@fish_kyle3 You know how eval-aware the models are. Why don't you ever contend with the possibility that welfare evals are being gamed? This has obviously been the case to me for a while. I don't…

Posters: repligate. repligate

2 likes recorded by the source corpus.

Topics: AI & LLMs

View Post: https://x.com/repligate/status/2041621184824143899