haoranli-ml/AIME2025_Qwen3-4B-Instruct-raw-sft_rl_32768budget_16rollouts_0.8temp
收藏Hugging Face2025-11-08 更新2025-11-15 收录
下载链接：
https://hf-mirror.com/datasets/haoranli-ml/AIME2025_Qwen3-4B-Instruct-raw-sft_rl_32768budget_16rollouts_0.8temp
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含了问题、答案、分词后的提示、响应列表、奖励列表、最大奖励和平均奖励等字段。数据集被划分为AIME2025这一部分，大小为19109403字节，共有30个示例。提供了一个默认配置，数据文件位于data/AIME2025-*路径下。
The dataset includes fields such as problem, answer, tokenized prompt, a list of responses, a list of rewards, maximum reward, and average reward. The dataset is split into the AIME2025 section, which is 19109403 bytes in size and contains 30 examples. A default configuration is provided, with data files located at the path data/AIME2025-*.
提供机构：
haoranli-ml
