{"id":"W4411119063","doi":"10.18653/v1/2025.nlp4dh-1.25","title":"Evaluating Large Language Models for Narrative Topic Labeling","year":2025,"lang":"en","type":"article","venue":"","topic":"Topic Modeling","field":"Computer Science","cited_by":3,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"Social Sciences and Humanities Research Council of Canada","keywords":"Computer science; Narrative; Natural language processing; Linguistics; Philosophy","routes":{"ca_aff":true,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":false},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.02049813,0.002295603,0.001352343,0.002970815,0.001423078,0.002871562,0.00229536,0.002221402,0.002305634],"category_scores_gemma":[0.04903544,0.0007775653,0.001245385,0.001595454,0.0009400867,0.003954964,0.002318912,0.002195018,0.001726281],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.002643887,"about_ca_system_score_gemma":0.00170714,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01089106,"about_ca_topic_score_gemma":0.0182257,"domain_scores_codex":[0.9873039,0.009083004,0.0005825428,0.001841156,0.0009152361,0.0002741064],"domain_scores_gemma":[0.9430758,0.04974148,0.001253784,0.002727268,0.002388973,0.0008127436],"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":[0.003522445,0.001458769,0.02687952,0.001415092,0.001338762,0.0003691376,0.002229299,0.4201687,0.01141976,0.005584759,0.01490818,0.5107056],"study_design_scores_gemma":[0.00009964107,0.0002333441,0.001663791,0.00004506754,0.00007756399,0.00005279185,0.0004075357,0.9880224,0.003906856,0.003698011,0.001758999,0.00003383382],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.5437117,0.005896426,0.4232285,0.002163096,0.000534638,0.001190655,0.00393212,0.01162902,0.007713916],"genre_scores_gemma":[0.8010924,0.0005338225,0.1870659,0.0003528987,0.0002009278,0.0006317113,0.007258351,0.0004355785,0.00242838],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.02049813,"threshold_uncertainty_score":0.1084058,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0665263816001478,"score_gpt":0.3842538718380301,"score_spread":0.3177274902378823,"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."}}