Fetching the paper…

Defending Large Language Models against Jailbreak Attacks via Semantic Smoothing · Around