{"id":"W4414078785","doi":"10.1097/ijg.0000000000002627","title":"Comparing Performance of Large Language Model-Based Tools on Patient-Driven Glaucoma Inquiries","year":2025,"lang":"en","type":"article","venue":"Journal of Glaucoma","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Health Care Foundation","funders":"","keywords":"Glaucoma; Public health; MEDLINE; Occupational safety and health","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":[],"consensus_categories":[],"category_scores_codex":[0.0002353257,0.0001063138,0.0003625872,0.0002790316,0.00008103524,0.00001588191,0.0001208292,0.0000769013,0.00002988358],"category_scores_gemma":[0.0002262385,0.00008684673,0.0001217019,0.0002251429,0.00008558557,0.000149303,0.00002094946,0.0002951576,0.000007865261],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001301244,"about_ca_system_score_gemma":0.0004145301,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00002122777,"about_ca_topic_score_gemma":0.00001568149,"domain_scores_codex":[0.9986618,0.000043032,0.0007185506,0.00009559155,0.0002850927,0.0001959505],"domain_scores_gemma":[0.9988105,0.0001593012,0.0003590719,0.0002100476,0.000372843,0.00008824578],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.001144208,0.0004644633,0.961935,0.0003071683,0.0000464818,0.00001544676,0.002553515,0.0001611427,0.00337872,0.0006492605,0.001397179,0.02794734],"study_design_scores_gemma":[0.0006924992,0.002909201,0.5274026,0.00285824,0.0001565248,0.00003834342,0.004160944,0.1862988,0.2737367,0.0002169056,0.001337479,0.0001917829],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9946523,0.0002862471,0.002507182,0.001152588,0.0004714029,0.0001581884,0.000003389956,0.00001144493,0.0007572899],"genre_scores_gemma":[0.9976874,0.00003204928,0.001074695,0.0009624277,0.0001374323,0.000003274528,0.000006326753,0.000009089416,0.00008729201],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4345325,"threshold_uncertainty_score":0.3541507,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.08528478988284796,"score_gpt":0.3906838551475755,"score_spread":0.3053990652647276,"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."}}