{"slug": "stability-aware-feature-design-for-robust-watermark-detection-in-machine-text", "title": "Stability-Aware Feature Design for Robust Watermark Detection in Machine-Generated Text", "summary": "Researchers introduced Pattern Stability Score (PSS), a watermark detection framework for machine-generated text that improves detection AUC by over 10-15 percentage points across different token lengths compared to prior baselines. The method, evaluated on PG-19, CNN/DailyMail, and WikiText using Llama-3-8B and Qwen2-7B, maintains above 87.8% AUC even when all components differ from training.", "body_md": "arXiv:2608.18102v1 Announce Type: new\nAbstract: The widespread adoption of large language models (LLMs) has intensified the demand for principled methods to distinguish human from machine-generated text. Watermarking provides a promising avenue, yet existing detectors exhibit sharp performance deterioration under multiple paraphrasing and when applied to shorter texts. We introduce Pattern Stability Score (PSS), a novel detection framework that leverages local statistical features and stability dynamics across paraphrased variants. Specifically, the proposed method combines global and local z-score features with higher-order statistics of run-length patterns, enriched by autocorrelation signals and stability scores computed over paraphrase depth. Numerical evaluations are performed on three benchmark datasets (PG-19, CNN/DailyMail, and WikiText) using multiple LLMs (Llama-3-8B, Qwen2-7B) and paraphrasers (Mistral-7B, Qwen2-7B, Gemma-7B), systematically stress-testing robustness under up to eight rounds of paraphrasing. Compared to prior z-score thresholding baselines and some state-of-the-art deep learning methods, our approach improves detection AUC (area under the receiver operating characteristic curve) by over 10-15 percentage points across different token lengths. Additionally, extensive cross-domain experiments demonstrate that a single universal classifier generalizes across different LLMs, paraphrasers, and text domains without retraining, maintaining above 87.8% AUC even when all components differ from training.", "url": "https://wpnews.pro/news/stability-aware-feature-design-for-robust-watermark-detection-in-machine-text", "canonical_source": "https://arxiv.org/abs/2608.18102", "published_at": "2026-08-20 04:00:00+00:00", "updated_at": "2026-08-20 04:13:04.985341+00:00", "lang": "en", "topics": ["artificial-intelligence", "machine-learning", "large-language-models", "ai-research"], "entities": ["Pattern Stability Score", "Llama-3-8B", "Qwen2-7B", "Mistral-7B", "Gemma-7B", "PG-19", "CNN/DailyMail", "WikiText"], "alternates": {"html": "https://wpnews.pro/news/stability-aware-feature-design-for-robust-watermark-detection-in-machine-text", "markdown": "https://wpnews.pro/news/stability-aware-feature-design-for-robust-watermark-detection-in-machine-text.md", "text": "https://wpnews.pro/news/stability-aware-feature-design-for-robust-watermark-detection-in-machine-text.txt", "jsonld": "https://wpnews.pro/news/stability-aware-feature-design-for-robust-watermark-detection-in-machine-text.jsonld"}}