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

TL;DR

This paper introduces a structured framework for AI-text detection that disentangles semantic content from generator artifacts, improving robustness and generalization across diverse language models.

Contribution

The proposed method advances AI-text detection by effectively separating semantic information from generator-specific artifacts, enhancing performance on unseen models.

Findings

Achieves up to 24.2% accuracy gain over state-of-the-art methods.

Demonstrates strong scalability with increasing generator diversity.

Maintains robust performance in open-set scenarios.

Abstract

As large language models (LLMs) generate text that increasingly resembles human writing, the subtle cues that distinguish AI-generated content from human-written content become increasingly challenging to capture. Reliance on generator-specific artifacts is inherently unstable, since new models emerge rapidly and reduce the robustness of such shortcuts. This generalizes unseen generators as a central and challenging problem for AI-text detection. To tackle this challenge, we propose a progressively structured framework that disentangles AI-detection semantics from generator-aware artifacts. This is achieved through a compact latent encoding that encourages semantic minimality, followed by perturbation-based regularization to reduce residual entanglement, and finally a discriminative adaptation stage that aligns representations with task objectives. Experiments on MAGE benchmark, covering 20 representative LLMs across 7 categories, demonstrate consistent improvements over state-of-the-art methods, achieving up to 24.2% accuracy gain and 26.2% F1 improvement. Notably, performance continues to improve as the diversity of training generators increases, confirming strong scalability and generalization in open-set scenarios. Our source code will be publicly available at https://github.com/PuXiao06/DRGD.