GPT-4大语言模型对护理知识理解的测试研究  

Testing research of GPT-4 large language model on nursing knowledge comprehension

作　　者：徐文博 陈凤敏 王超 陈洁[3] 侯辉[3] Xu Wenbo;Chen Fengmin;Wang Chao;Chen Jie;Hou Hui(Department of General surgery,The First Affiliated Hospital of Jinzhou Medical University,Jinzhou 121000,China;不详)

机构地区：[1]锦州医科大学附属第一医院普外科,辽宁锦州121000 [2]辽宁工业大学图书馆 [3]桂林理工大学图书馆

出　　处：《护理学杂志》2024年第19期93-96,共4页Journal of Nursing Science

基　　金：2023年度教育部人文社会科学研究规划基金(23YJA870002)。

摘　　要：目的探讨GPT-4大语言模型在护理教育中的应用潜力。方法选用GPT-4对主管护师考试真题进行量化测试,并对答案准确率进行分类评价。结果GPT-4的整体准确率为81.00%。在知识记忆和简单选项题目上准确率较高,分别为82.64%和82.52%;在解答知识应用和复杂选项题目时,GPT-4的准确率较低,分别为76.60%和70.97%。结论GPT-4展现出作为护理教学和临床护理辅助工具的巨大潜力。未来研究应探索如何将大语言模型与外部知识源结合并创新应用方法,提升大模型生成内容的准确性。同时,护理教育工作者还应积极探索大模型提升学生自学能力和自我评价能力的方法。Objective To explore the application potential of GPT-4 large language model in nursing education.Methods GPT-4 was used to quantitatively test the Supervisory Nurse Examination,and the accuracy of the answers was classified and evaluated.Results The overall accuracy of GPT-4 was 81.00%.The accuracy of knowledge memorization and simple choice questions was 82.64%and 82.52%,respectively.The accuracy of GPT-4 was 76.60%and 70.97%respectively when solving knowledge application and complex choice questions.Conclusion GPT-4 shows great potential as an auxiliary tool for nursing teaching and clinical nursing.Future research should explore how to combine large language models with external knowledge sources and innovate application methods to improve the accuracy of content generated by large models.At the same time,nursing educators should actively explore ways to improve students′self-study ability and self-evaluation ability with large language models.

关 键 词：大语言模型 人工智能 GPT-4 护理教育 护理教学 护理知识 试题 测试 

分 类 号：G642.0[文化科学—高等教育学] TP181[文化科学—教育学]