{"id":"W4412629089","doi":"10.2196/66476","title":"Improving Large Language Models’ Summarization Accuracy by Adding Highlights to Discharge Notes: Comparative Evaluation","year":2025,"lang":"en","type":"article","venue":"JMIR Medical Informatics","topic":"Machine Learning in Healthcare","field":"Computer Science","cited_by":13,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"","keywords":"Automatic summarization; Terminology; Computer science; Process (computing); Interface (matter); Artificial intelligence; Test (biology); Information retrieval; Natural language processing; Data science; Linguistics; Programming language","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":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001372428,0.0001944322,0.0002699185,0.0002362826,0.0002794082,0.000226314,0.0009372256,0.0001764588,0.00007222359],"category_scores_gemma":[0.001259573,0.0001652681,0.00004722968,0.0008102324,0.00002518102,0.001364196,0.000550457,0.0004790211,0.00007951225],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002071786,"about_ca_system_score_gemma":0.000423474,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00009002256,"about_ca_topic_score_gemma":0.00004818345,"domain_scores_codex":[0.9970763,0.0001987717,0.0007473134,0.0002209418,0.001323689,0.0004330249],"domain_scores_gemma":[0.9980581,0.0005928929,0.0002661271,0.0005182031,0.0002693209,0.0002953234],"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":[0.00003711033,0.0003681026,0.000999791,0.001341466,0.00008516567,0.000005708312,0.2447692,0.005594133,0.0002387638,0.3459041,0.04173735,0.3589191],"study_design_scores_gemma":[0.0005129424,0.00004398751,0.0001242012,0.0002303426,0.00000764326,0.000001380025,0.00106505,0.9913244,0.00021728,0.0003476077,0.005952902,0.0001722532],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.03041535,0.0001019201,0.9607515,0.003689304,0.0004107261,0.001003176,0.00002248913,0.0002820142,0.003323517],"genre_scores_gemma":[0.9655904,0.000009243912,0.02943414,0.004270243,0.00007888091,0.0002232365,0.0002440841,0.000008991465,0.0001408294],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.9857303,"threshold_uncertainty_score":0.6739438,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02851150607305516,"score_gpt":0.375509264051678,"score_spread":0.3469977579786229,"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."}}