Revealing a Hidden Class of Task-in-Prompt Adversarial Attacks on LLMs

• Posted 1 hour ago by aghuang
• 1 points