# EXPOSE: Explainable and Domain-Robust Embeddings from Pathology Vision Foundation Models using Sparse Autoencoders

arXiv: https://arxiv.org/abs/2608.28191 · HTML: https://mlsanity.com/paper/2608.28191/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Placed in Interpreting biomedical and science models, one of 13 directions in Mechanistic interpretability.
- Builds on 25 of its 41 references that are in the map.
- First posted on arXiv 28 Aug 2026.

## Placed in

- Mechanistic interpretability / [Interpreting biomedical and science models](https://mlsanity.com/mechinterp/interpreting-biomedical-and-science-models/)

## Builds on (most cited first, 25 of 25)

- [Sparse Autoencoders Find Highly Interpretable Features in Language Models](https://mlsanity.com/paper/2309.08600/) (2023)
- [Scaling and evaluating sparse autoencoders](https://mlsanity.com/paper/2406.04093/) (2024)
- [Domain Adaptation for Medical Image Analysis: A Survey](https://mlsanity.com/paper/2102.09508/) (2021)
- [Jumping Ahead: Improving Reconstruction Fidelity with JumpReLU Sparse Autoencoders](https://mlsanity.com/paper/2407.14435/) (2024)
- [Improving Dictionary Learning with Gated Sparse Autoencoders](https://mlsanity.com/paper/2404.16014/) (2024)
- [Virchow2: Scaling Self-Supervised Mixed Magnification Models in Pathology](https://mlsanity.com/paper/2408.00738/) (2024)
- [Sparse Autoencoders Learn Monosemantic Features in Vision-Language Models](https://mlsanity.com/paper/2504.02821/) (2025)
- [A Clinical Benchmark of Public Self-Supervised Pathology Foundation Models](https://mlsanity.com/paper/2407.06508/) (2024)
- [Hibou: A Family of Foundational Vision Transformers for Pathology](https://mlsanity.com/paper/2406.05074/) (2024)
- [Current Pathology Foundation Models are unrobust to Medical Center Differences](https://mlsanity.com/paper/2501.18055/) (2025)
- [Steering CLIP's vision transformer with sparse autoencoders](https://mlsanity.com/paper/2504.08729/) (2025)
- [Distilling foundation models for robust and efficient models in digital pathology](https://mlsanity.com/paper/2501.16239/) (2025)
- [SAE-SSV: Supervised Steering in Sparse Representation Spaces for Reliable Control of Language Models](https://mlsanity.com/paper/2505.16188/) (2025)
- [Do Multiple Instance Learning Models Transfer?](https://mlsanity.com/paper/2506.09022/) (2025)
- [Interpretable and Testable Vision Features via Sparse Autoencoders](https://mlsanity.com/paper/2502.06755/) (2025)
- [Foundation Models in Computational Pathology: A Review of Challenges, Opportunities, and Impact](https://mlsanity.com/paper/2502.08333/) (2025)
- [A Survey on Computational Pathology Foundation Models: Datasets, Adaptation Strategies, and Evaluation Tasks](https://mlsanity.com/paper/2501.15724/) (2025)
- [A Survey of Pathology Foundation Model: Progress and Future Directions](https://mlsanity.com/paper/2504.04045/) (2025)
- [Beyond Interpretability: The Gains of Feature Monosemanticity on Model Robustness](https://mlsanity.com/paper/2410.21331/) (2024)
- [CytoSAE: Interpretable Cell Embeddings for Hematology](https://mlsanity.com/paper/2507.12464/) (2025)
- [Pathology Foundation Models are Scanner Sensitive: Benchmark and Mitigation with Contrastive ScanGen Loss](https://mlsanity.com/paper/2507.22092/) (2025)
- [Analyzing Hierarchical Structure in Vision Models with Sparse Autoencoders](https://mlsanity.com/paper/2505.15970/) (2025)
- [Prototype-Based Multiple Instance Learning for Gigapixel Whole Slide Image Classification](https://mlsanity.com/paper/2503.08384/) (2025)
- [Beyond Interpretability: When, Why, and How Sparse Autoencoders Enable Label-Free Visual Steering](https://mlsanity.com/paper/2506.01247/) (2025)
- [Towards Explainable End-to-End Prostate Cancer Relapse Prediction from H&E Images Combining Self-Attention Multiple Instance Learning with a Recurrent Neural Network](https://mlsanity.com/paper/2111.13439/) (2021)

## Abstract (arXiv)

Vision Foundation Models (VFMs) are widely used in computational pathology but remain sensitive to domain shifts arising from variations in staining, tissue preparation, and scanner hardware. A key limitation is that VFM embeddings entangle biological with domain-specific information, hindering cross-domain generalization. We propose Explainable Probing of Cross-Domain Sparse Embeddings (EXPOSE), a framework that uses Sparse Autoencoders (SAEs) as an explainable bottleneck to identify and suppress domain-specific components in VFM embeddings. We train a sparse representation of VFM features, use a linear classifier to identify domain-specific latent dimensions, and mask these features prior to downstream relapse prediction without retraining the backbone model. Experiments on a large prostate cancer dataset with multiple acquisition domains show that SAE features capture both domain- and task-specific information, which are partially disentangled in the latent space. Removing domain-specific features improves cross-domain performance and increases embedding robustness as measured by the Domain Robustness Index (DoRI). Code is available at https://github.com/imsb-uke/expose .