lighteval/MATH-Hard
收藏资源简介:
Mathematics Aptitude Test of Heuristics (MATH)数据集包含来自数学竞赛的问题,如AMC 10、AMC 12、AIME等。每个问题都有详细的步骤解答,解答使用LaTeX和自然语言编写,可用于教授模型生成答案推导和解释。MATH-Hard子集仅保留了最困难的问题(Level 5)。数据集分为不同的配置,如代数、几何、数论等,每个配置包含训练和测试数据。
The Mathematics Aptitude Test of Heuristics (MATH) dataset consists of problems from mathematics competitions, including the AMC 10, AMC 12, AIME, and more. Each problem in MATH has a full step-by-step solution, which can be used to teach models to generate answer derivations and explanations. For MATH-Hard, only the hardest questions were kept (Level 5). The dataset includes fields such as problem, solution, level, and type.
数据集概述
数据集名称
Mathematics Aptitude Test of Heuristics (MATH)
数据集描述
MATH数据集包含来自数学竞赛的问题,包括AMC 10、AMC 12、AIME等。每个问题都附有详细的逐步解决方案,可用于训练模型生成答案推导和解释。MATH-Hard子集仅保留了最难的问题(Level 5)。
支持的任务
- 文本生成
语言
- 英语
数据集结构
数据实例
每个数据实例包含一个竞赛数学问题及其逐步解决方案,解决方案以LaTeX和自然语言编写,并包含最终答案。
数据字段
problem: 竞赛数学问题。solution: 逐步解决方案。level: 仅保留Level 5的问题。type: 问题的主题,包括代数、计数与概率、几何、中级代数、数论、初等代数和微积分预备。
配置
default: 包含所有问题的训练和测试数据。algebra: 代数问题的训练和测试数据。counting_and_probability: 计数与概率问题的训练和测试数据。geometry: 几何问题的训练和测试数据。intermediate_algebra: 中级代数问题的训练和测试数据。number_theory: 数论问题的训练和测试数据。prealgebra: 初等代数问题的训练和测试数据。precalculus: 微积分预备问题的训练和测试数据。
许可证
MIT许可证
引用信息
bibtex @article{hendrycksmath2021, title={Measuring Mathematical Problem Solving With the MATH Dataset}, author={Dan Hendrycks and Collin Burns and Saurav Kadavath and Akul Arora and Steven Basart and Eric Tang and Dawn Song and Jacob Steinhardt}, journal={arXiv preprint arXiv:2103.03874}, year={2021} }




