{"id":"W4412945237","doi":"10.18653/v1/2025.acl-long.657","title":"Code-Switching Red-Teaming: LLM Evaluation for Safety and Multilingual Understanding","year":2025,"lang":"en","type":"article","venue":"","topic":"Software Reliability and Analysis Research","field":"Computer Science","cited_by":2,"is_retracted":false,"has_abstract":true,"ca_institutions":"Kootenay Association for Science & Technology","funders":"","keywords":"Computer science; Code-switching; Code (set theory); Programming language; Software engineering; Linguistics","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":[0.002588894,0.00008118453,0.0001440286,0.0001836522,0.0003510599,0.0002062077,0.0002862064,0.00005516904,0.000009784171],"category_scores_gemma":[0.001123896,0.00007088909,0.00006249681,0.0004314885,0.00002950401,0.0002473475,0.000188222,0.00009822808,0.000001248488],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0002262755,"about_ca_system_score_gemma":0.0001695215,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003463421,"about_ca_topic_score_gemma":0.00007270071,"domain_scores_codex":[0.9987397,0.0001032168,0.0002267267,0.0003951415,0.0003267502,0.000208473],"domain_scores_gemma":[0.9984322,0.001015293,0.0000352597,0.000296408,0.0001700776,0.0000507465],"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":[0.00006397431,0.000111622,0.004090408,0.0001346449,0.0001227088,0.000001100253,0.00111459,0.002024952,0.001328544,0.1779965,0.0003203087,0.8126907],"study_design_scores_gemma":[0.0005600927,0.00003908704,0.0006409047,0.00003280889,0.00001965385,6.328522e-7,0.0003698484,0.9599566,0.0008642204,0.03656495,0.000868529,0.00008262239],"study_design_candidate":"simulation_or_modeling","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01369374,0.0001224559,0.9818333,0.001927965,0.00008991839,0.0003206496,9.323562e-7,0.0001003867,0.001910661],"genre_scores_gemma":[0.97045,0.00003152066,0.02893784,0.00009758592,0.00001856542,0.00001487419,0.000003076722,0.000003171515,0.0004433264],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9579317,"threshold_uncertainty_score":0.2890773,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.08500175203487303,"score_gpt":0.383077898694948,"score_spread":0.298076146660075,"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."}}