{"id":"W4310413652","doi":"10.48550/arxiv.2211.14402","title":"An Analysis of Social Biases Present in BERT Variants Across Multiple Languages","year":2022,"lang":"en","type":"preprint","venue":"arXiv (Cornell University)","topic":"Natural Language Processing Techniques","field":"Computer Science","cited_by":5,"is_retracted":false,"has_abstract":true,"ca_institutions":"McGill University","funders":"","keywords":"Computer science; Natural language processing; Persian; Artificial intelligence; Set (abstract data type); Linguistics; Adjective; Noun","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":"metacan-v3-hybrid-931329e0061c","candidate_categories":[],"consensus_categories":[],"category_scores_codex":[0.003278295,0.0004695849,0.0003202,0.0008920549,0.0005768618,0.001153888,0.0005009469,0.0004119818,0.001181748],"category_scores_gemma":[0.0131544,0.0002669363,0.0004047212,0.0009751391,0.0008404151,0.001686089,0.001132993,0.0006744916,0.0003372052],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0007637274,"about_ca_system_score_gemma":0.0004981732,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.006322127,"about_ca_topic_score_gemma":0.01230135,"domain_scores_codex":[0.9986963,0.0006912196,0.00007796865,0.0001815665,0.0002631239,0.00008990301],"domain_scores_gemma":[0.9900563,0.006413202,0.0006914636,0.00127296,0.001342085,0.0002240837],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001131386,0.0002541763,0.645119,0.0003556875,0.0005539572,0.001743449,0.01577297,0.1247775,0.03916134,0.02021063,0.004520679,0.1463993],"study_design_scores_gemma":[0.00004961384,0.000394827,0.1856961,0.00008241462,0.0002309738,0.0013939,0.007718267,0.7428669,0.02697757,0.02422752,0.01018224,0.0001797449],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9845575,0.0001027016,0.01185049,0.0001602681,0.00001385049,0.00002091036,0.0002942652,0.0002022488,0.002797765],"genre_scores_gemma":[0.995139,0.0000411738,0.003678157,0.0000258843,0.000005812689,0.0000154684,0.0004580411,0.00007702725,0.0005594649],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.006322127,"threshold_uncertainty_score":0.0173375,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0646907881846757,"score_gpt":0.2729201547505337,"score_spread":0.2082293665658579,"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."}}