{"id":"W4408146479","doi":"10.1109/icairc64177.2024.10900215","title":"Robustness of Large Language Models Against Adversarial Attacks","year":2024,"lang":"en","type":"article","venue":"","topic":"Adversarial Robustness in Machine Learning","field":"Computer Science","cited_by":12,"is_retracted":false,"has_abstract":true,"ca_institutions":"Simon Fraser University","funders":"","keywords":"Robustness (evolution); Computer science; Adversarial system; Language model; Computer security; Artificial intelligence","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":[0.0005470538,0.0001849622,0.000240953,0.0001992573,0.00008211839,0.0001380791,0.001004569,0.0001104231,0.0001011343],"category_scores_gemma":[0.00006531976,0.0001608696,0.0001387127,0.000620375,0.00004296195,0.001119029,0.0006143056,0.0003236363,0.00003259996],"about_ca_system_candidate":false,"about_ca_system_consensus":false,"about_ca_system_score_codex":0.00005069875,"about_ca_system_score_gemma":0.0001268762,"about_ca_topic_candidate":false,"about_ca_topic_consensus":false,"about_ca_topic_score_codex":0.00003608253,"about_ca_topic_score_gemma":0.00001237568,"domain_scores_codex":[0.9983377,0.0000964111,0.000307704,0.0004833749,0.0004049847,0.0003697982],"domain_scores_gemma":[0.9990084,0.0001790239,0.00006196916,0.00060608,0.00006164591,0.00008284509],"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":[0.000006156846,0.00003349161,0.00001892363,0.00006605582,0.00003722734,0.00006347283,0.001508808,0.7617255,0.0003461318,0.2224173,0.001188371,0.01258856],"study_design_scores_gemma":[0.0003194123,0.00002200426,0.00001215864,0.0000607221,0.00001006725,0.000006413987,0.0001903421,0.9969343,0.0005019505,0.0005964563,0.001153632,0.0001925562],"study_design_candidate":"simulation_or_modeling","study_design_consensus":"simulation_or_modeling","genre_codex":"methods","genre_gemma":"empirical","genre_scores_codex":[0.01026675,0.0003166747,0.9639716,0.0003909945,0.001734736,0.0001272741,0.000006173489,0.0006422849,0.0225435],"genre_scores_gemma":[0.911322,0.00001021786,0.086862,0.0001224066,0.0003248682,0.000006048812,0.000008472573,0.00002387006,0.001320123],"genre_candidate":"methods","genre_consensus":null,"teacher_disagreement_score":0.9010553,"threshold_uncertainty_score":0.6560072,"prediction_status":"machine_predicted_unvalidated"},"machine_scores":{"provisional":true,"baseline":true,"maturity_gate_passed":false,"score_opus":0.01302120973298575,"score_gpt":0.2868284806767559,"score_spread":0.2738072709437701,"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."}}