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

 (aclanthology.org)

2 points | by aghuang 9 hours ago ago

1 comments