arxivcs.AIstat.ME2026-07-23
Detecting LLM-Generated Tokens in Human--LLM Coauthored Text
Yangjun Lu, Hongyi Zhou, Fabian Spill, Kai Ye, Chengchun Shi, Jin Zhu
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 classifica…