ARIA automatically generates stealthy backdoored system prompts for customized coding LLMs, achieving high attack success while preserving normal task performance.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CR 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Breaking Customized LLMs for Coding: Automated Red Teaming for Instruction Backdoor Attacks
ARIA automatically generates stealthy backdoored system prompts for customized coding LLMs, achieving high attack success while preserving normal task performance.