{"id":"W7125948853","doi":"10.1109/ase63991.2025.00086","title":"Coverage-Based Harmfulness Testing for LLM Code Transformation","year":2025,"lang":"","type":"article","venue":"","topic":"Software Engineering Research","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"Concordia University","funders":"","keywords":"Moderation; Code (set theory); Set (abstract data type); Source code; Harm; Key (lock); Software; Code review","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":[0.004898579,0.001679824,0.0008261545,0.002073958,0.0006181037,0.001108286,0.002283653,0.001383676,0.001915072],"category_scores_gemma":[0.05194725,0.0007079,0.001209093,0.0006694965,0.00223822,0.002803846,0.002296186,0.001737975,0.0006529441],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00129568,"about_ca_system_score_gemma":0.002915322,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.004755412,"about_ca_topic_score_gemma":0.006258623,"domain_scores_codex":[0.9881187,0.003496401,0.0006724148,0.002101157,0.005025486,0.0005859187],"domain_scores_gemma":[0.9472181,0.03529661,0.004550388,0.006975554,0.005186506,0.0007729405],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"design_other","study_design_gemma":"bench_or_experimental","study_design_scores_codex":[0.002173131,0.001718142,0.08573601,0.001842691,0.0003355408,0.001320501,0.003317941,0.1467094,0.1927231,0.01207137,0.007255949,0.5447963],"study_design_scores_gemma":[0.0001541774,0.001311957,0.01320989,0.0001779268,0.0001574795,0.0007755693,0.0004153993,0.8695294,0.0974845,0.01136491,0.005302481,0.0001162831],"study_design_candidate":"bench_or_experimental","study_design_consensus":null,"genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.3617323,0.0008434494,0.5917413,0.0006462155,0.00008669402,0.0009694138,0.0009922122,0.0389977,0.003990769],"genre_scores_gemma":[0.7383638,0.0001319289,0.2574942,0.0002761186,0.00001986373,0.0004011171,0.00102678,0.001312602,0.0009735443],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.004898579,"threshold_uncertainty_score":0.02590644,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04030988784248639,"score_gpt":0.3051761466031735,"score_spread":0.2648662587606871,"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."}}