mlfoundations-dev/AMC23_evalchemy_gpt-4o-mini

该数据集是通过Curator工具构建的，包含了一个数学问题的样本数据点，并通过Python代码展示了如何加载数据集。数据集与AMC23任务相关的解决问题数据。评估结果显示，平均准确度为54.00%±2.19%，五次运行中的准确度从45.00%到60.00%不等。

This dataset was constructed by the Curator tool, containing a sample data point of a mathematical problem, and demonstrating how to load the dataset using Python code. The dataset is related to the AMC23 task and contains problem-solving data. The evaluation results show that the average accuracy is 54.00% ± 2.19%, and the accuracy rate varies from 45.00% to 60.00% over five runs.