NeuralNews

NeuralNews 是用于机器生成的新闻检测的数据集。它由人工生成的文章和机器生成的文章组成。人工生成的文章是从 GoodNews 数据集中提取的，该数据集是从纽约时报中提取的。它包含 4 种类型的文章：Real Articles 和 Real Captions Real Articles and Generated Captions Generated Articles 和 Real Captions Generated Articles 和 Generated Captions 总共包含每种文章类型的大约 32K 样本（总共大约 128K）。

NeuralNews is a dataset for machine-generated news detection. It consists of human-written articles and machine-generated articles. The human-written articles are extracted from the GoodNews dataset, which is sourced from The New York Times. The dataset includes four types of article-caption combinations: Real Articles and Real Captions, Real Articles and Generated Captions, Generated Articles and Real Captions, and Generated Articles and Generated Captions. Each type contains approximately 32,000 samples, resulting in a total of around 128,000 samples overall.

搜集汇总

背景与挑战

背景概述

NeuralNews是一个用于机器生成新闻检测的数据集，包含约128K篇人工和机器生成的文章样本，分为4种类型。该数据集由波士顿大学于2020年发布，旨在帮助识别神经假新闻。