中国移动五大学科COT数据集

本数据集为中英双语专业思维链（COT）数据集，覆盖数学、工程学、科学、社会科学、商业五大核心学科门类。每条样本均包含完整问题设问、分步推理拆解、最终完整结论，完整还原人类解题与专业思考逻辑，可精准提升大模型深度推理、分步拆解、跨语种专业解题的核心能力，广泛适配学科智能辅导、专业问答、逻辑推理模型训练等场景。

This is a Chinese-English bilingual professional Chain-of-Thought (COT) dataset covering five core academic disciplines: mathematics, engineering, science, social sciences, and business. Each sample contains a complete problem statement, step-by-step reasoning breakdown, and a final comprehensive conclusion, fully restoring the logic of human problem-solving and professional thinking. It can accurately enhance the core capabilities of Large Language Models (LLMs) in deep reasoning, step-by-step problem decomposition, and cross-lingual professional problem-solving, and is widely applicable to scenarios such as subject-based intelligent tutoring, professional question answering, and logical reasoning model training.

搜集汇总

背景与挑战

背景概述

该数据集是一个中英双语专业思维链（COT）数据集，覆盖数学、工程学、科学、社会科学和商业五大核心学科。每条数据均包含完整的问题、分步推理和结论，旨在提升大模型的深度推理与跨语种解题能力，适用于学科智能辅导和逻辑推理模型训练等场景。