{"id":"W7130308397","doi":"10.48448/k9e2-8102","title":"Revisiting Intermediate-Layer Matching in Knowledge Distillation: Layer-Selection Strategy Doesn’t Matter (Much)","year":2025,"lang":"","type":"other","venue":"Underline Science Inc.","topic":"","field":"","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Alberta","funders":"","keywords":"Matching (statistics); Phenomenon; Focus (optics); Interpretation (philosophy); Work (physics); Layer (electronics)","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":[0.004767042,0.0006146895,0.0006977414,0.000632362,0.001166657,0.003318975,0.002039959,0.001341529,0.005179915],"category_scores_gemma":[0.01818863,0.0004436862,0.0005876925,0.0009211786,0.002671669,0.009417721,0.004593295,0.002318544,0.001625154],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001128129,"about_ca_system_score_gemma":0.002163303,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003444418,"about_ca_topic_score_gemma":0.003968578,"domain_scores_codex":[0.9966838,0.001398406,0.0002475814,0.0008557005,0.0005426836,0.0002717374],"domain_scores_gemma":[0.9927084,0.003119678,0.0005037207,0.002483986,0.0008068287,0.0003773691],"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":[0.0006572089,0.0004400003,0.007523032,0.0005496423,0.0001582588,0.0002220766,0.002138393,0.06205264,0.02001823,0.3407368,0.006710195,0.5587936],"study_design_scores_gemma":[0.0001074328,0.0002605656,0.001690409,0.000135184,0.00009301422,0.000212401,0.0006151867,0.5357494,0.03425305,0.4075231,0.0192747,0.00008540553],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.08772399,0.0005973594,0.8924975,0.00229663,0.00012149,0.0001597489,0.00019912,0.001896506,0.01450768],"genre_scores_gemma":[0.7552651,0.0002524989,0.2386944,0.0004994022,0.00003869099,0.00011203,0.0002527628,0.0002845695,0.004600611],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.005179915,"threshold_uncertainty_score":0.02521086,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0281727157441371,"score_gpt":0.3306146019196956,"score_spread":0.3024418861755585,"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."}}