Fairness

HELM (Holistic Evaluation of Language Models)

A standardized framework evaluating language models across many scenarios and multiple metrics including accuracy, robustness, fairness, and efficiency.