{"id":"W4412075512","doi":"10.1101/2025.07.06.25330972","title":"Benchmarking Multimodal Large Language Models for Forensic Science and Medicine: A Comprehensive Dataset and Evaluation Framework","year":2025,"lang":"en","type":"preprint","venue":"medRxiv","topic":"Topic Modeling","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"McMaster University; Loyalist College; Queen's University","funders":"","keywords":"Benchmarking; Computer science; Data science; Forensic science; Natural language processing; Artificial intelligence; Medicine; Business","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":[0.02017716,0.003323047,0.001305158,0.005701229,0.001402864,0.003011841,0.005394757,0.0035781,0.003614666],"category_scores_gemma":[0.04021333,0.0006022285,0.00283007,0.00367773,0.00199349,0.003372334,0.004612555,0.003129138,0.002794935],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.004140098,"about_ca_system_score_gemma":0.003255188,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.02252788,"about_ca_topic_score_gemma":0.02713411,"domain_scores_codex":[0.9863528,0.008557905,0.001055836,0.001961368,0.001686916,0.0003850846],"domain_scores_gemma":[0.9746595,0.01568617,0.001058783,0.004510453,0.003276946,0.0008080167],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.002444,0.004338684,0.07568038,0.00424357,0.002474803,0.001310695,0.001352628,0.3466519,0.004636257,0.007393782,0.2078277,0.3416456],"study_design_scores_gemma":[0.0007421466,0.001765422,0.03627633,0.00128089,0.0007256126,0.001228985,0.001818609,0.8741311,0.01047274,0.01377266,0.05744185,0.0003436909],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"empirical","genre_gemma":"methods","genre_scores_codex":[0.6317766,0.01403319,0.1365347,0.00623743,0.001031237,0.005280181,0.1671169,0.02185828,0.01613134],"genre_scores_gemma":[0.5109302,0.001557418,0.1204327,0.001256216,0.0002113139,0.003330868,0.35817,0.0009370788,0.003174129],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.02252788,"threshold_uncertainty_score":0.1067083,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05501007735612667,"score_gpt":0.3614847540115834,"score_spread":0.3064746766554567,"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."}}