{"id":"W4415299179","doi":"10.3390/make7040121","title":"Small or Large? Zero-Shot or Finetuned? Guiding Language Model Choice for Specialized Applications in Healthcare","year":2025,"lang":"en","type":"article","venue":"Machine Learning and Knowledge Extraction","topic":"Topic Modeling","field":"Computer Science","cited_by":6,"is_retracted":false,"has_abstract":true,"ca_institutions":"Provincial Health Services Authority; University of British Columbia","funders":"","keywords":"Task (project management); Language model; Selection (genetic algorithm); Exploit; Health care; Language understanding; Task analysis","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":[],"consensus_categories":[],"category_scores_codex":[0.0005115966,0.0001471271,0.0002047015,0.000231641,0.0003428802,0.0001061881,0.0002157173,0.0001012265,0.000008383967],"category_scores_gemma":[0.0003433321,0.0001232899,0.00004276591,0.000402425,0.000009426263,0.0001696435,0.0001161043,0.0002778913,0.000003725165],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00008925523,"about_ca_system_score_gemma":0.000147939,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000386648,"about_ca_topic_score_gemma":0.00501388,"domain_scores_codex":[0.9988254,0.0001007959,0.0002946201,0.0004403066,0.00006733966,0.0002715455],"domain_scores_gemma":[0.99913,0.0004206529,0.00009331154,0.0002247865,0.00007306241,0.00005814887],"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":[0.000399227,0.0006035588,0.02430178,0.001213567,0.00005570584,0.00001014672,0.008722996,0.01712664,0.004822893,0.09361891,0.0008033315,0.8483213],"study_design_scores_gemma":[0.001038741,0.00003948559,0.0005890257,0.00007373002,0.000009244661,0.000003905101,0.0001116043,0.9364247,0.000154678,0.0008833042,0.06053272,0.0001389007],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.0148935,0.001707445,0.9798022,0.001082123,0.0002149903,0.0004560252,0.000004434678,0.0001939215,0.001645376],"genre_scores_gemma":[0.9104095,0.000153839,0.05885646,0.0001687771,0.0002367261,0.0002561503,0.0000398154,0.00002123474,0.02985751],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9209457,"threshold_uncertainty_score":0.5027618,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07084545954893032,"score_gpt":0.3752476515817091,"score_spread":0.3044021920327787,"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."}}