{"id":"W4320853823","doi":"10.48550/arxiv.2302.05895","title":"Discourse Structure Extraction from Pre-Trained and Fine-Tuned Language Models in Dialogues","year":2023,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Topic Modeling","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada; Agence Nationale de la Recherche","keywords":"Computer science; Task (project management); Exploit; Natural language processing; Artificial intelligence; Sentence; Language model; Machine learning","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.002681622,0.001817847,0.001071153,0.002631361,0.0008976506,0.001976791,0.001304294,0.00143001,0.002178216],"category_scores_gemma":[0.0116428,0.0009123517,0.00132897,0.00136393,0.0006311833,0.002873794,0.001808717,0.002603529,0.002699345],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001147969,"about_ca_system_score_gemma":0.001522617,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004686167,"about_ca_topic_score_gemma":0.009444249,"domain_scores_codex":[0.9972407,0.00143095,0.0001297533,0.0008320909,0.0002087502,0.0001578798],"domain_scores_gemma":[0.9944527,0.003842859,0.000245305,0.0005570119,0.0007204923,0.0001816551],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001045737,0.0004781267,0.007129886,0.001151013,0.0003797693,0.0004674388,0.003309898,0.1218221,0.08897983,0.006377671,0.01690555,0.7519531],"study_design_scores_gemma":[0.00008199045,0.0001708777,0.003021924,0.00007784577,0.0001139906,0.0001015445,0.0006856611,0.9487398,0.0281502,0.009907348,0.008894955,0.00005388199],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1343239,0.001869767,0.8426067,0.0007637042,0.0002143448,0.0002674046,0.002198236,0.01472212,0.003033849],"genre_scores_gemma":[0.6327156,0.0005958906,0.3529111,0.0001718487,0.0002103666,0.0004122221,0.007987593,0.001041236,0.003954144],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004686167,"threshold_uncertainty_score":0.01418197,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.08395320489432903,"score_gpt":0.2229728679408234,"score_spread":0.1390196630464944,"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."}}