Proceedings

VaryBalance: Detecting LLM-Generated Text Through Variation

Xuecong Li, Xiaohong Li, Qiang Hu, Yao Zhang, Junjie Wang

Download BibTeX Cite

Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence. Main Track. Pages 5819-5827.

https://doi.org/10.24963/ijcai.2026/648

Keywords

  • Natural Language Processing: Language models
  • Natural Language Processing: Text classification
  • Natural Language Processing: Interpretability and analysis of models for NLP

Abstract

Detecting text generated by large language models~(LLMs) is crucial but challenging. Existing detectors depend on impractical assumptions, such as white-box settings, or solely rely on text-level features, leading to imprecise detection ability. In this paper, we propose a simple but effective and practical LLM-generated text detection method, VaryBalance. The core of VaryBalance is that, compared to LLM-generated texts, there is a greater difference between human texts and their rewritten version via LLMs. Leveraging this observation, VaryBalance quantifies this through mean standard deviation and distinguishes human texts and LLM-generated texts. Comprehensive experiments demonstrated that VaryBalance outperforms the state-of-the-art detectors, i.e., Binoculars, by up to 34.3% in terms of AUROC, and maintains robustness against multiple generating models and languages.

Xuecong Li, Xiaohong Li, Qiang Hu, Yao Zhang, and Junjie Wang. 2026. VaryBalance: Detecting LLM-Generated Text Through Variation. In Proceedings of the Thirty-Fifth International Joint Conference on Artificial Intelligence, IJCAI-26. IJCAI, 5819–5827. https://doi.org/10.24963/ijcai.2026/648
Welcome Future Conferences AI Journal Division IJCAI Proceedings About IJCAI IJCAI Awards