bethgelab/CuratedThoughts
收藏Hugging Face2025-02-26 更新2025-04-12 收录
下载链接：
https://hf-mirror.com/datasets/bethgelab/CuratedThoughts
下载链接
链接失效反馈官方服务：
资源简介：
这个数据集是用于强化学习训练的数据集，包含了数学问题和相应的解答。数据集分为几个不同的配置，每个配置都包含了问题、解决方案、答案等不同的特性。数据集旨在通过强化学习的方法来生成正确的推理轨迹，但在实际使用中存在一些问题，例如问题包含子问题、需要证明的问题、引用图片的问题以及没有 ground truth 的情况。
This dataset is for reinforcement learning training, containing math problems and corresponding solutions. The dataset is divided into several configurations, each with different features such as problems, solutions, answers, etc. The dataset aims to generate correct reasoning traces through reinforcement learning methods, but there are some issues in practical use, such as problems containing sub-questions, problems requiring proof, problems referring to figures, and cases without ground truth.
提供机构：
bethgelab
