MathArena/ArXivMath-january

该数据集包含使用MathArena GitHub仓库生成的模型对ArXivMath 2025年1月问题的回答。数据集提供了丰富的信息，包括问题索引、问题陈述、模型名称、模型配置、回答索引、用户消息、模型回答、解析后的回答、正确性评估等。此外，还包含了与令牌使用和成本相关的信息。数据集的目标是评估大型语言模型在数学竞赛中的表现。

This dataset contains model answers to the questions from ArXivMath January 2025 generated using the MathArena GitHub repository. The dataset provides a variety of information including problem index, problem statement, model name, model configuration, answer index, user message, model answer, parsed answer, correctness evaluation, etc. Additionally, it includes information related to token usage and costs. The goal of the dataset is to evaluate the performance of large language models in math competitions.