CUDRT
收藏资源简介:
CUDRT数据集是由中国人民大学信息学院等机构创建,旨在评估人工智能生成文本检测器的性能。该数据集包含中英文两种语言,涵盖新闻文章和学术论文,总计约40000条数据。创建过程中,使用了多种主流大型语言模型生成文本,以模拟真实世界中的文本生成场景。CUDRT数据集的应用领域主要集中在信息安全、版权保护和伦理问题,特别是解决人工智能生成文本与人类创作文本难以区分的问题。
The CUDRT dataset was developed by the School of Information, Renmin University of China and other institutions, with the aim of evaluating the performance of AI-generated text detectors. It contains texts in both Chinese and English, spanning news articles and academic papers, with a total of approximately 40,000 instances. During its development, multiple mainstream large language models were utilized to generate texts, simulating real-world text generation scenarios. The CUDRT dataset is primarily applied in the fields of information security, copyright protection and ethical issues, particularly addressing the challenge of distinguishing between AI-generated texts and human-authored works.

- 1CUDRT: Benchmarking the Detection of Human vs. Large Language Models Generated Texts中国人民大学信息学院 · 2024年



