mlsanityData to 14 Oct 2026

Mammo-SAE: Interpreting Breast Cancer Concept Learning with Sparse Autoencoders

arXiv 2507.15227 ↗First posted 21 Jul 2025cs.CVcited by 2

On the map: Mechanistic interpretability → Interpreting biomedical and science models

Abstract (arXiv)

Interpretability is critical in high-stakes domains such as medical imaging, where understanding model decisions is essential for clinical adoption. In this work, we introduce Sparse Autoencoder (SAE)-based interpretability to breast imaging by analyzing Mammo-CLIP, a vision--language foundation model pretrained on large-scale mammogram image--report pairs. We train a patch-level textttMammo-SAE on Mammo-CLIP to identify and probe latent features associated with clinically relevant breast concepts such as textitmass and textitsuspicious calcification. Our findings reveal that top activated class level latent neurons in the SAE latent space often tend to align with ground truth regions, and also uncover several confounding factors influencing the model's decision-making process. Additionally, we analyze which latent neurons the model relies on during downstream finetuning for improving the breast concept prediction. This study highlights the promise of interpretable SAE latent representations in providing deeper insight into the internal workings of foundation models at every layer for breast imaging. The code will be released at https://krishnakanthnakka.github.io/MammoSAE/

Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Placed in Interpreting biomedical and science models, one of 13 directions in Mechanistic interpretability. Builds on 16 of its 26 references that are in the map. First posted on arXiv 21 Jul 2025.

Builds on

16 of 26 references.

References · 16

  1. 2023LLaMA: Open and Efficient Foundation Language Modelscited by 14,132
  2. 2023Sparse Autoencoders Find Highly Interpretable Features in Language Modelscited by 1,420
  3. 2024Gemma: Open Models Based on Gemini Research and Technologycited by 916
  4. 2024Scaling and evaluating sparse autoencoderscited by 553
  5. 2024Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2cited by 394
  6. 2024Llama Scope: Extracting Millions of Features from Llama-3.1-8B with Sparse Autoencoderscited by 143
  7. 2024Discover-then-Name: Task-Agnostic Concept Bottlenecks via Automated Concept Discoverycited by 91
  8. 2024Towards Principled Evaluations of Sparse Autoencoders for Interpretability and Controlcited by 86
  9. 2022VinDr-Mammo: A large-scale benchmark dataset for computer-aided diagnosis in full-field digital mammographycited by 84
  10. 2024Sparse autoencoders reveal selective remapping of visual concepts during adaptationcited by 68
  11. 2024Mammo-CLIP: A Vision Language Foundation Model to Enhance Data Efficiency and Robustness in Mammographycited by 38
  12. 2024PURE: Turning Polysemantic Neurons Into Pure Features by Identifying Relevant Circuitscited by 29
4 more
  1. 2023Disentangling Neuron Representations with Concept Vectorscited by 28
  2. 2024Understanding polysemanticity in neural networks through coding theorycited by 17
  3. 2023Breast Cancer Classification with Enhanced Interpretability: DALAResNet50 and DT Grad-CAMcited by 4
  4. 2024Prototype-based Interpretable Breast Cancer Prediction Models: Analysis and Challengescited by 3

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 1; 2026: 1.

2 papers.

  1. 2025ConceptScope: Characterizing Dataset Bias via Disentangled Visual Conceptscited by 5
  2. 2026Mechanistic Interpretability of EEG Foundation Models via Sparse Autoencoderscited by 5

Similar, not cited

Close in wording, no citation either way

  1. 2025Mammo-CLIP Dissect: A Framework for Analysing Mammography Concepts in Vision-Language Models
  2. 2026Sparse Autoencoders for Interpretable Medical Image Representation Learning
  3. 2026MAM-CLIP: Vision-Language Pretraining on Mammography Atlases for BI-RADS Classification
  4. 2025MV-MLM: Bridging Multi-View Mammography and Language for Breast Cancer Diagnosis and Risk Prediction
  5. 2025Resurrecting the Salmon: Rethinking Mechanistic Interpretability with Domain-Specific Sparse Autoencoders
  6. 2022Knowledge Distillation to Ensemble Global and Interpretable Prototype-Based Mammogram Classification Models
  7. 2024Learning biologically relevant features in a pathology foundation model using sparse autoencoders
  8. 2025MedSAE: Dissecting MedCLIP Representations with Sparse Autoencoders