Large Language Models are Vulnerable to Bait-and-Switch Attacks for Generating Harmful Content

Open in new window