AdaDetectGPT: Adaptive Detection of LLM-Generated Text with Statistical Guarantees
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866911412710801408 |
|---|---|
| author | Zhou, Hongyi Zhu, Jin Su, Pingfan Ye, Kai Yang, Ying Gavioli-Akilagun, Shakeel A O B Shi, Chengchun |
| author_facet | Zhou, Hongyi Zhu, Jin Su, Pingfan Ye, Kai Yang, Ying Gavioli-Akilagun, Shakeel A O B Shi, Chengchun |
| contents | We study the problem of determining whether a piece of text has been authored by a human or by a large language model (LLM). Existing state of the art logits-based detectors make use of statistics derived from the log-probability of the observed text evaluated using the distribution function of a given source LLM. However, relying solely on log probabilities can be sub-optimal. In response, we introduce AdaDetectGPT -- a novel classifier that adaptively learns a witness function from training data to enhance the performance of logits-based detectors. We provide statistical guarantees on its true positive rate, false positive rate, true negative rate and false negative rate. Extensive numerical studies show AdaDetectGPT nearly uniformly improves the state-of-the-art method in various combination of datasets and LLMs, and the improvement can reach up to 37\%. A python implementation of our method is available at https://github.com/Mamba413/AdaDetectGPT. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2510_01268 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | AdaDetectGPT: Adaptive Detection of LLM-Generated Text with Statistical Guarantees Zhou, Hongyi Zhu, Jin Su, Pingfan Ye, Kai Yang, Ying Gavioli-Akilagun, Shakeel A O B Shi, Chengchun Computation and Language Artificial Intelligence Machine Learning We study the problem of determining whether a piece of text has been authored by a human or by a large language model (LLM). Existing state of the art logits-based detectors make use of statistics derived from the log-probability of the observed text evaluated using the distribution function of a given source LLM. However, relying solely on log probabilities can be sub-optimal. In response, we introduce AdaDetectGPT -- a novel classifier that adaptively learns a witness function from training data to enhance the performance of logits-based detectors. We provide statistical guarantees on its true positive rate, false positive rate, true negative rate and false negative rate. Extensive numerical studies show AdaDetectGPT nearly uniformly improves the state-of-the-art method in various combination of datasets and LLMs, and the improvement can reach up to 37\%. A python implementation of our method is available at https://github.com/Mamba413/AdaDetectGPT. |
| title | AdaDetectGPT: Adaptive Detection of LLM-Generated Text with Statistical Guarantees |
| topic | Computation and Language Artificial Intelligence Machine Learning |
| url | https://arxiv.org/abs/2510.01268 |