Return to Issue Details Path-Aware Defense Against Prompt Injection Attacks in Retrieval-Augmented Large Language Models Download Download PDF