{"id":"W6891607569","doi":"10.48448/fvef-7072","title":"Distilling Linguistic Context for Language Model Compression","year":2021,"lang":"en","type":"other","venue":"Underline Science Inc.","topic":"","field":"","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Kootenay Association for Science & Technology","funders":"","keywords":"Language model; Representation (politics); Context (archaeology); Knowledge representation and reasoning; Relation (database); Pruning; Universal Networking Language; Word (group theory); Artificial neural 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":[0.000741231,0.0004639497,0.000578196,0.0006715319,0.000348559,0.0002362859,0.001116072,0.0002425784,0.0007054684],"category_scores_gemma":[0.001576137,0.0004175464,0.0001264172,0.0007466311,0.001065861,0.0000673529,0.0003382278,0.0003366232,0.0002950968],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002757924,"about_ca_system_score_gemma":0.001036082,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0003621505,"about_ca_topic_score_gemma":0.001060703,"domain_scores_codex":[0.9966527,0.00003954156,0.000405041,0.001164293,0.0009447771,0.0007936665],"domain_scores_gemma":[0.997802,0.0001798075,0.0004257594,0.0009239949,0.0003854159,0.0002830624],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.0001509623,0.001187242,0.000117031,0.001861668,0.0002249346,0.0002188252,0.006244285,0.02538835,0.1631929,0.06289633,0.6902797,0.0482378],"study_design_scores_gemma":[0.0006510912,0.00003546097,0.000002202819,0.001184478,0.00006862223,0.00000707256,0.0004767317,0.8827105,0.001081102,0.0007557665,0.1124087,0.0006183262],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"other","genre_gemma":"other","genre_scores_codex":[0.0004030329,0.009543029,0.4116036,0.0001151862,0.002874255,0.002283344,0.003018771,0.001536379,0.5686225],"genre_scores_gemma":[0.3050635,0.00004487328,0.1264804,0.0004555337,0.00190708,0.0001193833,0.00159216,0.001916555,0.5624205],"genre_candidate":"other","genre_consensus":"other","teacher_disagreement_score":0.8573222,"threshold_uncertainty_score":0.9998276,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04778031652179233,"score_gpt":0.3565567001907823,"score_spread":0.30877638366899,"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."}}