大语言模型辅助国际中文教育文化教学研究——以ChatGPT和文心一言为例
Research on Large Language Models Assisting in Cultural Teaching of International Chinese Language Education:A Case Study of ChatGPT and ERNIE Bot
蔡汶池 1马世 1郑晓兰 1毛海莹1
作者信息
- 1. 宁波大学人文与传媒学院,中国宁波 315211
- 折叠
摘要
大语言生成模型的出现给国际中文教育数字化转型带来新的机遇,但由于其本身不具有认知动机和鉴别数据源真伪的功能,使其在生成教学文本时缺少一定的准确性,在文本难度上也与预期猜想不符.本文基于ChatGPT和文心一言,依托汉语二语文本难度自动分级平台和spss对其生成的文化点解释进行分等级的阅读难度和准确性测试,发现:1.二者所生成的本文阅读难度都偏高;2.文心一言所生成的文本内容难度要高于ChatGPT;3.文心一言对于文化点的解释准确度高于ChatGPT.ChatGPT为代表的大语言模型已经初步具备了辅助国际中文教育文化教学的潜力,但由于其内容存在准确性问题和伦理风险,教师在使用过程中应不断提升自身数字素养,充分利用大语言辅助模型优势,警惕其风险和问题,促进国际中文智能教育教学.
Abstract
The rise of large language models offers new opportunities for the digital transformation of In-ternational Chinese Language Education.However,because these models lack the ability to understand and verify data sources,the texts they generate can be inaccurate and unexpectedly difficult.This study,using ChatGPT and ERNIE Bot,employs an automatic grading platform for the difficulty of Chinese as a second language texts and SPSS to test the reading difficulty and accuracy of their cultural explanations.The findings are:1.Both models generate texts that are generally too difficult;2.ERNIE Bot's texts are harder than those of ChatGPT;3.ERNIE Bot provides more accurate cultural explanations than ChatGPT.Large language models like ChatGPT have shown potential to help in cultural teaching within International Chinese Language Education.However,because of issues with accuracy and ethical risks,teachers should improve their digital skills,make the best use of these models,be aware of their risks,and pro-mote intelligent education and teaching in International Chinese Language Education.
关键词
文化教学/ChatGPT/文心一言/数字化转型Key words
cultural teaching/ChatGPT/ERNIE Bot/digital transformation引用本文复制引用
出版年
2023
NETL
NSTL
万方数据