{"id":"W4412955598","doi":"10.1088/1742-5468/adefb1","title":"How feature learning can improve neural scaling laws<sup>*</sup>","year":2025,"lang":"en","type":"article","venue":"Journal of Statistical Mechanics Theory and Experiment","topic":"Neural Networks and Applications","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"","keywords":"Feature (linguistics); Scaling law; Scaling; Artificial neural network; Statistical physics; Artificial intelligence; Computer science; Law; Mathematics; Physics; Political science; Philosophy; Geometry","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":[0.001169698,0.0005479878,0.0004900984,0.0002752227,0.0002768328,0.0008755461,0.001013469,0.0008014155,0.002698862],"category_scores_gemma":[0.01078525,0.0002594797,0.000424616,0.0002673862,0.0008641474,0.002674018,0.00083868,0.0009837975,0.0006392599],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006778509,"about_ca_system_score_gemma":0.0004417234,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001186094,"about_ca_topic_score_gemma":0.0007818647,"domain_scores_codex":[0.9996957,0.00008482897,0.00001946203,0.00007161883,0.0000775472,0.00005092732],"domain_scores_gemma":[0.9968815,0.001691236,0.0003091393,0.0006800893,0.0003221509,0.0001159777],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0003054084,0.0002377375,0.003479513,0.0001995145,0.00005289068,0.0002311135,0.0002594218,0.7274184,0.03289728,0.1028117,0.005335886,0.1267713],"study_design_scores_gemma":[0.00000497224,0.00003280327,0.0002523659,0.000004398889,0.000003443651,0.00002416507,0.000008274524,0.9752075,0.00232387,0.02179771,0.0003353625,0.000005133708],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.2775436,0.0003117985,0.7115891,0.001341464,0.0001135014,0.00006708344,0.0001499276,0.001649391,0.007234186],"genre_scores_gemma":[0.9452918,0.00009798852,0.05296982,0.0001318126,0.00003876821,0.00006246244,0.00008944584,0.0001524478,0.001165551],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.002698862,"threshold_uncertainty_score":0.009028614,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.007363415210228173,"score_gpt":0.259062483169425,"score_spread":0.2516990679591969,"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."}}