0.0155776795],"category_scores_gemma":[0.0014417116,0.00018340189,0.00017352594,0.0051136054,0.0015312765,0.00077734195,0.000491448,0.00079876155,0.0010551935],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":true,"about_ca_topic_consensus":true,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00027862503,0.00004991757,0.033225134,0.0010688369,0.00008515365,0.0009320411,0.048874304,0.0017785609,0.009506048,0.18335022,0.2804515,0.44039968],"study_design_scores_gemma":[0.000009711881,0.000013120877,0.08584839,0.00021241354,0.0000139852555,0.0003539819,0.005577195,0.0004007523,0.0010003466,0.00115954,0.9053744,0.000036120025],"about_ca_topic_score_codex":0.9680604,"about_ca_topic_score_gemma":0.972128,"teacher_disagreement_score":0.031939626,"about_ca_system_score_codex":0.022521447,"about_ca_system_score_gemma":0.011737171,"threshold_uncertainty_score":0.16340536},"labels":[],"label_agreement":null},{"id":"W4387736150","doi":"10.1075/scl.113.06lab","title":"run away!","year":2023,"lang":"en","type":"book-chapter","venue":"Studies in corpus linguistics","topic":"Categorization, perception, and language","field":"Psychology","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"","keywords":"Vocabulary; Verb; Path (computing); Motion (physics); Word (group theory); Computer science; Expression (computer science); Linguistics; Translation (biology); Natural language processing; Artificial intelligence; Chemistry; Philosophy; Programming language","score_opus":0.12279469422098577,"score_gpt":0.3941716250231471,"score_spread":0.27137693080216135,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4387736150","genre_codex":"other","genre_gemma":"other","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"other","genre_consensus":"other","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.005486526,0.0049038264,0.004749823,0.008372913,0.004945996,0.000053063344,0.00084084057,0.0007439041,0.9699033],"genre_scores_gemma":[0.03654902,0.0029548935,0.003198625,0.0030870289,0.00093574164,0.000052393,0.0008149519,0.0010960782,0.9513114],"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","domain_scores_codex":[0.99981266,0.000049186663,0.0000063354096,0.000044786713,0.00005662128,0.00003040344],"domain_scores_gemma":[0.99981517,0.000052725605,0.00001091915,0.000022601625,0.000059471757,0.00003914376],"candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.00027510803,0.00047242775,0.00024591893,0.0007366212,0.0016980934,0.002272202,0.00039444095,0.0006000172,0.13076478],"category_scores_gemma":[0.00079286314,0.0001219839,0.0003131781,0.0007042875,0.0010093575,0.002412711,0.001398878,0.0014385832,0.061070703],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00005108564,0.000027684311,0.0007068001,0.0003065416,0.000006135724,0.0002663093,0.010789961,0.000050858915,0.0010367953,0.09728037,0.7335107,0.15596679],"study_design_scores_gemma":[0.0000018738267,0.0000070075203,0.0004112055,0.00003588543,0.0000011072311,0.00012036638,0.0008984107,0.000016286229,0.00006915825,0.0012814285,0.99715436,0.0000029101661],"about_ca_topic_score_codex":0.004159219,"about_ca_topic_score_gemma":0.011631679,"teacher_disagreement_score":0.8692352,"about_ca_system_score_codex":0.000741878,"about_ca_system_score_gemma":0.00060887105,"threshold_uncertainty_score":0},"labels":[],"label_agreement":null},{"id":"W4405077419","doi":"10.1075/scl.118.09mil","title":"Modeling fine-grained sociolinguistic variation","year":2024,"lang":"en","type":"book-chapter","venue":"Studies in corpus linguistics","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"","keywords":"Variation (astronomy); Computer science; Linguistics; Astrophysics; Physics; Philosophy","score_opus":0.05598935463493161,"score_gpt":0.3361349588186589,"score_spread":0.2801456041837273,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4405077419","genre_codex":"empirical","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.75732505,0.0004555604,0.22903244,0.0005382867,0.000045098415,0.0000856229,0.0037187503,0.000646351,0.008152733],"genre_scores_gemma":[0.9693927,0.00008760555,0.02744446,0.000045462642,0.000010073519,0.000058318154,0.001587796,0.00009273202,0.0012807489],"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","domain_scores_codex":[0.999673,0.00012953972,0.000008943383,0.00009521322,0.000058544312,0.00003484916],"domain_scores_gemma":[0.9977036,0.0015995294,0.00017194256,0.0002308325,0.00023889124,0.000055133314],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0007324419,0.00028983486,0.00026752337,0.0011016484,0.000492536,0.001177925,0.0005789989,0.00032663002,0.0023319698],"category_scores_gemma":[0.004581069,0.00021722006,0.0002568989,0.001230131,0.0009451016,0.0010101589,0.00068837084,0.0007475683,0.0003378739],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00027826885,0.0002337101,0.15381843,0.00053955143,0.00025903023,0.0009920141,0.007373995,0.50498366,0.042155694,0.11189312,0.011900185,0.16557235],"study_design_scores_gemma":[0.0000067716896,0.000022232227,0.050261576,0.000027258098,0.000012805292,0.000047298658,0.00078106456,0.91679496,0.0016198911,0.023725245,0.0066731516,0.00002776742],"about_ca_topic_score_codex":0.15589188,"about_ca_topic_score_gemma":0.2614704,"teacher_disagreement_score":0.15589188,"about_ca_system_score_codex":0.0019029025,"about_ca_system_score_gemma":0.0011891887,"threshold_uncertainty_score":0.309969},"labels":[],"label_agreement":null},{"id":"W4416401069","doi":"10.1075/scl.122.14joa","title":"Examining coherence and cohesion errors in writing Catalan as an additional language","year":2025,"lang":"en","type":"book-chapter","venue":"Studies in corpus linguistics","topic":"Second Language Acquisition and Learning","field":"Psychology","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":"Université de Montréal; Université TÉLUQ","funders":"","keywords":"Catalan; Cohesion (chemistry); Coherence (philosophical gambling strategy); Perception; Error analysis; Language proficiency","score_opus":0.06872773850521371,"score_gpt":0.37796157656266277,"score_spread":0.3092338380574491,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4416401069","genre_codex":"empirical","genre_gemma":"empirical","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"empirical","genre_consensus":"empirical","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.99568933,0.00030447755,0.0014391975,0.00007587123,0.00002374891,0.000044564298,0.00012349458,0.000034050387,0.0022651257],"genre_scores_gemma":[0.9915249,0.00023632502,0.0040818374,0.000040511266,0.000016084681,0.00007451773,0.00043374233,0.000039962466,0.0035521649],"study_design_codex":"design_other","study_design_gemma":"observational","domain_scores_codex":[0.99757904,0.0012930046,0.00017422973,0.00028094256,0.0005761621,0.000096625416],"domain_scores_gemma":[0.9706303,0.023241023,0.0023134723,0.00093031453,0.0026081838,0.00027677033],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0021330502,0.00044204103,0.00045779612,0.00207881,0.0006981625,0.0013963929,0.0005098704,0.00047697985,0.0018608973],"category_scores_gemma":[0.014305502,0.0001454971,0.00011817653,0.0015346484,0.0008316562,0.0006372639,0.0009125367,0.0004391889,0.0002893244],"study_design_candidate":"observational","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.0009087032,0.0013177771,0.25191587,0.0021762773,0.00011969052,0.005117238,0.20157124,0.0032166804,0.075488426,0.001999594,0.0066215633,0.44954678],"study_design_scores_gemma":[0.000047771133,0.00089562475,0.84477913,0.0004306107,0.00006881759,0.0028511062,0.07772949,0.0071837143,0.028300032,0.0012500685,0.03637969,0.00008402649],"about_ca_topic_score_codex":0.004585296,"about_ca_topic_score_gemma":0.00885658,"teacher_disagreement_score":0.004585296,"about_ca_system_score_codex":0.0009392841,"about_ca_system_score_gemma":0.00075572345,"threshold_uncertainty_score":0.011280835},"labels":[],"label_agreement":null},{"id":"W4416401093","doi":"10.1075/scl.122.04gag","title":"Crosslinguistic influences and L3/Ln teaching","year":2025,"lang":"en","type":"book-chapter","venue":"Studies in corpus linguistics","topic":"EFL/ESL Teaching and Learning","field":"Arts and Humanities","cited_by":1,"is_retracted":false,"has_abstract":true,"route_ca_aff":true,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":false,"ca_institutions":"Université TÉLUQ","funders":"","keywords":"Set (abstract data type); Language proficiency; Error analysis; Catalan; German; Language acquisition; Romance languages","score_opus":0.09240474182370162,"score_gpt":0.3478189198680716,"score_spread":0.25541417804437,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W4416401093","genre_codex":"empirical","genre_gemma":"methods","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"methods","genre_consensus":null,"domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.9951892,0.00014058687,0.00021443605,0.000047749098,0.000004206874,0.000005728823,0.000030702737,0.00001242961,0.0043549836],"genre_scores_gemma":[0.99826056,0.00006471935,0.00021656914,0.000022908609,0.0000022080624,0.0000063320626,0.000041695435,0.000010011753,0.0013749225],"study_design_codex":"observational","study_design_gemma":"not_applicable","domain_scores_codex":[0.9979656,0.00079415855,0.0001090356,0.0003592278,0.000620785,0.00015119874],"domain_scores_gemma":[0.9877243,0.0076710014,0.0020154435,0.00060765765,0.0012785068,0.0007031369],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0014182343,0.000323262,0.00031618963,0.0007878127,0.0005810479,0.0019269264,0.00043928815,0.00032099188,0.0038349184],"category_scores_gemma":[0.007941783,0.00011583448,0.0001329112,0.0007257949,0.0008951874,0.000417539,0.0012700447,0.00067588536,0.0006073887],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.00051758165,0.0008088433,0.68168795,0.00029123772,0.000066831155,0.003413123,0.110010095,0.0006194801,0.032343272,0.0020085336,0.0008818234,0.1673512],"study_design_scores_gemma":[0.00001627472,0.00038762056,0.9420906,0.0001234328,0.00004163857,0.0018760316,0.03679355,0.00091119466,0.009472174,0.0010683293,0.007185713,0.000033534292],"about_ca_topic_score_codex":0.010081822,"about_ca_topic_score_gemma":0.012164702,"teacher_disagreement_score":0.010081822,"about_ca_system_score_codex":0.0013198698,"about_ca_system_score_gemma":0.0010612605,"threshold_uncertainty_score":0.020046294},"labels":[],"label_agreement":null},{"id":"W612974530","doi":"10.1075/scl.67","title":"Grammatical Change in English World-Wide","year":2015,"lang":"en","type":"book","venue":"Studies in corpus linguistics","topic":"Linguistic Variation and Morphology","field":"Social Sciences","cited_by":104,"is_retracted":false,"has_abstract":true,"route_ca_aff":false,"route_ca_fund":false,"route_ca_venue":false,"route_about_ca":true,"ca_institutions":"","funders":"","keywords":"Computer science; Linguistics; Natural language processing; Philosophy","score_opus":0.16084686923901154,"score_gpt":0.4076607891824785,"score_spread":0.24681391994346696,"validation_status":"score_only:v0-immature-baseline","prediction":{"id":"W612974530","genre_codex":"other","genre_gemma":"other","domain_codex":null,"domain_gemma":null,"model_version":"metacan-v3-hybrid-931329e0061c","genre_candidate":"other","genre_consensus":"other","domain_candidate":null,"domain_consensus":null,"prediction_status":"machine_predicted_unvalidated","genre_scores_codex":[0.17986037,0.02401812,0.003232284,0.0049914024,0.0007679398,0.000031300046,0.00033799632,0.00013266545,0.78662807],"genre_scores_gemma":[0.90930223,0.008525837,0.0015322159,0.0004922982,0.00027157672,0.000027012285,0.0003813207,0.00009388968,0.079373576],"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"not_applicable","domain_scores_codex":[0.9995796,0.00013168376,0.000020756741,0.00007481943,0.00014687443,0.000046203048],"domain_scores_gemma":[0.99971586,0.00013502077,0.000024921703,0.00004064139,0.00007001553,0.00001355246],"candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.00040810695,0.00014716055,0.00016313599,0.0011582738,0.0012835036,0.0027178416,0.00026865953,0.00032112535,0.002142559],"category_scores_gemma":[0.00095023325,0.00009209843,0.00011337743,0.0016719598,0.002301346,0.002185231,0.0009896146,0.0006508004,0.00024958208],"study_design_candidate":"not_applicable","study_design_consensus":null,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_system_candidate":false,"about_ca_system_consensus":false,"study_design_scores_codex":[0.000018548717,0.000019700261,0.0033253897,0.0001566097,0.00001249692,0.00042525565,0.08041542,0.00035750776,0.0011902836,0.73844826,0.025124405,0.15050618],"study_design_scores_gemma":[0.000004138699,0.000027576214,0.014360142,0.00017038158,0.000010466671,0.00087762665,0.017628558,0.00044544903,0.00051066215,0.06954728,0.8964021,0.000015624531],"about_ca_topic_score_codex":0.0033482763,"about_ca_topic_score_gemma":0.005129461,"teacher_disagreement_score":0.0033482763,"about_ca_system_score_codex":0.001591213,"about_ca_system_score_gemma":0.00038258248,"threshold_uncertainty_score":0.011545181},"labels":[],"label_agreement":null}]}