mlfoundations-dev/OpenThinker-7B_1743026501_eval_71e5
收藏官方服务：
资源简介：
该数据集是OpenThinker-7B模型在AIME25和MATH500两个评估任务上的预计算输出结果。AIME25任务的平均准确度为23.33%，MATH500任务的准确度为84.00%。
This dataset contains precomputed model outputs for OpenThinker-7B evaluation on AIME25 and MATH500 tasks. The average accuracy on AIME25 is 23.33%, and the accuracy on MATH500 is 84.00%.
提供机构：
mlfoundations-dev
