mlfoundations-dev/global_batchsize_512_lradjusted8_1743609353_eval_0981
收藏Hugging Face2025-04-02 更新2025-04-12 收录
下载链接：
https://hf-mirror.com/datasets/mlfoundations-dev/global_batchsize_512_lradjusted8_1743609353_eval_0981
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含了对预计算模型在不同数学和编程任务（如AIME24、AIME25、AMC23、MATH500、GPQADiamond和LiveCodeBench）上的评估结果。每个任务都有多个运行次数的准确度记录，以及解决的问题数和总问题数。
The dataset consists of evaluation results of precomputed model outputs on various mathematical and programming tasks (such as AIME24, AIME25, AMC23, MATH500, GPQADiamond, and LiveCodeBench). Each task has records of accuracy for multiple runs, along with the number of questions solved and the total number of questions.
提供机构：
mlfoundations-dev
