{"id":"W4406153408","doi":"10.1145/3711816","title":"LLM-Powered Static Binary Taint Analysis","year":2025,"lang":"en","type":"article","venue":"ACM Transactions on Software Engineering and Methodology","topic":"Web Application Security Vulnerabilities","field":"Computer Science","cited_by":20,"is_retracted":false,"has_abstract":true,"ca_institutions":"University of Waterloo","funders":"Ant Group","keywords":"Taint checking; Computer science; Firmware; Static analysis; Vulnerability (computing); Binary number; Scalability; State (computer science); Computer security; Software; Computer hardware; Operating system; Programming language","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":[0.0006817542,0.0001528002,0.0003115371,0.0007027992,0.0001231158,0.00004569585,0.0004669007,0.0001017122,0.00002160245],"category_scores_gemma":[0.0006839765,0.0001580269,0.0001086436,0.001246081,0.00005021356,0.0001171724,0.000027798,0.0002512641,0.000005439373],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00004626853,"about_ca_system_score_gemma":0.00004716675,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00004111612,"about_ca_topic_score_gemma":0.000006769384,"domain_scores_codex":[0.9988005,0.0002237104,0.0002397412,0.0004148266,0.00009743362,0.0002237878],"domain_scores_gemma":[0.9957763,0.003194663,0.00003382371,0.0008588262,0.00006225066,0.00007411557],"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":[0.00006107063,0.0004472359,0.0007617616,0.0003650429,0.002414963,0.0000172391,0.00538393,0.5014724,0.002679808,0.1163426,0.0002577319,0.3697962],"study_design_scores_gemma":[0.003770433,0.001341,0.05786921,0.000213092,0.002678689,0.0001042422,0.001582195,0.6942698,0.03568679,0.1144211,0.08505499,0.003008402],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"methods","genre_scores_codex":[0.01800827,0.0002046314,0.9797755,0.001106623,0.0002798432,0.000117312,0.000008608459,0.0004860692,0.00001316261],"genre_scores_gemma":[0.2352722,0.00005052113,0.764065,0.0002816413,0.000007079154,0.00008860539,0.000002823305,0.000006826221,0.0002252897],"genre_candidate":"methods","genre_consensus":"methods","teacher_disagreement_score":0.3667878,"threshold_uncertainty_score":0.6444152,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.04508240882382882,"score_gpt":0.3181407358867413,"score_spread":0.2730583270629125,"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."}}