RUC-AIBOX/STILL-3-Preview-RL-Data
收藏Hugging Face2025-01-26 更新2025-04-12 收录
下载链接：
https://hf-mirror.com/datasets/RUC-AIBOX/STILL-3-Preview-RL-Data
下载链接
链接失效反馈官方服务：
资源简介：
STILL-3-预览-RL数据集包含约30K个用于强化学习训练的数学问答对。该数据集基于MATH、NuminaMathCoT和AIME 1983-2023数据源进行策划。每个数据条目是一个包含问题、使用聊天模板格式化的输入消息和问题答案的字典。
STILL-3-Preview-RL-Data consists of ~30K math question-answer pairs for RL training. The dataset is curated based on MATH, NuminaMathCoT, and AIME 1983-2023 as the data source. Each entry is a dictionary containing the question, formatted input using a chat template, and the answer to the question.
提供机构：
RUC-AIBOX
