EUVIMEDEUVIMED entdecken
Open Access · CC BY-NC 4.0

Limitations in cultural context: systematic biases of LLMs in Chinese sentiment analysis

Ruili Geng, Jiwen Zhang, Ruixian Yang, Mingzhe Quan, Xiang Zheng, Yishuai Xu

Information Research an international electronic journal · 20.03.2026 · 31 (iConf) · 178-192

DOI: 10.47989/ir31iconf64278 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Information Research an international electronic journal, online erschienen am 20.03.2026. Verfasst von Ruili Geng und 5 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-NC 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Introduction

Large language models (LLMs) are increasingly applied in Chinese sentiment analysis; however, their ability to interpret sentiment in culturally specific contexts remains uncertain. This study systematically evaluates the potential biases of LLM-based Chinese sentiment analysis, providing an empirical basis for model optimisation and responsible application.

Method

A human-AI comparison experiment was conducted at both word and sentence levels. Two hundred Chinese words from four categories and 225 sentences were used as materials. Judgment data on sentiment polarity, intensity, valence, and arousal were collected from native Chinese speakers and representative Chinese and international LLMs. Data analysis was performed using Chi-square tests and Kruskal–Wallis tests.

Results

In sentiment polarity judgment, LLMs are highly consistent with humans and outperform the traditional sentiment lexicon. For continuous dimensions (intensity, valence, arousal), LLMs generally show an exaggeration bias, especially for sensory-perceptual words and ironic sentences. The bias stems from the disembodied cognitive nature of models, exaggerated linguistic patterns in training data, and technical characteristics such as the attention mechanism.

Conclusion

This study reveals a systematic limitation of LLMs in Chinese sentiment analysis, characterised by accurate classification but exaggerated quantitative evaluations. Consequently, in practical applications, outputs related to sentiment intensity should be interpreted with caution.

Zitieren

APA

Geng, R., Zhang, J., Yang, R., Quan, M., Zheng, X., & Xu, Y. (2026). Limitations in cultural context: systematic biases of LLMs in Chinese sentiment analysis. Information Research an international electronic journal, 31(iConf), 178-192. https://doi.org/10.47989/ir31iconf64278

Vancouver

Geng R, Zhang J, Yang R, Quan M, Zheng X, Xu Y. Limitations in cultural context: systematic biases of LLMs in Chinese sentiment analysis. Information Research an international electronic journal. 2026;31(iConf):178-192. doi:10.47989/ir31iconf64278

BibTeX
@article{Geng2026,
  author = {Geng, Ruili and Zhang, Jiwen and Yang, Ruixian and Quan, Mingzhe and Zheng, Xiang and Xu, Yishuai},
  title = {Limitations in cultural context: systematic biases of LLMs in Chinese sentiment analysis},
  journal = {Information Research an international electronic journal},
  year = {2026},
  volume = {31},
  number = {iConf},
  pages = {178--192},
  doi = {10.47989/ir31iconf64278}
}
Kostenlos starten →