Researchers have developed a new framework to detect and categorize manipulative political narratives found on social media. The system first uses a few-shot prompt with a reasoning model to filter out manipulative posts from legitimate critiques. These filtered posts are then processed using UMAP for dimensionality reduction and HDBSCAN for unsupervised clustering to identify distinct narrative groups without predefined categories. Applied to over 1.2 million social media posts, this method successfully uncovered 41 unique manipulative narrative clusters. AI
Summary written by gemini-2.5-flash-lite from 1 source. How we write summaries →
IMPACT Provides a novel method for identifying and structuring disinformation campaigns, potentially aiding in content moderation and political analysis.
RANK_REASON Academic paper detailing a new computational framework for detecting manipulative political narratives. [lever_c_demoted from research: ic=1 ai=1.0]