Refusal in Language Models Is Mediated by a Single Direction

For open-weights models, censorship removal is now a "solved" problem. If you wait a few days after a new model release, someone will have made a heretic ( https://github.com/p-e-w/heretic ) version with the censorship removed, so in a way the only use for censorship now is to avoid lawsuits, not reduce improper usage.

2024 which is ancient history. This is not true anymore, the models now are trained to prevent abliteration by spreading out the refusal encoding

See https://arxiv.org/abs/2505.19056

I have had LLMs refuse several of my requests. I still got my answers, but at least they tried.