stanford-crfm / helm

Holistic Evaluation of Language Models (HELM), a framework to increase the transparency of language models (https://arxiv.org/abs/2211.09110). This framework is also used to evaluate text-to-image models in Holistic Evaluation of Text-to-Image Models (HEIM) (https://arxiv.org/abs/2311.04287).
https://crfm.stanford.edu/helm
Apache License 2.0
1.8k stars 239 forks source link

Release MMLU v1.3.0 with GPT-4o, Gemini 1.5 Flash, Palmyra-X V3 #2652

Closed yifanmai closed 2 months ago