新闻文章数据集
收藏资源简介:
该数据集由340篇新闻文章组成,旨在评估意大利新闻媒体市场中新闻出版物的可靠性。每篇文章由两个人类专家和一个大型语言模型(LLM)进行标注,涵盖六个不同的可靠性标准,总计6,120个标注。数据集的创建过程包括选择和实施质量标准,收集代表性样本,并由专家和LLM进行评估。该数据集主要应用于新闻出版物可靠性评估,旨在通过自动化方法提高评估效率和准确性。
This dataset comprises 340 news articles developed to evaluate the reliability of news publications in the Italian news media market. Each article was annotated by two human experts and one large language model (LLM) across six distinct reliability criteria, resulting in a total of 6,120 annotations. The dataset creation process includes selecting and implementing quality standards, collecting representative samples, and conducting evaluations by experts and the LLM. This dataset is primarily applied to news publication reliability assessment, aiming to improve the efficiency and accuracy of evaluations via automated methods.

- 1Evaluation of Reliability Criteria for News Publishers with Large Language ModelsIIT-CNR, Pisa 和 IMT School for Advanced Studies Lucca · 2024年



