Continuous CAS-CS Dataset
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/gyc-nii/CAS-CS-and-dual-head-detector
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含了55,000篇在计算机科学领域由人类与AI协作生成的内容，这些内容反映了从0到1不同级别的人类参与度。此外，为了进行泛化能力测试，数据集还包含了1,000篇来自不同AI系统生成的内容。该数据集被用于评估基于人工智能生成内容中人类参与度的回归模型的有效性。规模上，数据集包含了55,000篇文章以及1,000篇来自各异AI系统生成的内容，其任务是衡量人类参与的程度。
This dataset contains 55,000 articles co-generated by humans and AI in the computer science domain, which cover varying levels of human involvement ranging from 0 to 1. To facilitate generalization capability testing, the dataset also includes 1,000 pieces of content generated by various AI systems. This dataset is designed to evaluate the effectiveness of regression models for assessing human involvement in AI-generated content, with its core task being to measure the degree of human participation.
