DetectRL-X: Towards Reliable Multilingual and Real-World LLM-Generated Text Detection
Fuente:
arXiv
Gespeichert in:
| Hauptverfasser: | , , , , , , , , , , |
|---|---|
| Format: | Preprint |
| Veröffentlicht: |
2026
|
| Schlagworte: | |
| Online-Zugang: | |
| Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
| _version_ | 1866916024177131520 |
|---|---|
| author | Wu, Junchao Liu, Yefeng Zhu, Chenyu Zhang, Hao Wu, Zeyu Shi, Tianqi Du, Yichao Wang, Longyue Luo, Weihua Su, Jinsong Wong, Derek F. |
| author_facet | Wu, Junchao Liu, Yefeng Zhu, Chenyu Zhang, Hao Wu, Zeyu Shi, Tianqi Du, Yichao Wang, Longyue Luo, Weihua Su, Jinsong Wong, Derek F. |
| contents | The effective detection and governance of Large Language Model (LLM) generated content has become increasingly critical due to the growing risk of misuse. Despite the impressive performance of existing detectors, their reliability and potential in multilingual, real-world scenarios remain largely underexplored. In this study, we introduce DetectRL-X, a comprehensive multilingual benchmark designed to evaluate advanced detectors across 8 dimensions. The benchmark encompasses 8 languages commonly used in commercial contexts and collects human-written texts from 6 domains highly susceptible to LLM misuse. To better aligned with real-world applications, We create LLM-generated texts using 4 popular commercial LLMs, and include typical AI-assisted writing operations such as polishing, expanding, and condensing to capture authentic usage patterns. Furthermore, we develop a multilingual framework for paraphrasing and perturbation attacks to simulate diverse human modifications and writing noise, enabling stress testing of detectors across languages. Experimental results on DetectRL-X reveal the strengths and limitations of current state-of-the-art detectors when applied to diverse linguistic resources. We further analyze how domains, generators, attack strategies, text length, and refinement operations influence performance in different languages, underscoring DetectRL-X as an effective benchmark for strengthening multilingual and language-specific detectors. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2605_15518 |
| institution | arXiv |
| publishDate | 2026 |
| record_format | arxiv |
| spellingShingle | DetectRL-X: Towards Reliable Multilingual and Real-World LLM-Generated Text Detection Wu, Junchao Liu, Yefeng Zhu, Chenyu Zhang, Hao Wu, Zeyu Shi, Tianqi Du, Yichao Wang, Longyue Luo, Weihua Su, Jinsong Wong, Derek F. Computation and Language The effective detection and governance of Large Language Model (LLM) generated content has become increasingly critical due to the growing risk of misuse. Despite the impressive performance of existing detectors, their reliability and potential in multilingual, real-world scenarios remain largely underexplored. In this study, we introduce DetectRL-X, a comprehensive multilingual benchmark designed to evaluate advanced detectors across 8 dimensions. The benchmark encompasses 8 languages commonly used in commercial contexts and collects human-written texts from 6 domains highly susceptible to LLM misuse. To better aligned with real-world applications, We create LLM-generated texts using 4 popular commercial LLMs, and include typical AI-assisted writing operations such as polishing, expanding, and condensing to capture authentic usage patterns. Furthermore, we develop a multilingual framework for paraphrasing and perturbation attacks to simulate diverse human modifications and writing noise, enabling stress testing of detectors across languages. Experimental results on DetectRL-X reveal the strengths and limitations of current state-of-the-art detectors when applied to diverse linguistic resources. We further analyze how domains, generators, attack strategies, text length, and refinement operations influence performance in different languages, underscoring DetectRL-X as an effective benchmark for strengthening multilingual and language-specific detectors. |
| title | DetectRL-X: Towards Reliable Multilingual and Real-World LLM-Generated Text Detection |
| topic | Computation and Language |
| url | https://arxiv.org/abs/2605.15518 |