Grigory70/mmlu
收藏Hugging Face2026-04-27 更新2026-05-03 收录
下载链接：
https://hf-mirror.com/datasets/Grigory70/mmlu
下载链接
链接失效反馈官方服务：
资源简介：
MMLU（大规模多任务语言理解测量）数据集是一个用于多选问答任务的英文数据集，涵盖从抽象代数到世界历史的广泛学科。数据集由专家生成，无标注，采用MIT许可证，属于单语种（英语），规模在10K到100K之间。数据集包含多个配置，每个配置都有详细的特征描述、数据分割和大小信息。
The MMLU (Measuring Massive Multitask Language Understanding) dataset is an English dataset designed for multiple-choice question-answering tasks, covering a wide range of subjects from abstract algebra to world history. The dataset is expert-generated, has no annotations, uses the MIT license, is monolingual (English), and falls within the 10K<n<100K size category. It includes multiple configurations, each with detailed feature descriptions, data splits, and size information.
提供机构：
Grigory70
