K12习题数据集

本数据集面向教育科技企业、智能教辅平台、大模型研发团队及基础教育教学研究机构，旨在满足K12全阶段智能学习、自动解题及能力评测场景的需求，解决传统题库数据格式不统一、内容分散且难以直接用于模型训练的问题。数据集采用统一的JSON结构，为每道习题提供题干、选项及完整解题过程，覆盖小学数学、初中数学、初中物理、初中化学、高中数学、高中生物等关键学科。与传统散装习题集不同，本数据集将正确答案背后的解题步骤以结构化字段形式呈现，而非仅提供最终答案，使模型能够学习从读题到结论的完整推演路径。

This dataset targets educational technology enterprises, intelligent teaching assistance platforms, large language model (LLM) R&D teams, and basic education teaching research institutions. It aims to meet the requirements of intelligent learning, automatic problem-solving, and ability evaluation scenarios across all K12 stages, and addresses the issues of inconsistent data formats, scattered content, and difficulty in directly applying traditional question banks for model training. The dataset adopts a unified JSON structure, providing each exercise with its question stem, options, and complete problem-solving process, covering key subjects including primary school mathematics, junior high school mathematics, junior high school physics, junior high school chemistry, senior high school mathematics, senior high school biology, and other key disciplines. Unlike traditional loose exercise collections, this dataset presents the problem-solving steps behind the correct answer in the form of structured fields, rather than only providing the final answer, enabling the model to learn the complete deduction path from reading the question to reaching the final conclusion.

搜集汇总

背景与挑战

背景概述

本数据集面向教育科技企业、智能教辅平台、大模型研发团队及基础教育教学研究机构，旨在满足K12全阶段智能学习、自动解题及能力评测场景的需求。它采用统一的JSON结构提供覆盖数学、物理、化学、生物等关键学科的习题，并包含结构化解题步骤，以支持模型学习完整的推演路径。