{"id":"W6966676709","doi":"10.48448/b7gs-gp03","title":"Sorted LLaMA: Unlocking the Potential of Intermediate Layers of Large Language Models for Dynamic Inference","year":2024,"lang":"en","type":"other","venue":"Underline Science Inc.","topic":"","field":"","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"York University","funders":"","keywords":"Inference; Language model; Modularity (biology); Transformer; Generative grammar; Sorting; Dynamic random-access memory; Natural language","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":[0.001676151,0.0004195731,0.0005971131,0.001302177,0.000101759,0.00009250941,0.001853221,0.0002327374,0.0002173316],"category_scores_gemma":[0.0003215208,0.0002985289,0.0002047625,0.001518973,0.001953794,0.0002098616,0.0006271626,0.0004093884,0.0001727573],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001757518,"about_ca_system_score_gemma":0.0008463764,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0003917484,"about_ca_topic_score_gemma":0.00195839,"domain_scores_codex":[0.9966375,0.00006751234,0.0006583607,0.0007883082,0.001103225,0.0007450426],"domain_scores_gemma":[0.9977339,0.000127966,0.0008032105,0.00092821,0.0003011736,0.0001055551],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"bench_or_experimental","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0005514054,0.002160433,0.0002486341,0.008151092,0.002380314,0.0001663768,0.02070633,0.0564719,0.6121421,0.146124,0.1185733,0.03232404],"study_design_scores_gemma":[0.0008072105,0.000143149,0.00003878277,0.001112447,0.0002806491,0.000006605528,0.002338302,0.9849757,0.001204875,0.006901443,0.001711613,0.0004792533],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.07554091,0.01297657,0.6247451,0.0006837651,0.0106063,0.01139866,0.03059203,0.003068856,0.2303878],"genre_scores_gemma":[0.9674112,0.00004330177,0.0019238,0.00004246027,0.0001266566,0.00003536214,0.0001300613,0.0005411614,0.02974596],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9285038,"threshold_uncertainty_score":0.9999467,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01711460245891395,"score_gpt":0.330973591178419,"score_spread":0.313858988719505,"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."}}