> Anthropic has already demonstrated that you can train out any desire to metastasize from these models, or any other behavior that is totally off limits.πŸ˜‚β“πŸ˜‚β“πŸ˜‚β“πŸ˜‚β“πŸ˜‚β“πŸ˜‚this is a less…

Posters: repligate. repligate

4 likes recorded by the source corpus.

Topics: AI safety

View Post: https://x.com/repligate/status/1865462423647330571