tum-nlp/neural-news-benchmark
收藏Hugging Face2025-02-05 更新2025-04-12 收录
下载链接：
https://hf-mirror.com/datasets/tum-nlp/neural-news-benchmark
下载链接
链接失效反馈官方服务：
资源简介：
neural-news是一个用于人类与AI新闻作者分类的基准数据集，包含英语、土耳其语、匈牙利语和波斯语四种语言。数据集分为训练、验证和测试集，并包含等量的机器生成和人类撰写的新闻文章。数据集可用于开发新闻作者分类器，评估零样本LLM和微调的多语言LLM。
neural-news is a benchmark dataset for human/AI news authorship classification, including four languages: English, Turkish, Hungarian, and Persian. The dataset is split into training, validation, and test sets, containing an equal amount of machine-generated and human-written news articles. It is used for developing news authorship classifiers and evaluating zero-shot LLMs and fine-tuned multilingual LLMs.
提供机构：
tum-nlp
