{"id":"W4416635488","doi":"10.1183/13993003.congress-2025.pa2023","title":"Language Models Lag Behind: Inconsistent Alignment with Recent Removal of Race from Clinical Algorithms","year":2025,"lang":"","type":"article","venue":"","topic":"Machine Learning in Healthcare","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of British Columbia","funders":"","keywords":"Race (biology); Construct (python library); Function (biology); Renal function; Language model; Lag","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":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.002425578,0.000534215,0.001176834,0.0002235829,0.0002271581,0.0001784057,0.001786581,0.0004056488,0.0003840206],"category_scores_gemma":[0.0002899579,0.000444033,0.0002797164,0.0006506444,0.0003274428,0.0003416464,0.00160925,0.001187463,0.0000373827],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002935655,"about_ca_system_score_gemma":0.001512462,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00922301,"about_ca_topic_score_gemma":0.001039215,"domain_scores_codex":[0.992506,0.001593815,0.002171767,0.001711892,0.001249582,0.000766923],"domain_scores_gemma":[0.9945374,0.001206866,0.0007871942,0.002494492,0.0005469619,0.0004270626],"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":[0.0003696814,0.001075296,0.01102319,0.0002725818,0.0005646066,0.0006473395,0.00888548,0.0129464,0.0000331481,0.02059427,0.001667357,0.9419206],"study_design_scores_gemma":[0.002392271,0.0009262266,0.006016524,0.000937449,0.0001571018,0.00005226807,0.001132592,0.9754583,0.0005293174,0.002423775,0.009418031,0.0005561916],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.1017226,0.007504385,0.8312296,0.02914217,0.003202677,0.00152482,0.00007777243,0.0002632323,0.02533283],"genre_scores_gemma":[0.5159858,0.001581835,0.4684758,0.004150822,0.0003033767,0.00003267189,0.00003276257,0.00004498054,0.009391976],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9625118,"threshold_uncertainty_score":0.9998012,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04179178006067401,"score_gpt":0.3593313345883819,"score_spread":0.3175395545277079,"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."}}