The Exploration of context sensitivity in large language models — A Comparative study of translation practice with Tencent Hunyuan and ChatGPT 4 as examples
Bo Han
Lecturer of Foreign Languages Department, Northwest Normal University
Hua Zhu
Assistant Professor, Director, AI Translation Lab TFSU, Tianjin Foreign Studies University
DOI: https://doi.org/10.59429/esta.v12i1.9652
Keywords: Large language model; Context sensitivity; ChatGPT 4; Translation practice
Abstract
This research compares Tencent Hunyuan and ChatGPT 4 in translation. While these models excel in long-text processing and multilingual translation, they still struggle with capturing deeper meanings, tones, and culturally specific expressions. Context sensitivity affects translation accuracy and fluency, and each model has pros and cons in complex contexts. The study offers references for improving model translation and understanding language - processing mechanisms.
References
[1] Victor Muntés-Mulero,Patricia Paladini Adell,Cristina España-Bonet,and Lluís Màrquez.Context -Aware Machine Trans lation for Software Localization[J].Proceedings of the 16th EAMT Conference,77–80,May
2012,Trento,Italy.European Association for Machine Translation.
[2] Zhao Mingli,Liu Yuhui.The Impact of Linguistic and Situational Context on Translation
[3] Zhou Shengnan.Application and Thinking of Machine Tra anslation in Different Texts against the Backdrop of Artificial Intelligence[J].The Guide Of Science & Education,2022,(7),39-41.
[4] Jin,Yucheng.Fan Hua.Shanghai:Shanghai Literature and Art Publishing House,2012.
[5] Jin,Yucheng.The Flowers of War.Translated by Michelle Yeh and Christopher Smith,Astra House,2020.