{"id":"W4287854991","doi":"10.18653/v1/2022.naacl-main.170","title":"Towards Understanding Large-Scale Discourse Structures in Pre-Trained and Fine-Tuned Language Models","year":2022,"lang":"en","type":"article","venue":"Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies","topic":"Topic Modeling","field":"Computer Science","cited_by":7,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Computer science; Redundancy (engineering); Natural language processing; Artificial intelligence; Language model; Scale (ratio); Simple (philosophy)","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0004205974,0.0001421762,0.0002911803,0.0001415296,0.000298167,0.0000398934,0.001732633,0.00003490539,0.000002451994],"category_scores_gemma":[0.001045869,0.00009741021,0.0001274993,0.0004144132,0.0002662362,0.00007132569,0.001502837,0.0002640145,2.027179e-8],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001784524,"about_ca_system_score_gemma":0.00006828069,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001561879,"about_ca_topic_score_gemma":0.000197901,"domain_scores_codex":[0.9985359,0.0000249574,0.0004015196,0.0002568528,0.0005796157,0.0002011998],"domain_scores_gemma":[0.9980678,0.0001361868,0.001212055,0.0002542737,0.0003164239,0.00001327525],"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":[0.00002898876,0.00007106656,0.04708419,0.0001118822,0.00009290618,2.255349e-7,0.02081461,0.02782863,0.001247715,0.9007897,0.00004230496,0.001887781],"study_design_scores_gemma":[0.000940208,0.0001772426,0.03959806,0.0001056374,0.00008475813,0.000002760094,0.03070738,0.3407156,0.004068311,0.5832394,0.00002279833,0.0003379533],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":"theoretical_or_conceptual","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.984985,0.00008340807,0.01202839,0.001542769,0.0001962245,0.0005201824,0.0002093337,0.00008561625,0.0003490685],"genre_scores_gemma":[0.9924249,0.000002371501,0.00739618,0.00003604464,0.00002096488,0.00002825387,0.000004036906,0.00001000742,0.00007722522],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.3175504,"threshold_uncertainty_score":0.3972273,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02328836485699962,"score_gpt":0.2689538667565817,"score_spread":0.2456655018995821,"validation_status":"score_only:v0-immature-baseline","note":"Baseline scores from an immature model (maturity gate not passed). Scores rank; they never assert a category."}}