Abstract
A computational framework combining prompt-based filtering and unsupervised clustering identifies manipulative political narrative clusters from social media posts without requiring predefined categories.
We present a new computational framework for detecting and structuring manipulative political narratives. A task that became more important due to the shift of political discussions to social media. One of the primary challenges thereby is differentiating between manipulative political narratives and legitimate critiques. Some posts may also reframe actual events within a manipulative context. To achieve good clustering results, we filter manipulative posts beforehand using a detailed few-shot prompt that combines documented campaign narratives with legitimate criticisms to differentiate them. This prompt enables a reasoning model to assign labels, retaining only manipulative narrative posts for further processing. The remaining posts are subsequently embedded and dimensionality-reduced using UMAP, before HDBSCAN is applied to uncover narrative groups. A key advantage of this unsupervised approach is its independence from a predefined list of target categories, enabling it to uncover new narrative clusters. Finally, a reasoning model is employed to uncover the narrative behind each cluster. This approach, applied to over 1.2 million social media posts, effectively identified 41 distinct manipulative narrative clusters by integrating prompt-based filtering with unsupervised clustering.
Community
We present a new computational framework for detecting and structuring manipulative political narratives. A task that became more important due to the shift of political discussions to social media. One of the primary challenges thereby is differentiating between manipulative political narratives and legitimate critiques. Some posts may also reframe actual events within a manipulative context.
To achieve good clustering results, we filter manipulative posts beforehand using a detailed few-shot prompt that combines documented campaign narratives with legitimate criticisms to differentiate them. This prompt enables a reasoning model to assign labels, retaining only manipulative narrative posts for further processing.
The remaining posts are subsequently embedded and dimensionality-reduced using UMAP, before HDBSCAN is applied to uncover narrative groups. A key advantage of this unsupervised approach is its independence from a predefined list of target categories, enabling it to uncover new narrative clusters.
Finally, a reasoning model is employed to uncover the narrative behind each cluster. This approach, applied to over 1.2 million social media posts, effectively identified 41 distinct manipulative narrative clusters by integrating prompt-based filtering with unsupervised clustering.
Hi everyone,
To make it as easy as possible to reproduce the results from our paper and build upon our results, we uploaded our datasets to HuggingFace. You'll find them at the "Datasets citing this paper" section.
This is an automated message from the Librarian Bot. I found the following papers similar to this paper.
The following papers were recommended by the Semantic Scholar API
- A Structured Clustering Approach for Inducing Media Narratives (2026)
- Reasoning-Based Refinement of Unsupervised Text Clusters with LLMs (2026)
- Mapping the Political Discourse in the Brazilian Chamber of Deputies: A Multi-Faceted Computational Approach (2026)
- A Community-Based Approach for Stance Distribution and Argument Organization (2026)
- Ideology Prediction of German Political Texts (2026)
- Fine-tuning with Hierarchical Prompting for Robust Propaganda Classification Across Annotation Schemas (2026)
- Retrieving Climate Change Disinformation by Narrative (2026)
Please give a thumbs up to this comment if you found it helpful!
If you want recommendations for any Paper on Hugging Face checkout this Space
You can directly ask Librarian Bot for paper recommendations by tagging it in a comment: @librarian-bot recommend
Get this paper in your agent:
hf papers read 2605.14354 Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash Models citing this paper 0
No model linking this paper
Datasets citing this paper 3
SinclairSchneider/tweets_about_german_politicians_jan_feb_2025_reddit_telegram_classified_embed_reduced_clustered
Spaces citing this paper 0
No Space linking this paper
Collections including this paper 0
No Collection including this paper