YTClickbait21K: Human-Annotated Multimodal Dataset for YouTube Clickbait Detection Across Diverse Channels and Content Categories
Researchers have introduced YTClickbait21K, a new dataset designed to improve the automated detection of clickbait on YouTube. This dataset contains over 21,000 videos, annotated by multiple human labelers to ensure reliability, and includes multimodal data such as titles, descriptions, engagement statistics, and thumbnail images. The goal is to provide a robust benchmark for machine learning models in content moderation and cross-modal understanding. AI