EAGLE
Preference alignment to reduce hallucination in pathology VLMs.
- Cost
- Free
- Hardware
- Needs a consumer GPU
- Offline
- Works without a connection
- Scanner
- Not required
- Your data
- Stays on your machine
- Licence
- unknown
- Maturity
- research
What it is
Builds a large set of preferred and rejected response pairs and applies iterative preference optimisation, targeting multimodal hallucination and biased answers — a failure mode that matters more in pathology than in general vision tasks.
Licence notes
Check the repository and model card before any reuse.
Activity: 2 stars · last commit 2025-12-16 — refreshed nightly
Related
Catalogued 2026-08-02, last checked 2026-08-02. View the source record · Report an error