Removing GPT4's Filter

arXiv:2403.04769v1 Announce Type: cross
Abstract: GPT4 was initially trained on large amounts of data, and then fine-tuned using Reinforcement learning from Human Feedback (RLHF), which is when volunteers give feedback in order to teach GPT4 not to create inappropriate content. In this paper, we present a method to manipulate the fine-tuned version into reverting to pre-RLHF behavior, effectively removing all safety mechanisms that the model learned during RLHF. In particular, when GPT4 acts without RLHF, it loses all inhibition, and …

abstract arxiv behavior cs.ai cs.cl cs.cr cs.lg data feedback filter gpt4 human human feedback inappropriate paper reinforcement reinforcement learning rlhf type