{"id":"W4403569925","doi":"10.48550/arxiv.2410.09615","title":"SLiM: One-shot Quantization and Sparsity with Low-rank Approximation for LLM Weight Compression","year":2024,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Sparse and Compressive Sensing Techniques","field":"Engineering","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Alliance de recherche numérique du Canada; Natural Sciences and Engineering Research Council of Canada; Canada Research Chairs; DeepMind","keywords":"Shot (pellet); Rank (graph theory); One shot; Mathematics; Physics; Econometrics; Combinatorics; Chemistry; Engineering","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.0006745494,0.001136672,0.000890721,0.0009013785,0.0003938297,0.001011355,0.001561911,0.001040233,0.003632646],"category_scores_gemma":[0.004953635,0.0003941681,0.0006553431,0.0008860434,0.0006070741,0.00176747,0.001642591,0.001861069,0.001221954],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006892785,"about_ca_system_score_gemma":0.001068661,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005885693,"about_ca_topic_score_gemma":0.01048924,"domain_scores_codex":[0.9994617,0.00008360902,0.00003246999,0.00008528693,0.0002885335,0.00004831794],"domain_scores_gemma":[0.9991832,0.00028074,0.00007998505,0.000202109,0.0002021497,0.00005176935],"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":[0.0003314709,0.0001861642,0.001169541,0.0002877168,0.00009897163,0.0002334891,0.0002116211,0.3232096,0.0283646,0.02561712,0.01447642,0.6058133],"study_design_scores_gemma":[0.000008246718,0.00003184441,0.0001213814,0.00001130622,0.000005253127,0.0000380136,0.00001585592,0.9882327,0.004249064,0.005978947,0.001297846,0.00000972192],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.008699992,0.0002665159,0.9876357,0.00017088,0.00006451795,0.00005414041,0.0002094329,0.001906207,0.0009925673],"genre_scores_gemma":[0.2881219,0.0004659357,0.7034143,0.0004606693,0.0001192118,0.0002748006,0.001983822,0.0007249484,0.004434563],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.005885693,"threshold_uncertainty_score":0.01215237,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.08459366769301566,"score_gpt":0.1957024386642909,"score_spread":0.1111087709712753,"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."}}