triospect

Code base for TACL 2026 paper Triospect: A Three-Dimensional Framework for Robust Statistical AI-Generated Text Detection Against Diverse Attacks

Project structure

├── scripts/          <- Scripts for baseline detectors and Triospect batch processing and evaluation.
├── data/             <- Humanize-16K and RAID subset with content/expression-preserving transformations.
└── exp_test/         <- Main experimental results.

Citation

If you find the content of this repo useful, please cite it as follows:

@article{bao2026triospect,
  title={Triospect: A Three-Dimensional Framework for Robust Statistical AI-Generated Text Detection Against Diverse Attacks},
  author={Bao, Guangsheng and Rong, Lihua and Zhao, Yanbin and Yu, Xiao and Zhou, Qiji and Zhang, Yue},
  journal={arXiv preprint arXiv:2606.31074},
  year={2026}
}