{"id":"W4413025253","doi":"10.1007/978-3-032-00891-6_28","title":"Enhancing Ultra-Low-Bit Quantization of Large Language Models Through Saliency-Aware Partial Retraining","year":2025,"lang":"en","type":"book-chapter","venue":"Lecture notes in computer science","topic":"Topic Modeling","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":false,"ca_institutions":"University of Toronto; University of New Brunswick","funders":"","keywords":"Computer science; Retraining; Quantization (signal processing); Bit (key); Artificial intelligence; Speech recognition; Computer vision; Computer network","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":[0.001108944,0.0004783464,0.0006525912,0.0005499262,0.0002357001,0.0002463029,0.002711643,0.0003857893,0.00001847423],"category_scores_gemma":[0.0001540917,0.0004739368,0.0001579442,0.000782743,0.0002601304,0.001075524,0.0008368469,0.0007130174,0.000005198524],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001938565,"about_ca_system_score_gemma":0.0007667744,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00005924764,"about_ca_topic_score_gemma":0.0001289096,"domain_scores_codex":[0.9956928,0.00005745185,0.0008755517,0.001553747,0.001080815,0.0007396495],"domain_scores_gemma":[0.9972681,0.000410109,0.0004353539,0.001510366,0.0002839047,0.0000921478],"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":[0.00001234761,0.00008473236,0.00004640432,0.0005107126,0.00003591517,0.0001076226,0.03072613,0.3665752,0.005416777,0.3483303,0.00002258686,0.2481313],"study_design_scores_gemma":[0.0002301183,0.00005464746,0.000003324012,0.001249552,0.000009229439,0.00001076293,0.000003622906,0.9131574,0.02478334,0.06002501,0.00004316352,0.0004298083],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0003575225,0.0004412952,0.9926651,0.0002624222,0.001561803,0.0003411975,0.00001357522,0.0001884112,0.004168657],"genre_scores_gemma":[0.7064903,0.00003382412,0.2923144,0.0006922267,0.000260955,0.000005709554,0.000009920916,0.00002287821,0.0001698497],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.7061328,"threshold_uncertainty_score":0.9997712,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02045388212035081,"score_gpt":0.2747283330984558,"score_spread":0.254274450978105,"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."}}