Saved in:
| Main Authors: | , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | https://arxiv.org/abs/2509.15550 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911199220727808 |
|---|---|
| author | Zhu, Xiaowei Ren, Yubing Fang, Fang Tan, Qingfeng Wang, Shi Cao, Yanan |
| author_facet | Zhu, Xiaowei Ren, Yubing Fang, Fang Tan, Qingfeng Wang, Shi Cao, Yanan |
| contents | The rapid advancement of large language models (LLMs) has blurred the line between AI-generated and human-written text. This progress brings societal risks such as misinformation, authorship ambiguity, and intellectual property concerns, highlighting the urgent need for reliable AI-generated text detection methods. However, recent advances in generative language modeling have resulted in significant overlap between the feature distributions of human-written and AI-generated text, blurring classification boundaries and making accurate detection increasingly challenging. To address the above challenges, we propose a DNA-inspired perspective, leveraging a repair-based process to directly and interpretably capture the intrinsic differences between human-written and AI-generated text. Building on this perspective, we introduce DNA-DetectLLM, a zero-shot detection method for distinguishing AI-generated and human-written text. The method constructs an ideal AI-generated sequence for each input, iteratively repairs non-optimal tokens, and quantifies the cumulative repair effort as an interpretable detection signal. Empirical evaluations demonstrate that our method achieves state-of-the-art detection performance and exhibits strong robustness against various adversarial attacks and input lengths. Specifically, DNA-DetectLLM achieves relative improvements of 5.55% in AUROC and 2.08% in F1 score across multiple public benchmark datasets. Code and data are available at https://github.com/Xiaoweizhu57/DNA-DetectLLM. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2509_15550 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | DNA-DetectLLM: Unveiling AI-Generated Text via a DNA-Inspired Mutation-Repair Paradigm Zhu, Xiaowei Ren, Yubing Fang, Fang Tan, Qingfeng Wang, Shi Cao, Yanan Computation and Language The rapid advancement of large language models (LLMs) has blurred the line between AI-generated and human-written text. This progress brings societal risks such as misinformation, authorship ambiguity, and intellectual property concerns, highlighting the urgent need for reliable AI-generated text detection methods. However, recent advances in generative language modeling have resulted in significant overlap between the feature distributions of human-written and AI-generated text, blurring classification boundaries and making accurate detection increasingly challenging. To address the above challenges, we propose a DNA-inspired perspective, leveraging a repair-based process to directly and interpretably capture the intrinsic differences between human-written and AI-generated text. Building on this perspective, we introduce DNA-DetectLLM, a zero-shot detection method for distinguishing AI-generated and human-written text. The method constructs an ideal AI-generated sequence for each input, iteratively repairs non-optimal tokens, and quantifies the cumulative repair effort as an interpretable detection signal. Empirical evaluations demonstrate that our method achieves state-of-the-art detection performance and exhibits strong robustness against various adversarial attacks and input lengths. Specifically, DNA-DetectLLM achieves relative improvements of 5.55% in AUROC and 2.08% in F1 score across multiple public benchmark datasets. Code and data are available at https://github.com/Xiaoweizhu57/DNA-DetectLLM. |
| title | DNA-DetectLLM: Unveiling AI-Generated Text via a DNA-Inspired Mutation-Repair Paradigm |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2509.15550 |