On the Salience of Low-Probability Tokens for AI-Generated Text Detection: A Multiscale Uncertainty Perspective

Yikai Guo, Bin Wang, Xilai Fan, Wenjun Ke, Haoran Luo
Proceedings of the 43rd International Conference on Machine Learning, PMLR 306:38379-38399, 2026.

Abstract

AI-generated text increasingly blends with human writing, raising practical risks such as misinformation, academic misuse, and corpora contamination. While statistical detectors are appealing for efficiency and generalization, they suffer from two key limitations. (i) Boilerplate dominance, boilerplate tokens shared across human and LLM writing can overwhelm discriminative signals. (ii) Brittle point estimates, relying on a single probability score yields unstable decisions under adversarial manipulations. To address these issues, we propose $\textbf{Uncertainty}$, a multiscale uncertainty estimator that focuses on informative low-probability tokens, which more clearly expose distributional discrepancies. Locally, it alleviates boilerplate dominance by averaging the log-probabilities of low-probability tokens; globally, it reduces brittleness by capturing the distributional shape of this low-probability region via Rényi entropy. We further extend the detector to $\textbf{Uncertainty++}$ via conditional independent sampling, yielding a more stable uncertainty estimation. Experiments across seven datasets and sixteen LLMs demonstrate high effectiveness, generalization, and robustness. Our code is available at github.com/guoyikai2000/Uncertainty-AIGT.

Cite this Paper

BibTeX
@InProceedings{pmlr-v306-guo26w, title = {On the Salience of Low-Probability Tokens for {AI}-Generated Text Detection: A Multiscale Uncertainty Perspective}, author = {Guo, Yikai and Wang, Bin and Fan, Xilai and Ke, Wenjun and Luo, Haoran}, booktitle = {Proceedings of the 43rd International Conference on Machine Learning}, pages = {38379--38399}, year = {2026}, editor = {Zhang, Tong and Dudik, Miroslav and Jaggi, Martin and Agarwal, Alekh and Li, Sharon and Schuurmans, Dale and Zhu, Jerry and Berkenkamp, Felix and Dong, Hanze and Bietti, Alberto}, volume = {306}, series = {Proceedings of Machine Learning Research}, month = {06--11 Jul}, publisher = {PMLR}, pdf = {https://raw.githubusercontent.com/mlresearch/v306/main/assets/guo26w/guo26w.pdf}, url = {https://proceedings.mlr.press/v306/guo26w.html}, abstract = {AI-generated text increasingly blends with human writing, raising practical risks such as misinformation, academic misuse, and corpora contamination. While statistical detectors are appealing for efficiency and generalization, they suffer from two key limitations. (i) Boilerplate dominance, boilerplate tokens shared across human and LLM writing can overwhelm discriminative signals. (ii) Brittle point estimates, relying on a single probability score yields unstable decisions under adversarial manipulations. To address these issues, we propose $\textbf{Uncertainty}$, a multiscale uncertainty estimator that focuses on informative low-probability tokens, which more clearly expose distributional discrepancies. Locally, it alleviates boilerplate dominance by averaging the log-probabilities of low-probability tokens; globally, it reduces brittleness by capturing the distributional shape of this low-probability region via Rényi entropy. We further extend the detector to $\textbf{Uncertainty++}$ via conditional independent sampling, yielding a more stable uncertainty estimation. Experiments across seven datasets and sixteen LLMs demonstrate high effectiveness, generalization, and robustness. Our code is available at github.com/guoyikai2000/Uncertainty-AIGT.} }
Endnote
%0 Conference Paper %T On the Salience of Low-Probability Tokens for AI-Generated Text Detection: A Multiscale Uncertainty Perspective %A Yikai Guo %A Bin Wang %A Xilai Fan %A Wenjun Ke %A Haoran Luo %B Proceedings of the 43rd International Conference on Machine Learning %C Proceedings of Machine Learning Research %D 2026 %E Tong Zhang %E Miroslav Dudik %E Martin Jaggi %E Alekh Agarwal %E Sharon Li %E Dale Schuurmans %E Jerry Zhu %E Felix Berkenkamp %E Hanze Dong %E Alberto Bietti %F pmlr-v306-guo26w %I PMLR %P 38379--38399 %U https://proceedings.mlr.press/v306/guo26w.html %V 306 %X AI-generated text increasingly blends with human writing, raising practical risks such as misinformation, academic misuse, and corpora contamination. While statistical detectors are appealing for efficiency and generalization, they suffer from two key limitations. (i) Boilerplate dominance, boilerplate tokens shared across human and LLM writing can overwhelm discriminative signals. (ii) Brittle point estimates, relying on a single probability score yields unstable decisions under adversarial manipulations. To address these issues, we propose $\textbf{Uncertainty}$, a multiscale uncertainty estimator that focuses on informative low-probability tokens, which more clearly expose distributional discrepancies. Locally, it alleviates boilerplate dominance by averaging the log-probabilities of low-probability tokens; globally, it reduces brittleness by capturing the distributional shape of this low-probability region via Rényi entropy. We further extend the detector to $\textbf{Uncertainty++}$ via conditional independent sampling, yielding a more stable uncertainty estimation. Experiments across seven datasets and sixteen LLMs demonstrate high effectiveness, generalization, and robustness. Our code is available at github.com/guoyikai2000/Uncertainty-AIGT.
APA
Guo, Y., Wang, B., Fan, X., Ke, W. & Luo, H.. (2026). On the Salience of Low-Probability Tokens for AI-Generated Text Detection: A Multiscale Uncertainty Perspective. Proceedings of the 43rd International Conference on Machine Learning, in Proceedings of Machine Learning Research 306:38379-38399 Available from https://proceedings.mlr.press/v306/guo26w.html.

Related Material