Enhancing LLM Agent Safety via Causal Influence Prompting

Open in new window