Invisible Prompts, Visible Threats: Malicious Font Injection in External Resources for Large Language Models

Open in new window