{"id":"W7011290330","doi":"","title":"Low-Bit Power-of-Two Quantization for Large Language Models","year":2025,"lang":"en","type":"dissertation","venue":"eScholarship@McGill (McGill)","topic":"History of Computing Technologies","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"","keywords":"Quantization (signal processing); Language model; Natural language; Language identification","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":[0.002044337,0.0007956056,0.001213104,0.0008858927,0.0009890764,0.001540393,0.001692075,0.001299241,0.01181421],"category_scores_gemma":[0.01370731,0.0005203751,0.000593485,0.001156782,0.001156471,0.004281147,0.002490934,0.002374198,0.00282861],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001271378,"about_ca_system_score_gemma":0.001832944,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004216325,"about_ca_topic_score_gemma":0.01069442,"domain_scores_codex":[0.9981232,0.0007080706,0.0001431109,0.0002497161,0.0005774013,0.0001984407],"domain_scores_gemma":[0.9947202,0.003142257,0.0001244658,0.001273029,0.0006188494,0.0001211132],"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":[0.001120278,0.0002327295,0.001252141,0.0003721615,0.00008456352,0.0002509813,0.0003672701,0.08828149,0.01829581,0.1760178,0.04152341,0.6722014],"study_design_scores_gemma":[0.00007997995,0.00006053311,0.0001635163,0.00003969843,0.00001618209,0.0000665213,0.00005879774,0.8538509,0.006843837,0.1348572,0.003923287,0.0000395273],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.02571264,0.001316939,0.959174,0.001895564,0.0004732827,0.0001469169,0.0007049115,0.005435163,0.00514057],"genre_scores_gemma":[0.4962539,0.0006569219,0.4877444,0.001364525,0.0003370142,0.0003934377,0.001219687,0.0008553916,0.01117472],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.01181421,"threshold_uncertainty_score":0.03952247,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01764006618815532,"score_gpt":0.2677293591335536,"score_spread":0.2500892929453983,"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."}}