{"id":"W6910373184","doi":"10.48448/k4b7-c679","title":"Metric-guided Distillation: Distilling Knowledge from the Metric to Ranker and Retriever for Generative Commonsense Reasoning","year":2022,"lang":"en","type":"other","venue":"Underline Science Inc.","topic":"","field":"","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"","keywords":"Metric (unit); Generalization; Relevance (law); Sentence; Benchmark (surveying); Labrador Retriever; Commonsense reasoning","routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true,"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":[0.001587324,0.001405279,0.0008371223,0.001156768,0.0005413993,0.001235462,0.002331714,0.001306342,0.006218981],"category_scores_gemma":[0.005799732,0.0004266973,0.001056583,0.0006929388,0.001213439,0.003165401,0.002565959,0.001757497,0.002062901],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009726074,"about_ca_system_score_gemma":0.001341271,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.003020536,"about_ca_topic_score_gemma":0.006098617,"domain_scores_codex":[0.9988903,0.0003898844,0.00006579432,0.0003218153,0.0002512352,0.00008083274],"domain_scores_gemma":[0.9985947,0.0006797757,0.0001043914,0.0004003783,0.000136866,0.00008378582],"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":[0.0003837931,0.0002760128,0.001025559,0.0005165869,0.0001035588,0.0005067496,0.0004989812,0.2006854,0.02113442,0.04941395,0.01604339,0.7094116],"study_design_scores_gemma":[0.00004920047,0.000100632,0.0002180021,0.0000323457,0.00003699544,0.0001800532,0.00004702844,0.9321539,0.01288863,0.04773366,0.006521467,0.00003806467],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.02099402,0.0006542729,0.9630898,0.0004299112,0.0000842856,0.0001912253,0.0005696579,0.009617221,0.004369567],"genre_scores_gemma":[0.5280848,0.0003785935,0.4600058,0.0005501528,0.0001053169,0.0003092413,0.002451769,0.001274158,0.006840187],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.006218981,"threshold_uncertainty_score":0.02080452,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05362116947968987,"score_gpt":0.3383028259425722,"score_spread":0.2846816564628823,"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."}}