Overview of the NLPCC 2025 Shared Task 1: LLM-Generated Text Detection

Download PDF
Open Webpage

Junchao Wu, Runzhe Zhan, Qianli Wang, Yulin Yuan, Lidia S. Chao, Derek F. Wong

Published: 2025, Last Modified: 12 Jan 2026NLPCC (4) 2025EveryoneRevisionsBibTeXCC BY-SA 4.0
Abstract: This paper introduces the achievements and findings of the shared task on Large Language Model (LLM) Generated Text Detection, organized as part of the 14th China National Conference on Natural Language Processing and Chinese Computing (NLPCC 2025). The primary objective of this shared task is to utilize machine learning techniques to develop detectors that can effectively distinguish between LLM-generated text and human-written text, thereby addressing the challenges posed by the rapid advancements of LLMs. The task garnered significant attention, attracting over 30 participating teams from both academia and industry, and 22 teams successfully submitted official results. In this paper, we provide a comprehensive overview of the task, including the dataset, task design, evaluation results, and an in-depth analysis of the submitted solutions. Furthermore, we introduce DetectRL-ZH, a newly released detection dataset, which aims to advance research in detecting LLM-generated Chinese texts (Shared Task Official Website: https://github.com/NLP2CT/NLPCC-2025-Task1).