Virus: Harmful Fine-tuning Attack for Large Language Models Bypassing Guardrail Moderation

Open in new window