Year: 2024
Type: article
Abstract: This study compares the paraphrasing abilities of ChatGPT and Kimi AI in Jane Eyre, based on a linguistic typology standard. Firstly, the data shows a general paraphrasing situation across each facet of the standard, demonstrating the two Large Language Models(LLMs) performed better in their original source languages, English and Chinese. Subsequently, qualitative case studies demonstrate the discrepancy of single quantitative study and delves deeper into the English-Chinese comparison behind the behaviour of the two models, which is analyzed with comparative Linguistics, advanced translation studies and even Literature theories. The results demonstrate that English is more hypotactic, using more variations of words or clause to express compared to Chinese. Chinese, on the other hand, is more paratactic, requiring more modifiers to concretize the same expression. In conclusion, this paper sheds light on and contributes to the linguistic characteristics of the source languages in developing and refining LLMs for NLP tasks. Meanwhile, concerns are raised about the limitations due to the limited case study and translation deviation. Additionally, this paper rasies suggestions for future AI modification and personalizes development of future LLMs and AI.
DOI: 10.54097/58nrg156
Author: Xiangyu He
Institution: Beijing Foreign Studies University
Language: English
Subfield: Language and Linguistics
Field: Arts and Humanities
Domain: Social Sciences
Open Access status: diamond
OpenAlex ID: W4405394454