We show user expectations influence model self-conception. We had one LLM primed with text about AI identity (such as "Stochastic Parrots" or "Simulators") chat to another LLM about physics. That…

Posters: jankulveit. jankulveit

2 likes recorded by the source corpus.

Topics: AI & LLMs, AI safety

View Post: https://x.com/jankulveit/status/2032520497511166356