{"id":"W4412888761","doi":"10.18653/v1/2025.findings-acl.111","title":"Drop Dropout on Single Epoch Language Model Pretraining","year":2025,"lang":"en","type":"article","venue":"","topic":"Speech and dialogue systems","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Canadian Institute for Advanced Research","keywords":"Dropout (neural networks); Drop out; Computer science; Drop (telecommunication); Language model; Artificial intelligence; Machine learning; Economics; Telecommunications","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":[0.002135451,0.001726506,0.0009782116,0.0004788547,0.0005731119,0.001030473,0.002274921,0.001853604,0.006977787],"category_scores_gemma":[0.01045398,0.0006883302,0.0008948415,0.0004135758,0.0006452515,0.002100804,0.00184437,0.003953754,0.003556486],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009232739,"about_ca_system_score_gemma":0.001520035,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00629871,"about_ca_topic_score_gemma":0.0130655,"domain_scores_codex":[0.9990655,0.0003158102,0.00005170266,0.0002279993,0.0002001881,0.000138856],"domain_scores_gemma":[0.9973888,0.001555198,0.00008991676,0.0005388927,0.0003221572,0.0001050829],"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":[0.000963405,0.000472247,0.002827891,0.0004998847,0.0003288284,0.0006504718,0.0004133985,0.3881052,0.04585959,0.008239701,0.03878067,0.5128587],"study_design_scores_gemma":[0.00005043357,0.0001587699,0.0006113924,0.00003492915,0.00003603039,0.00008645459,0.0000471729,0.9661058,0.02375058,0.004021977,0.005070997,0.00002546679],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.08625966,0.001682571,0.8624116,0.001470219,0.0007062432,0.0002428831,0.001006699,0.04071064,0.00550957],"genre_scores_gemma":[0.6892163,0.0006182582,0.2896969,0.001639497,0.0001556436,0.0004597323,0.003607443,0.002533527,0.01207279],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.006977787,"threshold_uncertainty_score":0.02334303,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01653676269129877,"score_gpt":0.2648106750246286,"score_spread":0.2482739123333298,"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."}}