MathVista
收藏arXiv2025-09-30 收录
下载链接：
https://mathvista.github.io/
下载链接
链接失效反馈官方服务：
资源简介：
该数据集旨在成为一个评估模型在数学推理能力方面的基准，它专门设计用来检测模型在数学推理任务中的表现。因此，该数据集的任务专注于数学推理。
This dataset is intended as a benchmark for evaluating models' mathematical reasoning abilities, and it is specifically designed to assess models' performance on mathematical reasoning tasks. Accordingly, the tasks included in this dataset focus on mathematical reasoning.
