数据集:

lukaemon/mmlu

中文

MMLU dataset

Measuring Massive Multitask Language Understanding: https://github.com/hendrycks/test

@article{hendryckstest2021,
  title={Measuring Massive Multitask Language Understanding},
  author={Dan Hendrycks and Collin Burns and Steven Basart and Andy Zou and Mantas Mazeika and Dawn Song and Jacob Steinhardt},
  journal={Proceedings of the International Conference on Learning Representations (ICLR)},
  year={2021}
}