The Dangers of Indirect Prompt Injection Attacks on LLM-based Autonomous Web Navigation Agents: A Demonstration.

Sam Johnson, Viet Pham, Thai Le

Browse the full EMNLP paper archive.