{"id":"W4410060872","doi":"10.1038/s41746-025-01589-z","title":"Expert of Experts Verification and Alignment (EVAL) Framework for Large Language Models Safety in Gastroenterology","year":2025,"lang":"en","type":"article","venue":"npj Digital Medicine","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":6,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"National Institute of Diabetes and Digestive and Kidney Diseases; National Institutes of Health","keywords":"Computer science; Medical physics; Medicine; Natural language processing; Internal medicine","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":[0.009250721,0.001875153,0.001377195,0.001664814,0.001066668,0.002150674,0.00318022,0.002360767,0.01276107],"category_scores_gemma":[0.03606043,0.0007911164,0.002005891,0.0007587763,0.0009759152,0.002972987,0.004018543,0.003615672,0.006465891],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001518843,"about_ca_system_score_gemma":0.002856454,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.005676797,"about_ca_topic_score_gemma":0.01151155,"domain_scores_codex":[0.992338,0.004406266,0.0004539587,0.001586133,0.0009009994,0.000314749],"domain_scores_gemma":[0.9833808,0.01139248,0.0007115519,0.002413313,0.001679825,0.0004219179],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.001379087,0.0005061082,0.004455989,0.0008595025,0.0003661488,0.0006669656,0.001045947,0.1611494,0.01157512,0.02270136,0.05544133,0.7398531],"study_design_scores_gemma":[0.0001926524,0.0001454973,0.0003845314,0.00005348546,0.00004357744,0.0001534962,0.0001337316,0.9462257,0.006097147,0.03519193,0.01134067,0.00003758767],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01112864,0.0006029986,0.9352186,0.0007146971,0.0001937337,0.0003597378,0.001293362,0.04796219,0.00252611],"genre_scores_gemma":[0.2482141,0.0002032719,0.7353705,0.001277528,0.000212673,0.0006524129,0.006642273,0.002948299,0.00447899],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01276107,"threshold_uncertainty_score":0.04892302,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0731186697190006,"score_gpt":0.4205977627554838,"score_spread":0.3474790930364832,"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."}}