mlfoundations-dev/DCFT-hero_run_2_fix_conversations-etash_1743569538_eval_f912
收藏Hugging Face2025-04-02 更新2025-04-12 收录
下载链接：
https://hf-mirror.com/datasets/mlfoundations-dev/DCFT-hero_run_2_fix_conversations-etash_1743569538_eval_f912
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含了使用GPQADiamond模型进行评估的预计算模型输出结果。评估结果显示了三次运行的准确率分别为25.25%、20.71%和19.19%，平均准确率为21.72%。每次运行都针对198个问题，解决的问题数量分别为50、41和38。
This dataset includes precomputed model outputs for evaluation using the GPQADiamond model. The evaluation results show accuracy rates of 25.25%, 20.71%, and 19.19% for three runs, with an average accuracy of 21.72%. Each run was conducted on 198 questions, with 50, 41, and 38 questions solved respectively.
提供机构：
mlfoundations-dev
