{"id":"W6979343015","doi":"","title":"Model Compression vs. Adversarial Robustness: An Empirical Study on Language Models for Code","year":2025,"lang":"en","type":"article","venue":"ArXiv.org","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":0,"is_retracted":false,"has_abstract":true,"ca_institutions":"","funders":"Natural Sciences and Engineering Research Council of Canada","keywords":"Adversarial system; Software; Source lines of code; Robustness (evolution); Source code; Software deployment; Uncompressed video; Empirical research; Language model","routes":{"ca_aff":false,"ca_fund":true,"ca_venue":false,"about_ca":false,"invisible_to_affiliation_only":true},"retraction":null,"screen":null,"direct_labels":[],"prediction":{"model_version":"codex-gemma-dda1882f352a","candidate_categories":["metaepi_narrow"],"consensus_categories":[],"category_scores_codex":[0.0006213009,0.0003373782,0.0004318766,0.0002331681,0.0004700547,0.000129014,0.00168728,0.0001769822,0.000006930594],"category_scores_gemma":[0.0002142914,0.0003068019,0.0001320834,0.000449286,0.00005999876,0.0009408586,0.0007031436,0.0005041881,0.00001383261],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.0001255132,"about_ca_system_score_gemma":0.0001804793,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00004706655,"about_ca_topic_score_gemma":0.00003277789,"domain_scores_codex":[0.9973449,0.0002870479,0.0004231521,0.001014699,0.0004260572,0.0005042142],"domain_scores_gemma":[0.9979721,0.0003511529,0.0001504088,0.001248619,0.0001334903,0.0001441981],"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":[0.000242192,0.0005809766,0.0303563,0.0000165592,0.0000393161,0.00001643932,0.002844154,0.9605941,0.0002270187,0.003019693,0.0006705719,0.001392685],"study_design_scores_gemma":[0.002395541,0.0003646781,0.009082255,0.00005100418,0.00003934444,9.227939e-7,0.0005457071,0.9857715,0.0001886917,0.001150294,0.00009907367,0.0003109345],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.441755,0.00001325631,0.5559258,0.0005588561,0.0004837618,0.0005010904,0.00000476739,0.0002762429,0.0004811824],"genre_scores_gemma":[0.9647692,0.000001245702,0.03331986,0.0009114182,0.0002043063,0.0001009779,0.00001638482,0.00003276694,0.0006438279],"genre_candidate":"empirical","genre_consensus":null,"teacher_disagreement_score":0.5230142,"threshold_uncertainty_score":0.9999384,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.07452659093562705,"score_gpt":0.3712633562769646,"score_spread":0.2967367653413375,"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."}}