Efficient Adversarial Training in LLMs with Continuous Attacks

Xhonneux, Sophie, Sordoni, Alessandro, Günnemann, Stephan · arXiv · 2024