{"id":"W4414760613","doi":"10.2196/82484","title":"From Context to Care: Rethinking Stigma Detection in Clinical Language Models","year":2025,"lang":"en","type":"article","venue":"JMIR Medical Informatics","topic":"Migration, Health and Trauma","field":"Psychology","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"U.S. National Library of Medicine","keywords":"Stigma (botany); Context (archaeology); Citizen journalism; Health records; Natural language; Situational ethics; Pragmatics; Language understanding","routes":{"ca_aff":false,"ca_fund":false,"ca_venue":true,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.01767871,0.0009659919,0.0007669941,0.002146922,0.002193412,0.007641247,0.002936077,0.002017668,0.003177621],"category_scores_gemma":[0.104542,0.0008223201,0.001765125,0.001438557,0.00369951,0.00931887,0.005875273,0.005834526,0.0007560347],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002560296,"about_ca_system_score_gemma":0.004176271,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.009472577,"about_ca_topic_score_gemma":0.01573314,"domain_scores_codex":[0.9810028,0.01511229,0.0009676379,0.001393452,0.00124758,0.0002762535],"domain_scores_gemma":[0.9060101,0.08053003,0.003736365,0.005208926,0.003756526,0.0007581104],"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":[0.0008419744,0.0003967129,0.06331636,0.002829461,0.0005747717,0.00127693,0.04443574,0.1286676,0.004355778,0.2557265,0.01799731,0.4795809],"study_design_scores_gemma":[0.00005634613,0.0001246092,0.004111831,0.002078125,0.0001980765,0.0006206225,0.008735261,0.567506,0.003043559,0.3805796,0.03277048,0.0001754523],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.05303764,0.002134262,0.906817,0.02777702,0.0004452238,0.0004273835,0.0008132379,0.001289423,0.007258922],"genre_scores_gemma":[0.6213761,0.0008866454,0.3738739,0.001804724,0.0001787493,0.0004477138,0.0005834707,0.0002259749,0.0006227648],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.01767871,"threshold_uncertainty_score":0.09349507,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.03895908008300044,"score_gpt":0.4138745977907626,"score_spread":0.3749155177077622,"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."}}