{"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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0002132514,0.0001848769,0.0002212508,0.0003353738,0.0000466915,0.0001444409,0.0007897069,0.0002199671,0.000009882597],"category_scores_gemma":[0.00005050887,0.0001977411,0.00009625965,0.0006673859,0.00005554247,0.0001710956,0.001845746,0.0003596202,0.0001322153],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002811315,"about_ca_system_score_gemma":0.000210403,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00007878742,"about_ca_topic_score_gemma":0.0002983227,"domain_scores_codex":[0.9987254,0.00008701941,0.0001587138,0.0007665352,0.00005747485,0.0002048995],"domain_scores_gemma":[0.999088,0.00008208551,0.00008415643,0.0006104515,0.00008693417,0.00004835622],"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":[0.00001399625,0.00005667118,0.007714886,0.000209303,0.00003771572,0.0003589171,0.0007836263,0.1799272,0.00001067445,0.8056587,0.0005931294,0.004635216],"study_design_scores_gemma":[0.000188433,0.00002006799,0.007768616,0.0001658147,0.00002498689,0.000001947075,0.00006453784,0.6887596,0.00005886763,0.3017592,0.0009158779,0.0002720674],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1684844,0.0004273448,0.8153376,0.0001593982,0.001195322,0.0001951964,0.000009228369,0.0003716315,0.01381991],"genre_scores_gemma":[0.9980959,0.0001331662,0.0007293092,0.00001224932,0.00005162445,5.551644e-7,0.00001072409,0.000008590939,0.000957831],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8296116,"threshold_uncertainty_score":0.8063648,"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."}}