{"id":"W4417229639","doi":"10.1061/9780784486139.076","title":"Can Large Vision-Language Models Understand Construction Safety? A Novel Benchmark Using Construction Safety Posters","year":2025,"lang":"","type":"article","venue":"","topic":"Occupational Health and Safety Research","field":"Health Professions","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia","funders":"","keywords":"Benchmarking; Task (project management); Software deployment; Benchmark (surveying); The Internet; Natural language; Key (lock); Work (physics)","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":[0.003331561,0.001828961,0.0006889566,0.001623141,0.0006403304,0.001874311,0.002275302,0.002400961,0.002920387],"category_scores_gemma":[0.01332473,0.0004612634,0.001215094,0.0007344203,0.001225063,0.003460354,0.001987578,0.002867388,0.001596373],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001882655,"about_ca_system_score_gemma":0.001799784,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.02134516,"about_ca_topic_score_gemma":0.02258997,"domain_scores_codex":[0.9975231,0.000922686,0.0001233113,0.000804158,0.000414178,0.0002125576],"domain_scores_gemma":[0.9938709,0.003687656,0.000232409,0.0009703612,0.0009061692,0.0003325085],"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":[0.001539687,0.003495599,0.03722247,0.001443913,0.0005668424,0.001015798,0.002678743,0.3966837,0.03542604,0.004876154,0.03287598,0.4821751],"study_design_scores_gemma":[0.00008192823,0.0006602927,0.01118712,0.0001231779,0.00008069359,0.0001775842,0.001178308,0.9599104,0.01343059,0.005669277,0.007416257,0.00008442401],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8071373,0.002379915,0.1594205,0.001975872,0.0006733896,0.0005110815,0.004721924,0.01098781,0.01219224],"genre_scores_gemma":[0.9052721,0.0003657615,0.07886422,0.0005250897,0.00009135275,0.0001916336,0.01125365,0.0003553675,0.003080669],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.02134516,"threshold_uncertainty_score":0.04244184,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.06443117792109902,"score_gpt":0.4304245028225262,"score_spread":0.3659933249014272,"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."}}