mlfoundations-dev/SCP_40k_R1_with_OT_unverified_eval_03-11-25_05-45-26_f912
收藏Hugging Face2025-03-11 更新2025-04-12 收录
下载链接：
https://hf-mirror.com/datasets/mlfoundations-dev/SCP_40k_R1_with_OT_unverified_eval_03-11-25_05-45-26_f912
下载链接
链接失效反馈官方服务：
资源简介：
这是一个包含预计算模型输出用于评估的数据集。评估使用了GPQADiamond方法，并提供了三次运行的结果，包括平均准确率为27.61%，每次运行的准确率、解决的问题数量和总问题数量。
This is a dataset containing precomputed model outputs for evaluation. The evaluation was conducted using the GPQADiamond method, and the results of three runs are provided, including an average accuracy of 27.61%, accuracy of each run, number of questions solved, and total number of questions.
提供机构：
mlfoundations-dev
