{"id":"W7125586192","doi":"10.1109/cascon66301.2025.00111","title":"ScenarioBench: Trace-Grounded Compliance Evaluation for Text-to-SQL and RAG","year":2025,"lang":"","type":"article","venue":"","topic":"Scientific Computing and Data Management","field":"Decision Sciences","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Ontario Tech University","funders":"","keywords":"Correctness; Benchmark (surveying); Artifact (error); Code (set theory); Matching (statistics); Aggregate (composite); Witness; Core (optical fiber); Compliance (psychology)","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":["metaepi_narrow","scholarly_communication","insufficient_payload"],"consensus_categories":[],"category_scores_codex":[0.01973557,0.0003285798,0.0004945191,0.0008655015,0.001173693,0.003363966,0.001492887,0.0001097979,0.001281098],"category_scores_gemma":[0.006068741,0.0002788248,0.0001527687,0.002864676,0.0002386207,0.0005461549,0.001112548,0.0001369356,0.0004716837],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000171858,"about_ca_system_score_gemma":0.0003920061,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.0001357314,"about_ca_topic_score_gemma":0.0002873504,"domain_scores_codex":[0.9933668,0.0003218607,0.001268674,0.002330187,0.002056412,0.0006561112],"domain_scores_gemma":[0.993999,0.002385877,0.0002610525,0.001871176,0.00121315,0.0002697461],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00008343891,0.000125592,0.0001843658,0.00004773956,0.00003227141,5.2364e-7,0.0004039822,0.001002252,0.00006391142,0.02026198,0.2126815,0.7651125],"study_design_scores_gemma":[0.001384289,0.0001210325,0.01057288,0.0002588184,0.0001138188,0.000001293633,0.001863692,0.6321278,0.00008988659,0.03043966,0.322744,0.0002829337],"study_design_candidate":"design_other","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.02728416,0.002189518,0.8698044,0.0221934,0.008218068,0.00445383,0.00008556474,0.00009916013,0.06567191],"genre_scores_gemma":[0.8988715,0.00003363598,0.02423344,0.00194573,0.0001125653,0.0001114365,0.0000203995,0.00001121086,0.07466004],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.8715874,"threshold_uncertainty_score":0.9999664,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.316861616734726,"score_gpt":0.4856278158292047,"score_spread":0.1687661990944787,"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."}}