SummEval

该数据集包含了基于CNN/Dailymail新闻文章数据集的人工评估模型生成摘要的判断。这些评估由专家评委和众包工作者共同完成，为确保更高的可靠性，仅使用了专家的标注。该数据集的任务是对摘要质量进行评价。

This dataset comprises judgments of model-generated summaries based on the CNN/DailyMail news article dataset. While these evaluations were jointly conducted by expert judges and crowd workers, only expert annotations were retained to ensure higher reliability. The core task of this dataset is to evaluate the quality of text summaries.