中华传统文化知识图谱实体本体数据

本数据集旨在为中华优秀传统文化领域大语言模型及智能应用提供高质量、结构化、可追溯的实体知识基座。具体应用包括：1. 大模型领域知识注入与精调：作为高质量SFT数据，用于对齐和优化传统文化垂域LLM的参数，使其在生成关于典籍、人物、器物等实体的解释时，能严格基于权威出处输出结构化知识，有效抑制事实性幻觉。2. 文化智能应用开发：可作为文化教育App、数字文博导览、文创内容生成等应用的底层知识库，支撑实体查询、关系探索、智能问答等功能。This dataset aims to provide a high-quality, structured, traceable entity knowledge base for large language models and intelligent applications in the field of outstanding traditional Chinese culture. Its specific applications include: 1. Knowledge injection and fine-tuning for large models: As high-quality Supervised Fine-Tuning (SFT) data, it is used to align and optimize the parameters of traditional Chinese culture vertical-domain LLMs, enabling them to output structured knowledge strictly based on authoritative sources when generating explanations for entities such as classics, figures and artifacts, and effectively suppressing factual hallucinations. 2. Development of cultural intelligent applications: It can serve as the underlying knowledge base for applications such as cultural education Apps, digital cultural heritage tours and cultural and creative content generation, supporting functions including entity query, relationship exploration, intelligent question answering and other related functions.

搜集汇总

背景与挑战

背景概述

该数据集是一个高质量的中华传统文化实体知识库，旨在为相关大语言模型和智能应用提供结构化、可追溯的知识基座。它通过混合流水线处理公开古籍资料，产出了超过10万条实体记录，涵盖典籍、人物、器物等十大类别，每条记录均包含标准名称、解释、原句和文献出处等字段，确保学术严谨性。该数据集可用于大模型精调以抑制事实性幻觉，并支持文化教育、数字文博等应用的开发。