llm-compe-2025-kato/step2-DeepMath-103K-Bespoke-Filtered-Test-200-eval-cp32
收藏Hugging Face2025-08-20 更新2025-09-13 收录
下载链接：
https://hf-mirror.com/datasets/llm-compe-2025-kato/step2-DeepMath-103K-Bespoke-Filtered-Test-200-eval-cp32
下载链接
链接失效反馈官方服务：
资源简介：
这是一个包含问题、答案以及对应解释生成（Chain of Thought）的数据集，用于训练和评估模型在生成解答过程中的推理能力。数据集由60个处理过的样本组成，每个样本都包含系统提示、问题文本、答案文本和链式推理。数据集遵循MIT许可证，并专注于文本生成任务。
This dataset consists of questions, answers, and corresponding explanation generations (Chain of Thought) for training and evaluating models on their reasoning ability in the process of generating answers. The dataset comprises 60 processed samples, each including a system prompt, question text, answer text, and chain of thought. The dataset is licensed under MIT and focuses on the text generation task category.
提供机构：
llm-compe-2025-kato
