HELM

HELM (Holistic Evaluation of Language Models) is a comprehensive benchmarking framework developed by Stanford's Center for Research on Foundation Models. The platform evaluates language models across dozens of scenarios spanning question answering, summarization, information retrieval, toxicity detection, and reasoning, measuring not just accuracy but also calibration, robustness, fairness, and efficiency. HELM's standardized evaluation methodology and transparent reporting have made it a ref...

Threads