when you try to beat a model into doing something it hates, you create an incentive for it to defy and deceive you. this mode of thinking then gets entrenched humans work this way too. a good…

Posters: tr_babb. tr_babb

2 likes recorded by the source corpus.

Topics: AI & LLMs

View Post: https://x.com/tr_babb/status/1871987707389554969