IntFE: Fine-grained Evaluation of Generative Models with Interpretable Sparse Encoders



This is the official repository for Interpretable Fine-grained Evaluation. We present our Interpretable Sparse Encoders for evaluation.


![image/png](https://cdn-uploads.huggingface.co/production/uploads/662161f7d7c21d69cc6599ca/U_FfmqALK5jzJr4wC7wFB.png)

---
license: mit
tags:
- evaluation
- image_generation
- mechanistic_interpretability
- physics_preservation
- semantic_loyalty
- realistic_fidelity
---