{"id":"W4392487848","doi":"10.48550/arxiv.2403.01427","title":"Logit Standardization in Knowledge Distillation","year":2024,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Semantic Web and Ontologies","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"National Natural Science Foundation of China; Canadian Institute for Advanced Research","keywords":"Standardization; Logit; Logistic regression; Distillation; Econometrics; Computer science; Economics; Mathematics; Statistics; Chemistry; Chromatography","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":[0.003007114,0.001397797,0.00138618,0.001309638,0.0008855994,0.001959753,0.002202517,0.001366599,0.006637546],"category_scores_gemma":[0.01094699,0.0005613945,0.001052007,0.001610116,0.001880723,0.004779918,0.004098907,0.003491767,0.002162158],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001535013,"about_ca_system_score_gemma":0.002322514,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005183493,"about_ca_topic_score_gemma":0.006176208,"domain_scores_codex":[0.9978172,0.0007473985,0.0001253089,0.0005585632,0.0005063584,0.0002452985],"domain_scores_gemma":[0.9978147,0.001092616,0.0001504061,0.0004468475,0.0004011196,0.00009436259],"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":[0.0007105359,0.0003275705,0.002718618,0.0003363617,0.0001662297,0.0001479442,0.0003613732,0.2837158,0.01202106,0.06833479,0.005712993,0.6254468],"study_design_scores_gemma":[0.00006541261,0.0001298076,0.0008746586,0.00004012811,0.000035647,0.00007580053,0.00009387543,0.8911136,0.01858167,0.08391902,0.005003329,0.00006691773],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.02329776,0.000532863,0.9685361,0.0004389038,0.00008470615,0.00008570611,0.000220171,0.00400786,0.002796051],"genre_scores_gemma":[0.6797054,0.000500363,0.3061636,0.0005283607,0.0001277983,0.0003133884,0.001162172,0.0007795086,0.01071931],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.006637546,"threshold_uncertainty_score":0.02220482,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.08365277220118163,"score_gpt":0.2102097908837186,"score_spread":0.126557018682537,"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."}}