{"id":"W4416372001","doi":"10.1162/tacl.a.50","title":"Objectifying the Subjective: Cognitive Biases in Topic Interpretations","year":2025,"lang":"en","type":"article","venue":"Transactions of the Association for Computational Linguistics","topic":"Topic Modeling","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Victoria","funders":"","keywords":"Representativeness heuristic; Interpretation (philosophy); Salient; Cognitive bias; Quality (philosophy); Heuristics; Context (archaeology); Coherence (philosophical gambling strategy); Cognition","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":[0.0004083251,0.00007549632,0.0001140907,0.0001323225,0.0002972237,0.0000552227,0.0003902423,0.00004962611,0.000001613349],"category_scores_gemma":[0.005029233,0.00005988994,0.0001199832,0.0005818768,0.00003042552,0.00005728954,0.0000236646,0.0001559024,9.287535e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002322242,"about_ca_system_score_gemma":0.0002448569,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00008242325,"about_ca_topic_score_gemma":0.0001609477,"domain_scores_codex":[0.999079,0.00009858641,0.0003238317,0.0001594382,0.0002070864,0.0001320973],"domain_scores_gemma":[0.9941946,0.004674987,0.0001908059,0.0001587728,0.000768732,0.00001213156],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00001012845,0.00009237645,0.007759156,0.00002284722,0.0001050536,9.780367e-8,0.001481706,0.8192902,0.000006784745,0.1676645,0.00004972922,0.003517384],"study_design_scores_gemma":[0.0005168439,0.0000185458,0.03330974,0.0001668725,0.00005944441,4.151779e-7,0.0002435469,0.8764617,0.0002978895,0.08840924,0.0004325226,0.00008321278],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.004096439,0.00004329795,0.9910735,0.001106034,0.001200813,0.000364375,0.00003658838,0.00003413301,0.002044816],"genre_scores_gemma":[0.9864897,0.000002416664,0.01262566,0.000192305,0.00004211275,0.00004791291,0.000005553131,0.000003842927,0.0005905411],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9823932,"threshold_uncertainty_score":0.6020826,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.02274245253839106,"score_gpt":0.3057547969801711,"score_spread":0.28301234444178,"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."}}