{"id":"W4412070257","doi":"10.1136/oemed-2024-epicohabstracts.77","title":"77 The use of a large language model to identify industry-funded research in carcinogenic hazard evaluations","year":2025,"lang":"en","type":"article","venue":"","topic":"Occupational Health and Safety Research","field":"Health Professions","cited_by":0,"is_retracted":false,"has_abstract":false,"ca_institutions":"McGill University; Occupational Cancer Research Centre","funders":"","keywords":"Hazard; Computer science; Hazard analysis; Risk analysis (engineering); Business; Engineering; Reliability engineering","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":["metaresearch"],"consensus_categories":[],"category_scores_codex":[0.01158762,0.001003737,0.000641123,0.005149602,0.001447612,0.003628183,0.001316111,0.001297744,0.01054931],"category_scores_gemma":[0.03706104,0.0004372942,0.001860337,0.002442466,0.001139713,0.00428071,0.002604985,0.001487089,0.002381424],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002728519,"about_ca_system_score_gemma":0.005313871,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01564927,"about_ca_topic_score_gemma":0.02151987,"domain_scores_codex":[0.9918207,0.005594575,0.0004399845,0.0009462361,0.0008259832,0.0003725523],"domain_scores_gemma":[0.9445143,0.04482941,0.002708485,0.002688862,0.004315727,0.000943266],"domain_codex":null,"domain_gemma":"evaluation","domain_candidate":"evaluation","domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"observational","study_design_scores_codex":[0.003063363,0.002062671,0.2284083,0.001365324,0.001212279,0.001641148,0.003556703,0.09381441,0.01141181,0.2465854,0.02585209,0.3810266],"study_design_scores_gemma":[0.0003348751,0.0006038923,0.03840478,0.0003643591,0.0008267339,0.0007061439,0.00191089,0.754469,0.007072425,0.1635641,0.03156134,0.0001813768],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1679105,0.0007289697,0.7845712,0.003750534,0.0002434029,0.0009088828,0.005273581,0.003074288,0.03353866],"genre_scores_gemma":[0.7663326,0.0002268315,0.219014,0.0004999323,0.00006297982,0.0007733559,0.003989065,0.000312906,0.008788476],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9884124,"threshold_uncertainty_score":0.06128192,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.5534436144800086,"score_gpt":0.6666495771738191,"score_spread":0.1132059626938106,"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."}}