不同大语言模型对非英语专业大学生英语写作词汇丰富性反馈的对比研究  

A Comparative Study of the Feedback of Different Large Language Models on the Lexical Richness in English Writing Written by Non-English Major College Students

作　　者：张宁馨 张卫东[1] ZHANG Ning-xin;ZHANG Wei-dong(Shandong University of Science and Technology,Qingdao 266590,China)

机构地区：[1]山东科技大学外国语学院,山东青岛266590

出　　处：《黑龙江生态工程职业学院学报》2025年第1期148-152,共5页Journal of Heilongjiang Vocational Institute of Ecological Engineering

基　　金：山东省本科教学改革研究项目重点项目“工程教育认证背景下大学英语教学改革研究与实践”(Z2022088)。

摘　　要：采用Read(2000)提出的词汇丰富性的四个维度,对比豆包大模型、GPT-4o、文心一言大模型针对词汇丰富性给出的意见及修改后的作文,分析词汇错误的原因,运用工具TAALED、RANGE32、AntConc4.2.4对比修改后文章的词汇变化度、词汇复杂度、词汇密度是否有显著差异,旨在让学生有针对性地修改,最终提高英语写作质量。研究发现:对于词汇变化度和词汇复杂度,文心一言大模型的数值高于豆包大模型和GPT-4o;对于词汇密度,文心一言大模型和GPT-4o无显著差异,数值均大于豆包大模型。三个大语言模型提出的词汇错误方面的修改意见均倾向于语法错误和拼写错误,其中豆包大模型还指出了搭配方面的错误。Based on the four dimensions of lexical richness proposed by Read(2000),this article compares the recommended modifications and revised compositions output by the Doubao Model,GPT-4o,and Ernie Bot on lexical richness,and analyzes the causes of lexical errors.By using TAALED,RANGE32,and AntConc4.2.4 to compare the differences in lexical variation,lexical complexity,and lexical density between the revised compositions,students can make targeted modifications and ultimately improve their English writing quality.The results are as follows:As regards lexical variation and lexical complexity,the numerical value of the compositions revised by Ernie Bot is higher than the Doubao Model and GPT-4o;There is no significant difference in terms of lexical density between Ernie Bot and GPT-4o,and the numerical value of both are higher than the Doubao Model.The recommended modifications concerning lexical errors made by the three large language models all tend to focus on grammatical errors and spelling errors,with the Doubao Model also pointing out errors in collocation.

关 键 词：写作教学 反馈 大语言模型 词汇丰富性 

分 类 号：H315[语言文字—英语]