arXiv Computation and Language August 28, 2026 By Xin Chen, Fuwei Zhang, Yiqi Tong, Wei Guo, Yutian Xiao, Fuzhen Zhuang

D2C-Routing: Dimension-to-Composition Evidence Routing for Mixed-Origin AI-Generated Text Detection

Read the original on arXiv Computation and Language →

The paper introduces D2C-Routing, a method for detecting mixed-origin AI-generated text by separately routing content-side and expression-side evidence to supervised dimension heads before combining them with a gated composition layer. On the MixD2C benchmark, the system achieves a four-way average true positive rate of 0.8603 at 1% false positive rate, outperforming a baseline by 6.5 points. Ablation studies confirm the effectiveness of the routing design, and error analysis highlights the difficulty of distinguishing AI-content with human expression from fully AI-generated text.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Computation and Language.

benchmarks

arXiv Machine Learning

2d ago

Steer-to-Detect: Probing Hidden Representations for Detection of LLM-Generated Texts

arXiv:2605.12890v2 Announce Type: replace-cross Abstract: The rapid advancement of large language models (LLMs) has made machine-generated text increasingly difficult to distinguish from human-writte...

By Luxu Liang, Xiang Li

llms safety

arXiv AI

Jun 16

StyleShield: Exposing the Fragility of AIGC Detectors through Continuous Controllable Style Transfer

arXiv:2605. 00924v2 Announce Type: replace-cross Abstract: AI-generated content (AIGC) detectors are increasingly deployed in high-stakes settings such as academic integrity screening, yet their reliability rests on a fundamental paradox: as language models are trained on human-written corpora, the statistical boundary between AI and human writing will inevitably dissolve as models improve.

By Guantian Zheng

llms rag diffusion benchmarks

arXiv Computation and Language

Sep 15

SlopShape: Identifying AI-Generated Commercial Web Content

arXiv:2609.15369v1 Announce Type: new Abstract: Word-level detectors identify unedited AI-generated text almost perfectly, but the literature documents their brittleness under rewording, and a word-l...

By Jochen Madler (Sitefire)

llms

arXiv Machine Learning

Aug 4

Learning Compositional Meta-Routing for Agentic Workflows: An Executable Benchmark

arXiv:2608. 00106v1 Announce Type: new Abstract: Agentic systems must decide not only what answer to produce, but which reasoning and execution operations should precede it.

By Natan Vidra, Alina Kapanova, Arun Kanhai, Spurthi Setty

llms agents benchmarks

arXiv AI

Jun 4

'Your AI Text is not Mine': Redefining and Evaluating AI-generated Text Detection under Realistic Assumptions

arXiv:2606. 04906v1 Announce Type: cross Abstract: Although it is generally agreed that AI-generated text poses a broad societal risk, there is no common understanding in the AI-generated text detection literature on what constitutes harmful use.

By Nils Dycke, Marina Sakharova, Nico Daheim, Iryna Gurevych

benchmarks

arXiv AI

Jun 8

SV-Detect: AI-generated Text Detection with Steering Vectors

arXiv:2606. 07313v1 Announce Type: cross Abstract: Detecting machine-generated text is especially difficult under distribution shift, such as transfer across domains, source models, and editing attacks.

By Mikhail Vishnyakov, Tatiana Gaintseva

llms safety

One email a morning, machine-written

Related stories

Steer-to-Detect: Probing Hidden Representations for Detection of LLM-Generated Texts

StyleShield: Exposing the Fragility of AIGC Detectors through Continuous Controllable Style Transfer

SlopShape: Identifying AI-Generated Commercial Web Content

Learning Compositional Meta-Routing for Agentic Workflows: An Executable Benchmark

'Your AI Text is not Mine': Redefining and Evaluating AI-generated Text Detection under Realistic Assumptions

SV-Detect: AI-generated Text Detection with Steering Vectors