{"id":"W6910296007","doi":"10.48448/yn34-7z07","title":"Efficiently Enhancing Zero-Shot Performance of Instruction Following Model via Retrieval of Soft Prompt","year":2023,"lang":"en","type":"other","venue":"Open MIND","topic":"","field":"","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Kootenay Association for Science & Technology","funders":"","keywords":"Task (project management); Benchmark (surveying); Embedding; Training set; Scaling; Training (meteorology); Task analysis","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":[0.001801421,0.002414983,0.001332068,0.0005345522,0.0005147557,0.001757102,0.002761867,0.002138701,0.007180705],"category_scores_gemma":[0.009230692,0.0005162147,0.00113675,0.0004885005,0.0008098797,0.004784613,0.002499925,0.003573734,0.004919081],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0009612807,"about_ca_system_score_gemma":0.001593373,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006184699,"about_ca_topic_score_gemma":0.008888608,"domain_scores_codex":[0.9990196,0.0002269989,0.00004900901,0.0004512724,0.0001111515,0.0001419276],"domain_scores_gemma":[0.9978158,0.001017708,0.00008391735,0.00065991,0.0002494864,0.0001732318],"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":[0.001776627,0.001056063,0.007704762,0.0007770354,0.0002292613,0.0003691675,0.000471966,0.1581012,0.02357638,0.005246888,0.04144584,0.7592448],"study_design_scores_gemma":[0.000085039,0.0004508048,0.001100501,0.00005402156,0.00005433814,0.0001345985,0.0002151605,0.9701837,0.01310604,0.01098309,0.003585741,0.00004694965],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3759613,0.005369359,0.5117738,0.001995943,0.001478415,0.0004271332,0.004157927,0.083811,0.015025],"genre_scores_gemma":[0.8750803,0.0004429515,0.1023695,0.0009735231,0.0001462729,0.0002273572,0.00857436,0.001213851,0.01097185],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.007180705,"threshold_uncertainty_score":0.0240218,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03759482533536243,"score_gpt":0.303201677765894,"score_spread":0.2656068524305315,"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."}}