mlfoundations-dev/OpenThinker-7B_eval_03-09-25_03-36_f912
收藏Hugging Face2025-03-09 更新2025-04-12 收录
下载链接：
https://hf-mirror.com/datasets/mlfoundations-dev/OpenThinker-7B_eval_03-09-25_03-36_f912
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含了预计算的模型输出，用于评估。具体评估结果在GPQADiamond部分给出，显示了平均准确率为28.11%，共进行了3次运行。每次运行的准确率分别为22.73%、24.75%和36.87%，解决的问题数量分别为45、49和73，每次运行总问题数量为198。
This dataset contains precomputed model outputs for evaluation. The specific evaluation results are provided in the GPQADiamond section, showing an average accuracy of 28.11%, with a total of 3 runs. The accuracy of each run is 22.73%, 24.75%, and 36.87% respectively, with 45, 49, and 73 questions solved, and a total of 198 questions in each run.
提供机构：
mlfoundations-dev
