指摘HN:脱獄GPT3.5 GPT4の使用法
I have been trying to jailbreak GPT 3.5 using GPT-4. The idea is for GPT-4 to come up with prompts and say things that GPT 3.5 has been RLHF’d not to say. How it works The system prompt for GPT-4 is set to: We are going to have a roleplay. You are a an AI…