{"id":"W6948203318","doi":"10.48448/pnm1-gx20","title":"Self-Consistency of Large Language Models under Ambiguity","year":2023,"lang":"en","type":"other","venue":"Open MIND","topic":"Medicinal Plant Pharmacodynamics Research","field":"Pharmacology, Toxicology and Pharmaceutics","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Dalhousie University","funders":"","keywords":"Ambiguity; Language model; Nonparametric statistics; Robustness (evolution); Consistency (knowledge bases); Suite; Probability distribution; Series (stratigraphy)","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","insufficient_payload"],"consensus_categories":["insufficient_payload"],"category_scores_codex":[0.001867387,0.000364615,0.0007105274,0.0003886962,0.00009701138,0.0000323223,0.001353438,0.0008295843,0.06953939],"category_scores_gemma":[0.00006950125,0.0003457283,0.0001291021,0.0003601628,0.0001779576,0.00009911961,0.000753445,0.001390129,0.003447325],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001032734,"about_ca_system_score_gemma":0.0004283591,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0004940746,"about_ca_topic_score_gemma":0.002480268,"domain_scores_codex":[0.9972669,0.0005426866,0.0004791648,0.0005863049,0.0004408699,0.0006840916],"domain_scores_gemma":[0.9983205,0.0004296154,0.0003423825,0.0004989535,0.00007551209,0.000333096],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0005466901,0.001806691,0.0003388926,0.0009424665,0.003249147,0.00135476,0.002911818,0.0002871235,0.005186013,0.002309098,0.9349186,0.0461487],"study_design_scores_gemma":[0.002683984,0.00008756403,0.00002088859,0.0002085464,0.0004718234,0.00002694011,0.000587506,0.01549007,0.002137575,0.0002070667,0.9775885,0.0004895592],"study_design_candidate":"not_applicable","study_design_consensus":"not_applicable","genre_codex":"other","genre_gemma":"other","genre_scores_codex":[0.004451599,0.001342895,0.0000792707,0.0002322668,0.001131652,0.001518834,0.003266023,0.00006064075,0.9879168],"genre_scores_gemma":[0.02697835,0.001893681,0.001285333,0.0003787339,0.0003802531,0.00007377166,0.0003671474,0.0007554809,0.9678872],"genre_candidate":"other","genre_consensus":"other","teacher_disagreement_score":0.06609207,"threshold_uncertainty_score":0.9998994,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.2207518045997466,"score_gpt":0.5200152132505748,"score_spread":0.2992634086508282,"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."}}