{"id":"W4388205631","doi":"10.1200/op.2023.19.11_suppl.511","title":"Large language models encode medical oncology knowledge: Performance on the ASCO and ESMO examination questions.","year":2023,"lang":"en","type":"article","venue":"JCO Oncology Practice","topic":"Artificial Intelligence in Healthcare and Education","field":"Medicine","cited_by":6,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Toronto; University Health Network; Princess Margaret Cancer Centre","funders":"","keywords":"Medicine; Confidence interval; Internal medicine; Oncology; Family 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":"codex-gemma-dda1882f352a","candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.004816356,0.0001159151,0.0002204061,0.0001639388,0.0003507389,0.00001449847,0.0001145079,0.0003935552,0.0002396364],"category_scores_gemma":[0.007666518,0.00008481021,0.00002955373,0.0004201392,0.0001665524,0.000296206,0.00007132832,0.0008161451,0.0008577434],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0003608768,"about_ca_system_score_gemma":0.001123282,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002163139,"about_ca_topic_score_gemma":0.0003815442,"domain_scores_codex":[0.997926,0.0008386311,0.0003526154,0.0002727813,0.000287994,0.0003219952],"domain_scores_gemma":[0.9931034,0.006012079,0.0001644929,0.0002656975,0.0002799733,0.0001743404],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"not_applicable","study_design_scores_codex":[0.0004658413,0.001591751,0.001480146,0.0001268312,0.00007295769,0.0002043917,0.04554836,0.00005429571,0.0001628282,0.02954779,0.0249105,0.8958343],"study_design_scores_gemma":[0.0005600307,0.004370068,0.0208346,0.0002655686,0.0002785335,0.001019388,0.07507686,0.1434313,0.00139309,0.001804518,0.7506688,0.0002973259],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8078057,0.0003609018,0.0003902847,0.1189889,0.0009616224,0.0004898388,0.00000259044,0.0001413988,0.07085874],"genre_scores_gemma":[0.9868215,0.002623261,0.0002722743,0.007443768,0.0006054897,0.0001621799,0.00003435585,0.00001635863,0.002020777],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.895537,"threshold_uncertainty_score":0.9999202,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1754834882070465,"score_gpt":0.5117673540927489,"score_spread":0.3362838658857024,"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."}}