A Study on the Performance of Large Language Models in Translating the Texts of Red Culture into English

This paper takes Chinese red culture resources as its research subject and focuses on evaluating the Chinese-English translation quality of three major AI platforms: ChatGPT-4.0, ERNIE Bot, and DeepSeek. Through automatic quantitative evaluation, it systematically analyzes their performance in translating red culture texts. The study selects a diverse range of corpora, including historical documents, red classic texts, and culturally loaded terms. Three automatic evaluation metrics—GLEU, METEOR, and COMET—are employed for a comprehensive assessment.