{"id":"W4385571495","doi":"10.18653/v1/2023.acl-long.26","title":"Knowledge of cultural moral norms in large language models","year":2023,"lang":"en","type":"article","venue":"","topic":"Cultural Differences and Values","field":"Psychology","cited_by":46,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Social Sciences and Humanities Research Council of Canada","keywords":"Globe; Morality; Cultural diversity; Inference; Variation (astronomy); Relevance (law); Psychology; Social psychology; Moral disengagement; Sociology; Epistemology; Computer science; Political science; Artificial intelligence","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.00009442124,0.00008191336,0.000154855,0.00006939916,0.00001727995,0.000008369907,0.0001245436,0.0000631283,0.00171456],"category_scores_gemma":[0.000006757445,0.00004957606,0.00006161063,0.0003280671,0.000023046,0.000089647,0.00005583432,0.00007710658,0.0007449008],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000009053795,"about_ca_system_score_gemma":0.000004606882,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.001038483,"about_ca_topic_score_gemma":0.001332608,"domain_scores_codex":[0.9993602,0.00003299342,0.0001641043,0.0001475749,0.00005375603,0.0002413528],"domain_scores_gemma":[0.9997624,0.00002250942,0.00002559417,0.00013063,0.00002601337,0.00003286662],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"theoretical_or_conceptual","study_design_gemma":"observational","study_design_scores_codex":[0.0001062786,0.0007160637,0.1302517,0.0001006622,0.0001016582,0.00005713184,0.3119216,0.0001321988,0.004892507,0.4705673,0.06737582,0.01377711],"study_design_scores_gemma":[0.00185403,0.0001347683,0.8714591,0.00004294269,0.00001263254,0.00000549507,0.1050174,0.01388178,0.0004976984,0.005967031,0.0007977036,0.0003294654],"study_design_candidate":"observational","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.8942545,0.0002045544,0.00001177462,0.00007770928,0.0001938754,0.0000799129,0.00002004171,0.00008165815,0.105076],"genre_scores_gemma":[0.9529054,0.00001316263,0.00002436508,0.00003280789,0.00004396146,0.00001261203,0.00003938746,0.000005820189,0.04692252],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.7412074,"threshold_uncertainty_score":0.999198,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1470952819041261,"score_gpt":0.4171879044689984,"score_spread":0.2700926225648723,"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."}}