Defending Large Language Models against Jailbreak Attacks via Semantic Smoothing

Open in new window