Evaluation (Evals)

say it: “EE-valz” /ˈivælz/ or “ih-val-yoo-AY-shun” /ɪˌvæljuˈeɪʃən/

Test sets that score AI output quality.

What is Evaluation (Evals)?

Evals are test sets for AI features: sample inputs with expected qualities, scored to see whether a prompt or model change made things better or worse.

Level: Journeyman · More AI Basics terms at this level

Updated