Relational Over-Regularization: Graph-Based AI-Generated Text Detection via Sentence Transition Deviation

Hyeonchu Park; Bugeun Kim

cs.AIAug 27, 2026

Relational Over-Regularization: Graph-Based AI-Generated Text Detection via Sentence Transition Deviation

Authors: Hyeonchu Park, Bugeun Kim

Organizations: Department of Artificial Intelligence, Chung-Ang University, Republic of Korea

Abstract

Detecting AI-generated text (AIGT) remains challenging because existing approaches rely on token-level statistical signals or independent stylometric features, causing them to overfit to specific generators and fail under distribution shift. We identify a structural signal at the sentence-pair level: LLMs produce inter-sentence transition variance that deviates from human writing through inflated variance driven by recurring similarity bursts at paragraph boundaries and templated transitions. We formalize this as Relational Over-Regularization (ROR) and validate it across four benchmarks (p < 0.001). The central contribution is this relational problem formulation, not a novel GNN architecture; CSFG is one concrete instantiation for operationalizing ROR. To exploit this signal, we propose the Cross-Source Stylometric Fingerprint Graph (CSFG), a graph-based framework that encodes positional, sequential, semantic, and transition deviation signals as learnable GNN edge features. The per-edge signed deviation δ_ij operationalizes ROR without hand-crafted thresholds and acts as a false-positive calibrator. CSFG achieves 97.14% accuracy under binary detection, outperforming the strongest graph-based baseline by 11.14 pp, with a false-positive rate of 1.57% and robust generalization to unseen LLMs in the inflated-variance regime; detection degrades for generators whose transition variance falls at or below the human baseline.

Figures & tables

Figure 1: Overview of the CSFG framework.

Dataset	$n_{\text{human}}$	$n_{\text{AI}}$	$\mu_{\text{human}}$	$\mu_{\text{AI}}$	$p$ -value	$r$
HC3	1000	1000	$5.72\times 10^{-7}$	$1.05\times 10^{-6}$	$3.34\times 10^{-45}$	$+0.38$
M4	1000	1000	$2.52\times 10^{-6}$	$3.06\times 10^{-6}$	$5.21\times 10^{-22}$	$+0.25$
MULTITuDE	1000	1000	$1.13\times 10^{-6}$	$2.12\times 10^{-6}$	$9.73\times 10^{-60}$	$+0.32$
AIGTBench	1000	1000	$1.95\times 10^{-6}$	$2.84\times 10^{-6}$	$1.34\times 10^{-73}$	$+0.22$

Table 1: Mann–Whitney

U

test results for

\mathrm{Var}(\mathcal{T}(D))

across four benchmarks.

\mu

denotes the per-group mean variance.

r

: rank-biserial correlation as effect size. Positive

r

indicates

\mathrm{Var}(\mathcal{T}(D))

is higher in AI-generated text than in human-written text. All differences are statistically significant (

p<0.001

Figure 2: Distribution of $\mathrm{Var}(\mathcal{T}(D))$ for human-written (blue) and AI-generated (orange) text across four benchmarks, as violin plots. The horizontal bar: median. AIGT consistently exhibits higher variance and a heavier upper tail than human-written text, driven by recurring similarity bursts at paragraph boundaries and templated transitions. Significance: ${}^{***}p<0.001$ . Effect sizes ( $r$ ) and exact $p$ -values below each panel.

Detector	ACC (%)	FPR (%)	FNR (%)
F-DetectGPT	70.45	17.30	17.30
DNAGPT	81.23	24.60	19.15
LASTDE	58.93	38.82	43.31
Likelhood	60.75	36.96	41.54
DetectAnyLLM	64.56	18.65	52.23
DeTeCtive	87.78	6.74	17.70

Table 2: Binary detection performance under standard evaluation (Setting 1). † RoBERTa is evaluated in an API-only setting via the Hugging Face inference endpoint.

Detector	Generator	Accuracy (%)	FPR (%)	FNR (%)
LASTDE	Claude Sonnet 4.6	50.50	72.00	26.40
	GPT-5	51.80	89.00	6.10
	Gemini 2.5 Flash Lite	52.30	18.50	77.80
Likelihood	Claude Sonnet 4.6	55.50	68.20	20.10
	GPT-5	55.70	37.40	51.40
	Gemini 2.5 Flash Lite	52.10	39.00	57.10

Table 3: Detection performance on unseen LLMs (Setting 2). Generators are excluded from training.

Detector	Perturbation	Accuracy (%)	FPR (%)	FNR (%)
LASTDE	Paraphrasing	49.45	19.61	53.77
	Humanization	54.16	33.33	47.14
	Back-translation	66.73	27.45	33.88
Likelihood	Paraphrasing	67.53	27.45	32.99
	Humanization	66.36	19.61	35.10
	Back-translation	74.31	25.49	25.71

Table 4: Detection performance under text perturbation (Setting 3) on Claude Sonnet 4.6. Results for GPT-5 and Gemini 2.5 Flash Lite are provided in Appendix D .

Appendix figures & tables12 assets

Supplementary material from the paper’s appendix.

Appendix

$\lambda$	Accuracy (%)		FPR (%)		FNR (%)
$\lambda$	$\mu$	$\sigma$	$\mu$	$\sigma$	$\mu$	$\sigma$
0.0 (no aux.)	96.25	0.42	2.27	0.31	5.23	0.81
0.1	96.95	0.26	1.68	0.53	4.42	0.83
0.3	97.02	0.24	1.51	0.40	4.44	0.58
0.5	97.19	0.17	1.60	0.38	4.02	0.46
1.0	97.22	0.22	1.48	0.39	4.08	0.45

Appendix

Table 5: Effect of auxiliary loss weight

\lambda

on binary detection performance. All results are averaged over 10 runs;

\sigma

denotes standard deviation in percentage points (pp). No aux. (

\lambda=0.0

) denotes training without the source attribution head. Bold indicates the best value per column.

$\theta$	Accuracy (%)		FPR (%)		FNR (%)
$\theta$	$\mu$	$\sigma$	$\mu$	$\sigma$	$\mu$	$\sigma$
0.3	97.20	0.24	1.48	0.30	4.11	0.55
0.4	97.22	0.21	1.51	0.29	4.05	0.51
0.5	97.14	0.21	1.58	0.36	4.14	0.48
0.6	97.23	0.17	1.53	0.43	4.01	0.45
0.7	97.19	0.22	1.45	0.30	4.16	0.44

Appendix

Table 6: Sensitivity to semantic edge threshold

\theta

. Lower

\theta

yields denser graphs with more long-range semantic edges. All results are averaged over 10 runs;

\sigma

denotes standard deviation in percentage points (pp). Bold indicates the best value per column.

Edge Features	Accuracy (%)		FPR (%)		FNR (%)
Edge Features	$\mu$	$\sigma$	$\mu$	$\sigma$	$\mu$	$\sigma$
Full (ours)	97.14	0.27	1.57	0.38	4.15	0.54
w/o sim	96.92	0.21	1.59	0.49	4.57	0.66
w/o dist	96.89	0.24	1.54	0.51	4.67	0.76
w/o is_seq	96.94	0.17	1.69	0.45	4.42	0.61
w/o trans_dev	96.97	0.23	1.65	0.34	4.40	0.67

Appendix

Table 7: Contribution of each edge feature component. Each variant zeros out the corresponding dimension of

\mathbf{e}_{ij}

at collation time; graph structure and model architecture remain unchanged. All results are averaged over 10 runs;

\sigma

denotes standard deviation in percentage points (pp). Bold indicates the best value per column.

# Layers	Accuracy (%)		FPR (%)		FNR (%)
# Layers	$\mu$	$\sigma$	$\mu$	$\sigma$	$\mu$	$\sigma$
1	96.59	0.19	2.06	0.31	4.75	0.63
2	97.09	0.26	1.53	0.28	4.29	0.71
3 (ours)	97.15	0.25	1.42	0.43	4.27	0.67
4	97.10	0.31	1.53	0.46	4.26	0.68

Appendix

Table 8: Effect of GNN depth on binary detection performance. All results are averaged over 10 runs;

\sigma

denotes standard deviation in percentage points (pp). All variants use the full edge feature vector and default configuration (

\theta=0.6

\lambda=1.0

). Bold indicates the best value per column.

Group	$n$	TV mean	TV std	$p$ -value	KS $D$	FNR (%)
Human	500	$1.1\times 10^{-3}$	$6\times 10^{-4}$	–	–	–
GPT-5	500	$0.8\times 10^{-3}$	$5\times 10^{-4}$	$6.05\times 10^{-13}$	0.2389	61.64
Claude Sonnet 4.6	500	$1.5\times 10^{-3}$	$5\times 10^{-4}$	$1.09\times 10^{-19}$	0.2915	20.46
Gemini 2.5 Flash Lite	500	$2.3\times 10^{-3}$	$6\times 10^{-4}$	$8.88\times 10^{-102}$	0.6482	14.81

Appendix

Table 9: Per-generator transition variance statistics on the Setting 2 test set. TV mean and TV std denote the mean and standard deviation of

\mathrm{Var}(T(D))

per group.

p

-value: Mann-Whitney

U

test against the human group. KS

D

: Kolmogorov-Smirnov distance against the human CDF. FNR: miss rate of CSFG from Table 3.

Figure 3: Cumulative distribution functions of $\mathrm{Var}(T(D))$ for human-written text and three unseen generators (Setting 2 test set). GPT-5 (red) lies to the left of human text (blue), indicating hyper-uniform generation with suppressed transition variance. Gemini 2.5 Flash Lite (green) is displaced far to the right, reflecting an amplified ROR signal. Claude Sonnet 4.6 (orange) occupies an intermediate position. The close overlap between GPT-5 and human CDFs (KS $D=0.2389$ ) directly explains the elevated FNR reported in Table 3.

Detector	Perturbation	Accuracy (%)	FPR (%)	FNR (%)
LASTDE	Paraphrasing	42.62	9.80	62.32
	Humanization	77.72	56.86	18.70
	Back-translation	54.14	23.53	48.17
Likelihood	Paraphrasing	60.70	33.33	39.92
	Humanization	46.41	21.57	56.91
	Back-translation	76.61	27.45	22.97

Appendix

Table 10: Detection performance under text perturbation (Setting 3) on Gemini 2.5 Flash Lite.

Detector	Perturbation	Accuracy (%)	FPR (%)	FNR (%)
LASTDE	Paraphrasing	43.51	22.45	60.00
	Humanization	75.62	60.00	20.73
	Back-translation	70.11	35.48	29.36
Likelihood	Paraphrasing	65.84	26.53	34.95
	Humanization	38.02	15.56	66.74
	Back-translation	65.92	16.13	35.78

Appendix

Table 11: Detection performance under text perturbation (Setting 3) on GPT-5.

Generator	$N$	Acc. (%)	FPR (%)	FNR (%)
Claude Sonnet 4.6	125	56.00	63.86	4.76
Gemini 2.5 Flash Lite	125	51.20	71.08	4.76
GPT-5	125	74.40	18.07	40.48

Appendix

Table 12: Detection performance on documents shorter than 500 characters. FPR increases sharply relative to the 0.66% observed on the full-length Setting 1 test set, while FNR remains low for Claude Sonnet 4.6 and Gemini 2.5 Flash Lite but increases substantially for GPT-5.

Dataset	$N$	Acc. (%)	FPR (%)	FNR (%)
XSum	100	85.00	25.00	6.00

Appendix

Table 13: CSFG performance on human-written news text (XSum). FPR is substantially elevated relative to the 1.57% observed on the balanced Setting 1 test set.

		SeqXGPT			CoCo			CSFG (Ours)
Generator	Perturbation	Acc	FPR	FNR	Acc	FPR	FNR	Acc	FPR	FNR
Claude Sonnet 4.6	Paraphrasing	70.48	9.80	31.57	91.55	84.75	1.67	90.59	88.24	1.22
	Humanization	71.90	29.41	27.96	89.22	81.54	4.76	87.43	84.31	5.10
	Back-translation	87.43	19.61	11.84	89.55	76.17	3.52	90.96	49.80	4.79
Gemini 2.5 Flash Lite	Paraphrasing	58.30	7.84	45.21	89.92	69.51	3.53	85.98	92.16	5.91
	Humanization	64.64	37.25	35.16	79.35	76.17	14.83	76.43	82.35	17.48

Appendix

Table 14: Cross-generator summary of Setting 3 robustness results (%). FPR is measured on perturbed human-written text. Detailed per-generator results are reported separately in Table 4 , 10 , and 11 .

Set.	Condition	Prompt
2	Generation	Please provide a continuation for the following content to make it coherent: { first 30 tokens }
3	Paraphrase	Paraphrase the following text: { Text }
	Humanization	Paraphrase the given sentence. Only return the paraphrased sentence in your response. Make it seem like a human wrote the article and that it is from the YOUR SECTION section of YOUR PUBLICATION. Sentence to paraphrase: YOUR SENTENCE Your paraphrase of the sentence: { Text }
	Back-translation	You are a professional translator. Translate the following English text into Chinese. Rules: (1) Preserve the original meaning and tone faithfully. (2) Do not add explanations or commentary. (3) Output only the translated Chinese text. Text: { input_text }
	Back-translation	You are a professional translator. Translate the following Chinese text back into English. Rules: (1) Translate naturally—do not attempt to recover the original wording. (2) Preserve meaning and tone. (3) Output only the translated English text. Text: { chinese_text }

Appendix

Table 15: Prompts used for text generation (Setting 2) and perturbation (Setting 3).

AI-Generated Text Detection

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

See all 171 papers

Rethinking AI-Generated Text Detection: A Strong Baseline and the Distribution-Shift Problem That Remains
Jul 4, 2026Zhuoer Shen, Mingyi Wang, Shaofeng Zou +1Distribution Shift Robustness AI-Generated Text Detection
Department of Computer Science, University of California, Santa Barbara · School of ECEE, Arizona State University
Breaking the Generator Barrier: Disentangled Representation for Generalizable AI-Text Detection
Apr 15, 2026Xiao Pu, Zepeng Cheng, Lin Yuan +2AI-Generated Text Detection Robustness of Machine-Generated Text Detection
Chongqing University of Posts and Telecommunications, China
EASE: Entropy-Adaptive Distribution Shaping for Evading AI-generated Text Detectors
Oct 7, 2026Jicheng Zhou, Kahim Wong, Jialong Wang +1Adversarial Text Attacks Robustness of Machine-Generated Text Detection
Department of Artificial Intelligence, Faculty of Information Science and Computing State Key Laboratory of Internet of Things for Smart City University of Macau, Macau, China

Jul 4, 2026J/K move · Enter open · S save

Rethinking AI-Generated Text Detection: A Strong Baseline and the Distribution-Shift Problem That Remains

Zhuoer Shen, Mingyi Wang, Shaofeng Zou, Yuheng Bu

Department of Computer Science, University of California, Santa Barbara · School of ECEE, Arizona State University

Distribution Shift Robustness AI-Generated Text Detection Domain Adaptation Distribution Shift Detection Robustness of Machine-Generated Text Detection

Open paper AskPDF arXiv

Recent AI-generated text detection work often introduces a new benchmark together with a specialized detector tailored to it. We revisit this practice from a baseline-first perspective. Across several benchmarks, we show that a plain, fully fine-tuned RoBERTa matches or exceeds the specialized detectors those benchmarks are built around. This suggests that much of the recent architectural complexity is not what drives strong in-distribution detection. The remaining challenge is the distribution shift. The same strong baseline degrades sharply when the topic domain or generating model changes at test time, and simply adding more source data does not close the gap. We identify a key failure mode: under distribution shift, the detector can assign high-confidence machine labels to human-written text from unseen domains. We then study two lightweight domain adaptation methods to address this problem:

K

-shot adaptation with first-order MAML over LoRA adapters, and a per-sample confidence-weighted ensemble built on top of the adapted detector. Overall, our results suggest that progress in AI-generated text detection should be measured not only by in-distribution performance, but also by robustness under distribution shift.

Abstract

AI-Generated Text Detection

Explore similar work

Rethinking AI-Generated Text Detection: A Strong Baseline and the Distribution-Shift Problem That Remains

Breaking the Generator Barrier: Disentangled Representation for Generalizable AI-Text Detection

EASE: Entropy-Adaptive Distribution Shaping for Evading AI-generated Text Detectors