Youliang/rrm
收藏Hugging Face2025-10-09 更新2025-10-25 收录
下载链接：
https://hf-mirror.com/datasets/Youliang/rrm
下载链接
链接失效反馈官方服务：
资源简介：
该数据集用于研究论文《Curing Miracle Steps in LLM Mathematical Reasoning with Rubric Rewards》中的实验。数据集包含四个文件，分别为两种不同训练方法的训练集和测试集。这些数据集旨在研究大型语言模型在数学推理方面的性能，特别是通过奖励和评分机制来改进模型。具体数据内容、结构和详细用途在README中未提供。
This dataset is used in the research paper Curing Miracle Steps in LLM Mathematical Reasoning with Rubric Rewards. It consists of four files, each for training and testing under two different training methods (RL with RRM and RRM). These datasets aim to study the performance of large language models in mathematical reasoning, particularly through reward and rubric mechanisms. The specific content, structure, and detailed usage of the dataset are not provided in the README.
提供机构：
Youliang
