GeneralReasoning/GeneralThought-195K

GeneralThought-195K数据集是一个开放推理数据集，来源于General Reasoning资源，包含问题、参考答案、推理痕迹、最终答案以及其他元数据。数据集包含了多个流行推理模型生成的数据，如DeepSeek-R1、DeepSeek-R1-Zero、OpenThoughts-32B等，以及用于比较和评估的o3-mini-2025-01-31、gemini-2-flash-thinking-exp-01-21和claude-3-7-sonnet-20250219模型的最终答案。数据集的最新版本包含了195,054行数据，推理痕迹的多样性得到了增强，覆盖了数学、自然科学、人文社会科学等多个领域。

The GeneralThought-195K dataset is an open reasoning dataset from the General Reasoning resource, containing questions, reference answers, reasoning traces, final answers, and other metadata. The dataset includes data generated by multiple popular reasoning models such as DeepSeek-R1, DeepSeek-R1-Zero, OpenThoughts-32B, etc., as well as final answers from models like o3-mini-2025-01-31, gemini-2-flash-thinking-exp-01-21, and claude-3-7-sonnet-20250219 for comparison and evaluation. The latest version of the dataset contains 195,054 rows of data, with enhanced diversity in reasoning traces covering various fields including mathematics, natural sciences, humanities, and social sciences.