EpochLex entry

Human Evaluation

HYOO-mun ih-val-yoo-AY-shun

Evaluation in which people directly judge, rate, compare, or otherwise assess AI outputs or behavior using defined criteria rather than relying only on automated metrics or model-based graders.