面向AI生成文本的鲁棒细粒度检测模型与2.4M多语言数据集
Robust and Fine-Grained Detection of AI Generated Texts
研究团队提出一组面向token分类的检测模型，专门针对人类与LLM共同撰写的文本训练，在未见领域、未见生成器、非母语者文本及对抗输入上均表现良好。团队同时发布了一个超过2.4M条、覆盖23种语言、主要由多个主流闭源LLM参与撰写的文本数据集，并给出模型在各领域与各生成器上的性能对比，以及不同对抗方法、输入长度对检测效果的影响分析。
Authors:Ram Mohan Rao Kadiyala, Siddartha Pullakhandam, Kanwal Mehreen, Drishti Sharma, Siddhant Gupta, Jebish Purbey, Ashay Srivastava, Subhasya TippaReddy, Arvind Reddy Bobbili, Suraj Telugara Chandrashekhar, Ankan Ghosh, Modabbir Adeeb, Srinadh Vura, Suman Debnath, Hamza Farooq
Abstract:An ideal detection system for machine generated content is supposed to work well on any generator as many more advanced LLMs come into existence day by day. Existing systems often struggle with accurately identifying AI-generated content over shorter texts. Further, not all texts might be entirely authored by a human or LLM, hence we focused more over partial cases i.e human-LLM co-authored texts. Our paper introduces a set of models built for the task of token classification which are trained on an extensive collection of human-machine co-authored texts, which performed well over texts of unseen domains, unseen generators, texts by non-native speakers and those with adversarial inputs. We also introduce a new dataset of over 2.4M such texts mostly co-authored by several popular proprietary LLMs over 23 languages. We also present findings of our models' performance over each texts of each domain and generator. Additional findings include comparison of performance against each adversarial method, length of input texts and characteristics of generated texts compared to the original human authored texts.
| Comments: | 18 pages, 6 figures |
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI); Machine Learning (cs.LG) |
| Cite as: | arXiv:2504.11952 [cs.CL] |
| (or arXiv:2504.11952v4 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2504.11952 arXiv-issued DOI via DataCite |
Submission history
From: Ram Mohan Rao Kadiyala [view email]
[v1]
Wed, 16 Apr 2025 10:29:30 UTC (7,712 KB)
[v2]
Thu, 22 May 2025 10:09:18 UTC (8,739 KB)
[v3]
Thu, 31 Jul 2025 09:14:25 UTC (7,480 KB)
[v4]
Sat, 24 Oct 2026 01:05:00 UTC (7,473 KB)
来源：arXiv cs.CL · arxiv.org