evaluation_0405
收藏资源简介:
该数据集包含多个配置版本,主要收录与学术论文相关的数据及其多维评估结果。每个数据条目包含论文ID、标题、生成输出内容以及包括相似度、新颖性、可行性等在内的12项评估指标。数据集分为ICLR_2026_oral子集,不同配置的样本量从222到2886不等,数据大小从1.4MB到55.6MB。适用于自然语言生成质量评估、学术文本分析等研究场景。
This dataset includes multiple configuration variants, primarily curating data related to academic papers and their multi-dimensional evaluation results. Each data entry contains paper ID, title, generated output content, alongside 12 evaluation metrics such as similarity, novelty, feasibility and others. The dataset is partitioned into the ICLR_2026_oral subset, where the sample sizes of different configurations range from 222 to 2886, and the data sizes vary between 1.4 MB and 55.6 MB. It is applicable to research scenarios including natural language generation quality assessment and academic text analysis.




