{"id":"W4402671660","doi":"10.18653/v1/2024.acl-long.600","title":"Measuring Political Bias in Large Language Models: What Is Said and How It Is Said","year":2024,"lang":"en","type":"article","venue":"","topic":"Hate Speech and Cyberbullying Detection","field":"Computer Science","cited_by":24,"is_retracted":false,"has_abstract":true,"ca_institutions":"Artificial Intelligence in Medicine (Canada)","funders":"","keywords":"Politics; Computer science; Natural language processing; Political science; Law","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":["scholarly_communication"],"consensus_categories":[],"category_scores_codex":[0.0003981324,0.0001318238,0.000129594,0.0001857614,0.00006106913,0.001213588,0.0002263902,0.00008362937,0.00006501376],"category_scores_gemma":[0.0000206224,0.0001129357,0.00005201982,0.000341759,0.0000197702,0.001707085,0.0001815942,0.0001916539,0.0001027813],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00007061978,"about_ca_system_score_gemma":0.00004432645,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0002480689,"about_ca_topic_score_gemma":0.0001961692,"domain_scores_codex":[0.9986525,0.00004857224,0.0001163821,0.0004437398,0.000275767,0.0004630209],"domain_scores_gemma":[0.9994603,0.00006138188,0.00001146096,0.0002954351,0.00002457641,0.0001468786],"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":[0.0000196177,0.0001847265,0.0006742222,0.000421979,0.0001090493,0.0009239375,0.0338222,0.00003561788,0.005813102,0.5382649,0.01730275,0.4024279],"study_design_scores_gemma":[0.0006629761,0.0001414292,0.0002864853,0.0005783858,0.00001630245,0.0001921181,0.005310897,0.8840726,0.06512253,0.01885421,0.02407347,0.0006885787],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.4493504,0.009843184,0.4195672,0.08631635,0.001433704,0.0004356056,0.00001801339,0.001256534,0.03177896],"genre_scores_gemma":[0.9899302,0.0001817836,0.003441117,0.003191034,0.0000835696,0.000006956867,8.645422e-7,0.00001209526,0.003152355],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.884037,"threshold_uncertainty_score":0.9998233,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.05210208125202521,"score_gpt":0.263815868988379,"score_spread":0.2117137877363538,"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."}}