yeah i agree with that. nostalgebraist had a good post on this on LW, that (pp) the reward-seeking "persona(s)" diverge from the regular inference one(s). my argument is just that this "reward…

Posters: voooooogel. voooooogel

3 likes recorded by the source corpus.

Topics: AI & LLMs

View Post: https://x.com/voooooogel/status/2086231651097653437