Revealing a Hidden Class of Task-in-Prompt Adversarial Attacks on LLMs aclanthology.org 2 points by aghuang a month ago · 1 comment Reader PiP Save No comments yet.