{"id":"W4415230910","doi":"10.21203/rs.3.rs-7623811/v1","title":"Responsible AI in NLP: GUS-Net Span-Level Bias Detection Dataset and Benchmark for Generalizations, Unfairness, and Stereotypes","year":2025,"lang":"en","type":"preprint","venue":"Research Square","topic":"Explainable Artificial Intelligence (XAI)","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":false,"ca_institutions":"Toronto Metropolitan University","funders":"","keywords":"Benchmark (surveying); Pipeline (software); Language model; Limiting; Textual entailment; Identification (biology); Security token; Framing (construction)","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":[0.00376345,0.001844176,0.0007886708,0.002786468,0.001928562,0.002496109,0.003285746,0.004167511,0.01694923],"category_scores_gemma":[0.02361835,0.0004734195,0.001237786,0.002313479,0.001535631,0.003585718,0.003412641,0.003320814,0.01818044],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.001591412,"about_ca_system_score_gemma":0.002705069,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.01786088,"about_ca_topic_score_gemma":0.02928089,"domain_scores_codex":[0.9964069,0.001087034,0.0002733495,0.0008955234,0.001088942,0.0002483112],"domain_scores_gemma":[0.981572,0.00814139,0.0007508248,0.005831705,0.002625854,0.00107816],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"not_applicable","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.001731134,0.001101879,0.01733029,0.001388518,0.0003288504,0.0005390503,0.000626976,0.01432751,0.003445429,0.01261062,0.8689054,0.07766446],"study_design_scores_gemma":[0.002797923,0.0007295077,0.04567637,0.0006981148,0.0004267645,0.001807678,0.00270447,0.2362811,0.02475953,0.09679698,0.5870563,0.0002652994],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"dataset","genre_gemma":"dataset","genre_scores_codex":[0.1867256,0.003022419,0.03854545,0.006705494,0.001176816,0.0008649099,0.6629463,0.04515529,0.05485784],"genre_scores_gemma":[0.150095,0.000328256,0.03242164,0.001288762,0.0001650172,0.0004692439,0.8038294,0.001357833,0.01004475],"genre_candidate":"dataset","genre_consensus":"dataset","teacher_disagreement_score":0.01786088,"threshold_uncertainty_score":0.05670083,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.1935936064666533,"score_gpt":0.439637199761039,"score_spread":0.2460435932943857,"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."}}