{"id":"W4394867428","doi":"10.1017/pan.2025.10007","title":"Measuring the Quality of Answers in Political Q&As with Large Language Models","year":2025,"lang":"en","type":"preprint","venue":"Political Analysis","topic":"Expert finding and Q&A systems","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"","keywords":"Operationalization; Quality (philosophy); Parallels; House of Commons; Context (archaeology); Relevance (law); Computer science; Period (music); Politics; Measure (data warehouse); Parliament; Epistemology; Political science; Data mining; Law; History","routes":{"ca_aff":false,"ca_fund":false,"ca_venue":false,"about_ca":true,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.001634916,0.0003185586,0.001064439,0.000646494,0.00009192605,0.0001464044,0.001597221,0.0002573455,0.00001467934],"category_scores_gemma":[0.000324564,0.0002095391,0.0005620116,0.001386967,0.0001758988,0.000102741,0.001092583,0.0006874245,0.00001011739],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002613511,"about_ca_system_score_gemma":0.0003603607,"about_ca_topic_candidate":true,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.03664519,"about_ca_topic_score_gemma":0.003291837,"domain_scores_codex":[0.9956645,0.0007371026,0.0008467938,0.0007943606,0.0008982433,0.001058965],"domain_scores_gemma":[0.9968624,0.0006770535,0.0001755289,0.001802901,0.0001854328,0.0002967015],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.000008882937,0.000131109,0.009648301,0.0001631157,0.0006315869,0.000009465315,0.002286409,0.002251992,0.00001076556,0.9847986,0.00001119575,0.00004865534],"study_design_scores_gemma":[0.001693229,0.000106367,0.02787671,0.001364091,0.002326624,0.00001859019,0.008673842,0.6365117,0.001818973,0.3170226,0.0003158162,0.002271389],"study_design_candidate":"theoretical_or_conceptual","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.267894,0.0005524246,0.6754288,0.007077771,0.0001939637,0.0003774538,0.0002220972,0.0001826277,0.04807083],"genre_scores_gemma":[0.9960006,0.000002817005,0.002644914,0.0005263023,0.00007390649,0.00004360362,0.00002376304,0.000008750606,0.0006753296],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.7281066,"threshold_uncertainty_score":0.9697699,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07669152858246686,"score_gpt":0.3559642944476354,"score_spread":0.2792727658651686,"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."}}