| # Acknowledgements |
|
|
| This package is based on the research and original implementation by Cywinski & Deja. All credit for the underlying methodology — concept unlearning in diffusion models using sparse autoencoders — belongs to the original authors. |
|
|
| ## Original Paper |
|
|
| **SAeUron: Interpretable Concept Unlearning in Diffusion Models with Sparse Autoencoders** |
| Bartosz Cywinski, Kamil Deja |
|
|
| - arXiv: https://arxiv.org/abs/2501.18783 |
| - Original repository: https://github.com/bcywinski/saeuron |
|
|
| ## Citation |
|
|
| If you use this package in your research, please cite the original work: |
|
|
| ```bibtex |
| @article{cywinski2025saeuron, |
| title={SAeUron: Interpretable Concept Unlearning in Diffusion Models with Sparse Autoencoders}, |
| author={Cywinski, Bartosz and Deja, Kamil}, |
| journal={arXiv preprint arXiv:2501.18783}, |
| year={2025} |
| } |
| ``` |
|
|
