Jigsaw Puzzles: Splitting Harmful Questions to Jailbreak Large Language Models

Open in new window