{"id":"W4386566845","doi":"10.18653/v1/2023.eacl-main.13","title":"Do we need Label Regularization to Fine-tune Pre-trained Language Models?","year":2023,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"Université de Montréal; University of Waterloo","funders":"","keywords":"Pascal (unit); Regularization (linguistics); Computer science; Natural language processing; Artificial intelligence; Theoretical computer science; Programming language","routes":{"ca_aff":true,"ca_fund":false,"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":[0.00730774,0.002170084,0.002242775,0.000820792,0.001196879,0.002884591,0.003559984,0.0042845,0.00510933],"category_scores_gemma":[0.03016037,0.001230701,0.001172033,0.001037557,0.001329989,0.007879703,0.001816353,0.007216375,0.00756046],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001039015,"about_ca_system_score_gemma":0.002293461,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01086218,"about_ca_topic_score_gemma":0.02955444,"domain_scores_codex":[0.9972171,0.001384415,0.0000967851,0.0008039669,0.0002468321,0.0002509929],"domain_scores_gemma":[0.9891734,0.006585806,0.0003656127,0.002085294,0.001344025,0.0004457722],"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":[0.0008716902,0.0008841802,0.008285618,0.0008449838,0.0008110793,0.0002080258,0.000512924,0.0939047,0.02617357,0.02134638,0.1270219,0.7191349],"study_design_scores_gemma":[0.0002286053,0.0001232856,0.001998886,0.0002325513,0.0002093845,0.0001902973,0.0003163149,0.8681186,0.009537104,0.09475431,0.02419326,0.00009725451],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02674464,0.004847805,0.9378315,0.01123768,0.001602955,0.0001179683,0.0009005013,0.01047112,0.006245766],"genre_scores_gemma":[0.3593691,0.002560941,0.6033698,0.009207517,0.001956205,0.0003972179,0.004645866,0.006197942,0.01229544],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01086218,"threshold_uncertainty_score":0.03864747,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03158016076292302,"score_gpt":0.2743124755449816,"score_spread":0.2427323147820586,"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."}}