13:40
2026-07-30
arxiv.org
artificial-intelligence
Adversarial Code Obfuscation for Defending Against LLM-Based Analysis
Researchers have developed Acoda, a genetic algorithm-based adversarial code obfuscation framework that defends against LLM-based code analysis by inducing LLMs to refuse or misinterpret code. On seve…