defense arXiv Jan 27, 2026 · 9w ago
Xuecong Li, Xiaohong Li, Qiang Hu et al. · Tianjin University
Detects LLM-generated text by measuring log-perplexity variation across multiple LLM rewrites, outperforming Binoculars by 34.3% AUROC
Output Integrity Attack nlp
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.
llm transformer Tianjin University
ML Security Papers