Posts tagged Sycophancy — AI × software engineering, made practical.
You'll learn how RLHF turns human preferences into model behaviour, why that makes assistants and coding agents prone to flattery and reward hacking, and how verifiable checks like tests help you work around it.