AI safety researcher working on LLM evaluation, model incentives, revealed preferences, and covert sycophancy.https://my-website-eight-zeta-82.vercel.app/en-us