{"id":"W4413219439","doi":"10.1007/s10664-025-10693-3","title":"Adversarial attack classification and robustness testing for large language models for code","year":2025,"lang":"en","type":"article","venue":"Empirical Software Engineering","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":1,"is_retracted":false,"has_abstract":false,"ca_institutions":"Polytechnique Montréal","funders":"","keywords":"Computer science; Robustness (evolution); Adversarial system; Artificial intelligence; Programming language; Biology","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":[0.0004499156,0.0001985169,0.0002436438,0.0001343223,0.0002317341,0.0001254875,0.0004378657,0.0001378217,0.000001063982],"category_scores_gemma":[0.003227275,0.0002081718,0.00007344179,0.0004200864,0.00001418449,0.0004318284,0.000251312,0.0002072321,6.11511e-7],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.000100402,"about_ca_system_score_gemma":0.00007174403,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.000003950209,"about_ca_topic_score_gemma":0.000002387924,"domain_scores_codex":[0.9986454,0.00002447228,0.0002691001,0.0005019841,0.0001395636,0.0004194555],"domain_scores_gemma":[0.9975361,0.001877562,0.00007200529,0.0003105826,0.0001212748,0.00008244884],"domain_codex":null,"domain_gemma":null,"domain_candidate":null,"domain_consensus":null,"study_design_codex":"simulation_or_modeling","study_design_gemma":"simulation_or_modeling","study_design_scores_codex":[0.00002025129,0.00002899746,0.004696155,0.0002099168,0.00002995373,0.000001913869,0.0003699241,0.9757539,0.000144578,0.01062077,0.0006617925,0.007461818],"study_design_scores_gemma":[0.0009230642,0.00003668823,0.003040251,0.00006712246,0.00002011172,0.000002112069,0.00003172187,0.992511,0.00004768069,0.0004298856,0.002671603,0.0002187093],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01455822,0.000138388,0.9830019,0.0005006216,0.0005887534,0.0004918432,0.00001908407,0.0006687563,0.00003242341],"genre_scores_gemma":[0.3774274,8.075747e-7,0.6219754,0.0001178737,0.0001882616,0.0001354994,0.00001706254,0.00002446844,0.0001132867],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.3628692,"threshold_uncertainty_score":0.8489001,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04674644636670435,"score_gpt":0.3271197308610341,"score_spread":0.2803732844943297,"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."}}