BriefGPT.xyz
Aug, 2024
学习重写:通用的LLM生成文本检测
Learning to Rewrite: Generalized LLM-Generated Text Detection
HTML
PDF
Wei Hao, Ran Li, Weiliang Zhao, Junfeng Yang, Chengzhi Mao
TL;DR
本研究解决了当前分类器在开放世界中检测LLM生成内容的能力不足问题。通过训练LLM进行文本重写,该方法能够有效区分LLM和人类创作的文本,从而在各领域中实现可推广的编辑距离差异。实验表明,该分类器在检测准确性上显著优于目前最先进的无监督分类器,具有重要的实际应用潜力。
Abstract
Large Language Models
(LLMs) can be abused at scale to create non-factual content and spread
Disinformation
. Detecting LLM-generated content is essential to mitigate these risks, but current classifiers often fai
→