{"id":"W7082283166","doi":"10.48448/8ygs-qg81","title":"Self-Training Elicits Concise Reasoning in Large Language Models","year":2025,"lang":"en","type":"other","venue":"Underline Science Inc.","topic":"Geochemistry and Geologic Mapping","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Kootenay Association for Science & Technology","funders":"","keywords":"Inference; Language model; Sampling (signal processing); Non-monotonic logic; Data modeling; Latent variable","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":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.001313186,0.0003060023,0.0003744097,0.0006294035,0.0001772956,0.0002064433,0.002373054,0.0002500649,0.0001253428],"category_scores_gemma":[0.000353457,0.0002961469,0.0000526831,0.001639822,0.0001760973,0.0003496026,0.0009619694,0.0004473125,0.00003836063],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001163603,"about_ca_system_score_gemma":0.001131172,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002071076,"about_ca_topic_score_gemma":0.0002695098,"domain_scores_codex":[0.9972399,0.00004975422,0.0003148882,0.001010822,0.0005350582,0.0008496068],"domain_scores_gemma":[0.9985601,0.00008366532,0.0002062391,0.0008901039,0.0001180436,0.0001418362],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000005320108,0.0003651774,0.0003809061,0.0005293718,0.00006359414,0.0005542517,0.01568243,0.00383948,0.00093929,0.893942,0.03763486,0.04606334],"study_design_scores_gemma":[0.0005522225,0.00002438805,0.00002448537,0.0007189002,0.000009447371,0.00002099857,0.0009052792,0.8839723,0.0003195014,0.008185324,0.1046655,0.0006016315],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"other","genre_gemma":"other","genre_scores_codex":[0.00005699581,0.0005728777,0.159231,0.0005726432,0.0003512346,0.0002195868,0.00001199781,0.0005972492,0.8383864],"genre_scores_gemma":[0.2219496,0.0001309697,0.2542763,0.001505324,0.0003489084,0.00004708717,0.00003619879,0.00006127459,0.5216443],"genre_candidate":"other","genre_consensus":"other","teacher_disagreement_score":0.8857567,"threshold_uncertainty_score":0.999949,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01684458539469767,"score_gpt":0.2691218664544908,"score_spread":0.2522772810597931,"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."}}