knoveleng/open-s1
收藏Hugging Face2026-03-18 更新2025-04-12 收录
下载链接:
https://hf-mirror.com/datasets/knoveleng/open-s1
下载链接
链接失效反馈官方服务:
资源简介:
open-s1数据集是从s1K数据集中筛选出的18615个数学推理问题,它是Open RS项目的一部分,旨在通过强化学习增强小型语言模型的推理能力。数据集包含问题、解决方案、答案、来源以及用户和助手之间的交互信息。
The open-s1 dataset consists of 18,615 mathematical reasoning problems filtered from the s1K dataset. It is part of the Open RS project aimed at enhancing reasoning in small LLMs using reinforcement learning. The dataset includes problems, solutions, answers, sources, and interactions between users and assistants.
提供机构:
knoveleng



