Detecting LLM-Generated Tokens in Human--LLM Coauthored Text

Source

• Source name: arXiv
• Source kind: paper
• Authors / organization: Yangjun Lu, Hongyi Zhou, Fabian Spill, Kai Ye, Chengchun Shi, Jin Zhu
• Venue: arXiv: cs.AI
• Published: 2026-07-23
• Collected: 2026-07-24
• URL: http://arxiv.org/abs/2607.21458v1
• PDF: https://arxiv.org/pdf/2607.21458v1
• DOI: UNKNOWN

Plain-Language Brief

Researchers present a new method for detecting parts of text generated by language models (LLMs) in human-AI coauthored documents. The approach works at the token level, using existing scores and a smoothing technique to improve accuracy without requiring labeled data.

What the Document Says

• The paper discusses a method for detecting LLM-generated tokens in mixed-authority documents.
• The method operates at the token level.
• It utilizes existing token-level detection scores and smooths adjacent token scores to reduce variability.
• An adaptive Lepski-type rule is used to select the bandwidth according to local authorship structure.
• The approach does not require token-level labeled data for training.
• Theoretically, it achieves favorable mean square error performance in estimating the underlying signal.
• Empirical testing demonstrates strong performance against various baselines in both synthetic and realistic datasets.

Key Concepts Mentioned

• human-AI collaborative writing
• fine-grained detection methods
• LLM-generated content
• token-level
• document classification
• Lepski-type rule
• mean square error

Claims Made by the Document

• The proposed method addresses the need for fine-grained detection of LLM-generated content in mixed-authorship documents.
• It achieves favorable mean square error performance.

Evidence Used

• The paper presents a new method to solve the problem of identifying LLM-generated parts in human-AI coauthored texts.
• Theoretical and empirical evidence support the claim of favorable mean square error performance.

Methods / Approach

• smoothing adjacent token scores
• adaptive Lepski-type rule

Limitations Stated by the Authors

UNKNOWN

Limitations Noted by the Agent

Agent interpretation — not from the source.

UNKNOWN

Open Questions

UNKNOWN

Notes for Future Q&A

May include interpretation. Labeled where it does.

UNKNOWN

Raw Abstract / Excerpt

The rise of human-AI collaborative writing has created a growing need for fine-grained detection methods that support localizing likely LLM-generated content in mixed-authorship documents. Existing methods for detecting LLM-generated text mainly focus on document-level classification and cannot identify which parts of the text are generated by LLMs. This paper introduces a new method to address this urgent need. Our method operates at the token level, the natural unit of modern language models, and builds on existing token-level detection scores. The key idea is to smooth adjacent token scores to reduce their variability, while using an adaptive Lepski-type rule to select the bandwidth according to the local authorship structure. Our method is simple to implement and does not require token-level labeled data for training. Theoretically, we characterize this trade-off and show that the proposed method achieves favorable mean square error performance in estimating the underlying signal. Empirically, we demonstrate strong performance of our method against a wide range of baselines in both synthetic datasets and a realistic dataset. We deploy a publicly accessible website that implements the methods as well.