库帕思高质量消防手册监督微调（Supervised Fine-Tuning）数据集

本数据集面向消防领域大模型及智能问答系统研发团队，旨在解决通用大模型在消防实务操作指导中存在的流程描述不准确、操作要点缺失等问题。以《中国消防手册》为基础，构建了经过专业校验的高质量问答对，覆盖火灾预防、灭火救援、装备使用、应急管理等核心章节。

通过监督微调，可显著提升模型对消防实务知识的掌握程度，增强其在操作流程说明、应急处置指导、装备使用方法等任务中的专业性与准确性。与传统通用语料不同，本数据集所有问答对均源自权威手册并经人工核验，确保答案的现实可操作性。

This dataset is developed for R&D teams of large language models (LLMs) and intelligent question answering systems in the fire protection domain, aiming to address issues such as inaccurate process descriptions and missing critical operation points in general LLMs when providing fire practice operation guidance.
Built upon the *Chinese Fire Protection Handbook*, this dataset features high-quality question-answer pairs that have undergone professional validation, covering core chapters including fire prevention, fire suppression and rescue, equipment operation, emergency management and other key areas.
Through supervised fine-tuning, the model's proficiency in fire practice-related knowledge can be significantly improved, and its professionalism and accuracy in tasks such as explaining operation procedures, offering emergency response guidance, and demonstrating equipment usage methods can be enhanced. Unlike conventional general-purpose corpora, all question-answer pairs in this dataset are sourced from authoritative handbooks and manually verified, ensuring the practical operability of the provided answers.

搜集汇总

背景与挑战

背景概述

该数据集基于《中国消防手册》构建，包含经专业校验的高质量问答对，覆盖火灾预防、灭火救援等核心消防领域。它旨在通过监督微调提升大模型在消防实务操作指导中的专业性和准确性，适用于消防培训、现场作业等场景。