inclusionAI/AReaL-RL-Data
收藏Hugging Face2025-03-28 更新2025-04-08 收录
下载链接：
https://hf-mirror.com/datasets/inclusionAI/AReaL-RL-Data
下载链接
链接失效反馈官方服务：
资源简介：
AReaL数据集是一个包含40k个高质量数学推理任务的数据集，适用于强化学习模型的训练，特别是用于训练大型推理模型。数据集包含了每个问题的解决方案索引，以及用于训练DeepSeek-R1-Distill-Qwen-1.5B模型和R1-Zero风格模型的提示数据集。
The AReaL dataset consists of 40k high-quality mathematical reasoning tasks, suitable for reinforcement learning model training, especially for training large reasoning models. The dataset includes solutions indexed by question ID and prompt datasets for training the DeepSeek-R1-Distill-Qwen-1.5B model and the R1-Zero style model.
提供机构：
inclusionAI
