Heretic: Automatic censorship removal for language models(github.com)
705 points by melded | 339 comments
GitHub repository for Heretic, an AGPL-licensed tool for fully automatic censorship removal (abliteration) in transformer language models.
1 story tagged censorship-removal, newest first.
705 points by melded | 339 comments
GitHub repository for Heretic, an AGPL-licensed tool for fully automatic censorship removal (abliteration) in transformer language models.