{"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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.006991529,0.001360195,0.001132606,0.001905517,0.0008165365,0.00339799,0.00191382,0.002267495,0.001202372],"category_scores_gemma":[0.03373142,0.001275801,0.0008792107,0.001136746,0.001184701,0.006936096,0.002041369,0.005855338,0.0007326081],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002517878,"about_ca_system_score_gemma":0.00181845,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.009543543,"about_ca_topic_score_gemma":0.01875103,"domain_scores_codex":[0.9970558,0.00173024,0.000109072,0.0007220481,0.0002228727,0.0001599651],"domain_scores_gemma":[0.9694751,0.02560361,0.0009741306,0.002462639,0.001105324,0.0003791635],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0006259146,0.0003573614,0.008395908,0.0002932838,0.0003951127,0.0001791079,0.002407898,0.7946162,0.01514704,0.01419311,0.0028092,0.1605798],"study_design_scores_gemma":[0.00001739682,0.00003991053,0.0006486861,0.00001832261,0.00001637153,0.00001258476,0.0001271437,0.9854036,0.001474125,0.01159953,0.0006265795,0.00001569377],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2608354,0.001811068,0.7304624,0.001325357,0.00009816852,0.0001323251,0.0007424067,0.002247247,0.00234556],"genre_scores_gemma":[0.8014104,0.0004378637,0.1925046,0.0003178235,0.0001459856,0.0002217414,0.002282134,0.0004284712,0.002250945],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.009543543,"threshold_uncertainty_score":0.0369752,"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."}}