Fetching the paper…

Evaluating the Instruction-Following Robustness of Large Language Models to Prompt Injection · Around