【论文笔记】Large Language Models for Code：Security Hardening and Adversarial Testing