marin-community/datashop-science-qa
收藏Hugging Face2025-05-16 更新2025-07-05 收录
下载链接：
https://hf-mirror.com/datasets/marin-community/datashop-science-qa
下载链接
链接失效反馈官方服务：
资源简介：
这是一个聚焦于科学知识的问答数据集，通过模型筛选DCLM Baseline数据集构建而成，包含约12.6B Llama-3标记的问答对。数据集能够显著提高MMLU STEM子集的分数。
This is a science-focused QA dataset constructed by filtering the DCLM Baseline dataset using a model, containing approximately 12.6B Llama-3 token QA pairs. The dataset significantly improves the MMLU STEM subset scores.
提供机构：
marin-community
