Exons-Detect: Identifying and Amplifying Exonic Tokens via Hidden-State Discrepancy for Robust AI-Generated Text Detection

Xiaowei Zhu; Yubing Ren; Fang Fang; Shi Wang; Yanan Cao; Li Guo

cs.CLMar 26, 2026

Exons-Detect: Identifying and Amplifying Exonic Tokens via Hidden-State Discrepancy for Robust AI-Generated Text Detection

Authors: Xiaowei Zhu, Yubing Ren, Fang Fang, Shi Wang, Yanan Cao, Li Guo

Organizations: Institute of Information Engineering, Chinese Academy of Sciences, Beijing, China · School of Cyber Security, University of Chinese Academy of Sciences, Beijing, China · Institute of Computing Science, Chinese Academy of Sciences, Beijing, China

Abstract

The rapid advancement of large language models has increasingly blurred the boundary between human-written and AI-generated text, raising societal risks such as misinformation dissemination, authorship ambiguity, and threats to intellectual property rights. These concerns highlight the urgent need for effective and reliable detection methods. While existing training-free approaches often achieve strong performance by aggregating token-level signals into a global score, they typically assume uniform token contributions, making them less robust under short sequences or localized token modifications. To address these limitations, we propose Exons-Detect, a training-free method for AI-generated text detection based on an exon-aware token reweighting perspective. Exons-Detect identifies and amplifies informative exonic tokens by measuring hidden-state discrepancy under a dual-model setting, and computes an interpretable translation score from the resulting importance-weighted token sequence. Empirical evaluations demonstrate that Exons-Detect achieves state-of-the-art detection performance and exhibits strong robustness to adversarial attacks and varying input lengths. In particular, it attains a 2.2% relative improvement in average AUROC over the strongest prior baseline on DetectRL.

Figures & tables

Figure 1: Advantages of Our Method Exons-Detect.

Detectors	M4		DetectRL Multi-LLM		DetectRL Multi-Domain		RealDet		Avg.
Detectors	AUROC	$\text{F}_{1}$	AUROC	$\text{F}_{1}$	AUROC	$\text{F}_{1}$	AUROC	$\text{F}_{1}$	AUROC	$\text{F}_{1}$
Training-based Methods
OpenAI-D	77.51	71.18	78.15	71.90	74.60	70.03	84.75	77.47	78.75	72.65
Biscope	79.74	73.08	79.97	73.20	76.52	71.64	92.88	86.90	82.28	76.21
R-Detect	61.91	67.14	67.40	66.56	79.19	73.38	65.93	67.72	68.61	68.70
Training-free Methods

Table 1: Detection performance (AUROC and F1 score) on public benchmark datasets.

Figure 3: Detection performance (AUROC curves) against various attacks.

Figure 4: Detection performance on different length.

Figure 5: Detection performance of Exons-Detect under different parameter settings.

Figure 6: Time costs for processing a single sample.

Setting $\downarrow$	M4	Multi-L	Multi-D	RealDet	Avg.
Exons-Detect	92.43	90.67	90.46	94.98	92.14
w/o $\log\mathrm{PPL}^{W}_{M}(\hat{s})$	91.28	85.32	80.46	93.97	87.76
w/o $g(\cdot)$	91.92	89.63	88.66	94.64	91.21
Model Family $\downarrow$
Falcon-7B	92.43	90.67	90.46	94.98	92.14
LLaMA-7B	88.64	92.34	90.31	94.47	91.44

Table 2: Ablation study results under different settings.

Appendix figures & tables3 assets

Supplementary material from the paper’s appendix.

Appendix

Detectors	M4		DetectRL Multi-LLM		DetectRL Multi-Domain		RealDet		Avg.
Detectors	AUROC	$\text{F}_{1}$	AUROC	$\text{F}_{1}$	AUROC	$\text{F}_{1}$	AUROC	$\text{F}_{1}$	AUROC	$\text{F}_{1}$
Nonlinear mapping
$\alpha=10$	92.43	88.05	90.67	84.95	90.46	86.59	94.98	91.30	92.14	87.72
$\alpha=20$	92.25	87.55	90.87	85.20	90.89	87.05	95.02	91.29	92.26	87.77
Linear mapping
$\alpha=10$	92.24	87.92	90.90	84.93	91.18	87.24	95.03	91.02	92.34	87.78

Appendix

Table 3: Detection performance (AUROC and F1 score) with different mapping functions (

\theta=0.15

Parameter Setting	DetectRL Multi-LLM		DetectRL Multi-Domain		Avg.
Parameter Setting	AUROC	$\text{F}_{1}$	AUROC	$\text{F}_{1}$	AUROC	$\text{F}_{1}$
Reverse
$L=2\quad(\theta=0.15,\alpha=10)$	89.09	84.39	88.24	84.53	88.67	84.46
$L=4\quad(\theta=0.15,\alpha=10)$	89.41	85.07	88.38	84.65	88.90	84.86
$L=8\quad(\theta=0.15,\alpha=10)$	89.13	84.36	88.56	84.70	88.85	84.53
$L=16\quad(\theta=0.15,\alpha=10)$	90.63	85.38	90.55	86.79	90.59	86.09

Appendix

Table 4: Detection performance with different hidden layers.

Setting ( $L=16$ ) $\downarrow$	$\theta=0.05$	$\theta=0.10$	$\theta=0.15$	$\theta=0.20$
M4
$\alpha=2$	92.03	92.09	92.23	92.09
$\alpha=6$	91.99	92.06	92.57	92.38
$\alpha=10$	91.89	91.88	92.60	92.50
Multi-LLM
$\alpha=2$	89.58	89.64	89.72	89.47

Appendix

Table 5: AUROC under different hyperparameters.

AI-Generated Text Detection

AI-generated text detection develops methods to distinguish human-written, machine-generated, and AI-edited text.

See all 170 papers

Findings of the Counter Turing Test: AI-Generated Text Detection
May 20, 2026Rajarshi Roy, Gurpreet Singh, Ashhar Aziz +16LLM Evaluation AI-Generated Text Detection
Kalyani Government Engineering College, India. · IIIT Guwahati, India. · IIIT Delhi, India. +10
On the Salience of Low-Probability Tokens for AI-Generated Text Detection: A Multiscale Uncertainty Perspective
Jun 1, 2026Yikai Guo, Bin Wang, Xilai Fan +2AI-Generated Text Detection LLM Uncertainty Estimation
Beijing Institute of Computer Technology and Application, Beijing, China · Academy of Mathematics and Systems Science, Chinese Academy of Sciences, Beijing, China · Southeast University, Nanjing, China +1
Automatic detection of Gen-AI texts: A comparative framework of neural models
Mar 19, 2026Cristian Buttaro, Irene AmeriniAI-Generated Text Detection Robustness of Machine-Generated Text Detection
Sapienza University in Rome Latina

May 20, 2026J/K move · Enter open · S save

Findings of the Counter Turing Test: AI-Generated Text Detection

Rajarshi Roy, Gurpreet Singh, Ashhar Aziz, Shashwat Bajpai+15

Kalyani Government Engineering College, India. · IIIT Guwahati, India. · IIIT Delhi, India.+10

LLM Evaluation AI-Generated Text Detection

Open paper AskPDF arXiv

The growing capability of large language models to produce fluent, contextually coherent text has created mounting pressure on the systems and institutions responsible for ensuring the authenticity of digital content. Advanced generative models such as GPT-4, Claude 3.5, and Llama can produce highly coherent and human-like text, making it increasingly difficult to differentiate between human-written and AI-generated content. While these models have transformative applications, their misuse has raised concerns about misinformation, biased narratives, and security threats. This paper provides a comprehensive analysis of state-of-the-art AI-generated text detection techniques and evaluates their effectiveness through the Counter Turing Test (CT2) shared tasks. Task A (Binary Classification) required participants to distinguish between human-written and AI-generated text, while Task B (Model Attribution) focused on identifying the specific language model responsible for generating a given text. The results demonstrated high performance in binary classification, with the top system achieving an F1 score of 1.0000, but significantly lower scores in model attribution, where the best system achieved 0.9531, highlighting the increased complexity of this task. The top-performing teams leveraged fine-tuned transformer models, ensemble learning, and hybrid detection approaches, with DeBERTa-based and BART-based methods demonstrating strong results. However, the lower scores in Task B underscore the challenges of distinguishing outputs from different LLMs, necessitating further research into adversarial robustness, feature extraction, and cross-domain generalization.

Abstract

AI-Generated Text Detection

Explore similar work

Findings of the Counter Turing Test: AI-Generated Text Detection

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

Automatic detection of Gen-AI texts: A comparative framework of neural models