Empirical Research on High School English Composition Evaluation via Large Language Models

DOI: https://doi.org/10.62517/jhet.202515506

Author(s)

Siyu Liu

Affiliation(s)

School of Foreign Languages, Northwest Minzu University, Lanzhou, Gansu, China

Abstract

In recent years, artificial intelligence technology has developed rapidly across various domains, with large language models being increasingly widely applied in the field of education. Traditional evaluation of high school English compositions is plagued by issues such as prolonged teacher feedback cycles, strong subjectivity, and insufficient personalized guidance. While some schools have adopted large language models to assist in composition grading, challenges still persist, making it particularly important to test and investigate the evaluation capabilities of these models. To examine the English composition evaluation abilities of large language models, this study selected 100 high school English compositions, designed 20 prompts, and employed two domestic large language models, DeepSeek and Wenxin Large Model, as testing tools to conduct a comprehensive assessment of their capabilities in both scoring and correction. Finally, this paper proposes recommendations for the application of large language models in English education, offering practical foundations for integrating intelligent educational tools with traditional teaching practices.

Keywords

Large Language Model; Senior High School English; Human-Machine Collaboration

References

[1] Ruan Fenghui. Personalized Teaching Path of High School English Writing Assisted by Digital Tools [J]. Campus English, 2024, (50): 43 - 45 [2] Formulated by the Ministry of Education of the People's Republic of China, the General high school English curriculum Standard (2017 Edition, Revised in 2020) [S] was published by the People's Education Press in Beijing in 2020. [3] Wei Shunping, Zhang Yue, Ran Rou. Testing the Chinese Essay Assessment Capabilities of Domestic large language models [J]. Modern Educational Technology, 2025, 35(03): 24-33. [4] Xia Zhiting. A Study on the Effects and Impacts of Senior High School English Essay Correction [J]. Overseas English, 2021, (17): 89-90+92. [5] Wu Jiahui. Research on Key Technologies of an Automatic Evaluation System for English Writing Learning [D]. Beijing University of Posts and Telecommunications, 2024. [6] Wang Fang. Research on Strategies for DeepSeek to Empower high school English teaching [N]. Shanxi Science and Technology News, 2025-03-18(A06). [7] Liu Zhenghui, Zhao Xiaoyan, Ruan Libin. Empirical Research on the Effect of the Generative AI - Enhanced English Teaching Method Course from the Students' Perspective[J]. Journal of Shanxi Institute of Energy, 2025, 38(03): 96 - 99. [8] Huang Jinchun. A Study on Cultivating Senior High School Students' English Writing Ability Supported by AI Technology[J]. Education Circle, 2025, (23): 41-43