Citation Contexts Dataset
收藏arXiv2025-09-30 收录
下载链接：
https://github.com/lamps-lab/ccair-ai-reproducibility
下载链接
链接失效反馈官方服务：
资源简介：
该数据集包含了针对原始研究及其可重复性尝试的引用语境，这些语境被标注了其可重复性情感，包括正面、负面和中性情感。此外，数据集还包含了一个平衡的地面真相子集，其中69个标注的引用语境被手动分为三个情感类别。规模上，该数据集涵盖了来自130项原始研究的41,244个引用语境。该数据集的任务是对与论文可重复性相关的引用语境进行情感分析。
This dataset comprises citation contexts associated with original research studies and their reproducibility attempts, with the contexts annotated for their reproducibility-related sentiments covering positive, negative and neutral categories. Additionally, the dataset includes a balanced ground truth subset, in which 69 annotated citation contexts have been manually assigned to the three sentiment classes. Spanning 41,244 citation contexts from 130 original studies, this dataset is designed for sentiment analysis of citation contexts related to academic paper reproducibility.
提供机构：
Lamps Lab
