Fetching the paper…

Chain-of-Scrutiny: Detecting Backdoor Attacks for Large Language Models · Around