Wang-ML-Lab/interpretable-foundation-models

[ICML 2024] Probabilistic Conceptual Explainers (PACE): Trustworthy Conceptual Explanations for Vision Foundation Models

31
/ 100
Emerging

This project helps AI researchers and practitioners understand why a vision AI model makes certain predictions. It takes a trained vision transformer model and a dataset of images, then outputs automatically discovered visual concepts (like 'petals' or 'sky') that explain the model's behavior at a dataset, image, and even specific image-patch level. This is for AI developers, machine learning engineers, and researchers who need to debug or validate their vision AI models.

No commits in the last 6 months.

Use this if you need to gain trustworthy insights into how your vision transformer models 'think' by identifying the underlying visual concepts driving their decisions.

Not ideal if you are looking to interpret traditional machine learning models or want explanations for language-based AI models, as this is specifically for vision transformers.

AI-explainability computer-vision model-debugging deep-learning-interpretation machine-learning-research
No License Stale 6m No Package No Dependents
Maintenance 2 / 25
Adoption 6 / 25
Maturity 8 / 25
Community 15 / 25

How are scores calculated?

Stars

18

Forks

4

Language

Python

License

Last pushed

Sep 25, 2025

Commits (30d)

0

Get this data via API

curl "https://pt-edge.onrender.com/api/v1/quality/llm-tools/Wang-ML-Lab/interpretable-foundation-models"

Open to everyone — 100 requests/day, no key needed. Get a free key for 1,000/day.