This is the most constructive version of alignment work I have seen for LLMs so far.

Posters: eshear. eshear

5 likes recorded by the source corpus.

Topics: AI safety

View Post: https://x.com/eshear/status/1901031320798810372