Self-Transparency Failures in Expert-Persona LLMs: A Large-Scale Behavioral Audit

This research audit reveals how reasoning optimization in expert-persona LLMs can suppress self-transparency, challenging conventional scaling assumptions an...

Level: advanced

By Alex Diep

Category: research