RACE What does RACE cover in an LLM output rubric? “We can judge quality by vibes in the demo.” Your line Demos find promise; evals find repeatability across normal, edge, and failure cases. A good sample can hide brittle behavior. It validates excitement while requiring launch criteria. Golden set or live feedback? Use both. What makes a rubric usable? Parallel criteria, scored examples, and a launch threshold. Otherwise scoring becomes an opinion meeting.
Read the full lesson
Sign up free: one personalized lesson every day, matched to your role and goals.
Already have an account? Sign in