{"id":"W4229440380","doi":"10.18653/v1/2022.naacl-main.192","title":"Necessity and Sufficiency for Explaining Text Classifiers: A Case Study in Hate Speech Detection","year":2022,"lang":"en","type":"article","venue":"Proceedings of the 2022 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies","topic":"Hate Speech and Cyberbullying Detection","field":"Computer Science","cited_by":27,"is_retracted":false,"has_abstract":true,"ca_institutions":"National Research Council Canada","funders":"","keywords":"Computer science; Computational linguistics; Voice activity detection; Artificial intelligence; Linguistics; Speech recognition; Natural language processing; Speech processing; Philosophy","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":[0.0007167261,0.0001309774,0.0002588078,0.000179699,0.0006088582,0.00005153409,0.00112057,0.00003267153,0.000001059121],"category_scores_gemma":[0.002168963,0.00009409559,0.0001219227,0.0006565716,0.0002137756,0.00006833997,0.0008800454,0.0002539293,4.579496e-8],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001418399,"about_ca_system_score_gemma":0.00004843205,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0004094903,"about_ca_topic_score_gemma":0.0005200715,"domain_scores_codex":[0.9986504,0.00003463233,0.0004142107,0.0002725295,0.0004480412,0.0001801589],"domain_scores_gemma":[0.9973777,0.0002453116,0.001501179,0.000204978,0.0006582609,0.0000125695],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"observational","study_design_gemma":"observational","study_design_scores_codex":[0.0002462248,0.001018747,0.7305447,0.0004545812,0.0004362992,0.000007751369,0.03920165,0.009314743,0.01029662,0.1053221,0.00009343179,0.1030632],"study_design_scores_gemma":[0.005569263,0.005394071,0.2911175,0.0003645533,0.0005468715,0.0001996089,0.2666568,0.2563078,0.06940231,0.1023818,0.0004889181,0.001570459],"study_design_candidate":"observational","study_design_consensus":"observational","genre_codex":"empirical","genre_gemma":"empirical","genre_scores_codex":[0.9970115,0.00002356905,0.001320968,0.000295799,0.0002418711,0.0008815886,0.00004927682,0.00007179364,0.0001035899],"genre_scores_gemma":[0.9974834,0.000001486779,0.002314937,0.00001778495,0.000019139,0.00009643019,0.000001079193,0.000008639719,0.0000571552],"genre_candidate":"empirical","genre_consensus":"empirical","teacher_disagreement_score":0.4394272,"threshold_uncertainty_score":0.4682907,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.0191876154053738,"score_gpt":0.2586639038828571,"score_spread":0.2394762884774833,"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."}}