Fetching the paper…

Benchmarking and Defending Against Indirect Prompt Injection Attacks on Large Language Models · Around