{"id":"W4391480596","doi":"10.21203/rs.3.rs-3882757/v1","title":"mCodeGPT: Enhancing Cancer Research through Zero-Shot Information Extraction from Clinical Free Text Data","year":2024,"lang":"en","type":"preprint","venue":"Research Square","topic":"Topic Modeling","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":false,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"","keywords":"Zero (linguistics); Shot (pellet); Extraction (chemistry); Information extraction; Computer science; Cancer; Text messaging; Information retrieval; Medicine; World Wide Web; Linguistics; Materials science; Chromatography; Chemistry; Philosophy; 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":[0.004303664,0.001722956,0.00123475,0.01070354,0.0009008255,0.002542846,0.001665057,0.00219442,0.004899578],"category_scores_gemma":[0.01892317,0.0007088971,0.001909361,0.006184312,0.0005540911,0.003163358,0.002987796,0.001894223,0.004539057],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0006891175,"about_ca_system_score_gemma":0.002901183,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00497267,"about_ca_topic_score_gemma":0.009102267,"domain_scores_codex":[0.9970319,0.001058791,0.0002590591,0.000860964,0.0006259084,0.0001633114],"domain_scores_gemma":[0.9877999,0.008855525,0.0004218418,0.001322811,0.001227292,0.0003727277],"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":[0.0008820479,0.0004394806,0.007649667,0.003087038,0.0008248953,0.000584638,0.0007898837,0.007617787,0.02950667,0.005549534,0.1209452,0.8221231],"study_design_scores_gemma":[0.0006492828,0.0009943031,0.02043791,0.0009169359,0.002074305,0.002487459,0.001048674,0.5184645,0.09354809,0.08427601,0.2747779,0.0003247032],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.04508899,0.01375552,0.8181401,0.004000551,0.001258162,0.001298952,0.05808881,0.05191338,0.006455441],"genre_scores_gemma":[0.1316284,0.003467956,0.7736562,0.0008491632,0.0009247556,0.00105591,0.08009613,0.002085413,0.006236076],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01070354,"threshold_uncertainty_score":0.02276021,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.4635973813089228,"score_gpt":0.5652224047004077,"score_spread":0.1016250233914849,"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."}}