# Sparse autoencoders reveal selective remapping of visual concepts during adaptation

arXiv: https://arxiv.org/abs/2412.05276 · HTML: https://mlsanity.com/paper/2412.05276/

## Key facts

- Cited by 68 papers in the map (data through 14 Oct 2026).
- Placed in Interpreting vision and multimodal models, one of 13 directions in Mechanistic interpretability.
- Builds on 18 of its 37 references that are in the map.
- First posted on arXiv 6 Dec 2024.

## Placed in

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

## Builds on (most cited first, 18 of 18)

- [Learning Transferable Visual Models From Natural Language Supervision](https://mlsanity.com/paper/2103.00020/) (2021)
- [High-Resolution Image Synthesis with Latent Diffusion Models](https://mlsanity.com/paper/2112.10752/) (2021)
- [LLaMA: Open and Efficient Foundation Language Models](https://mlsanity.com/paper/2302.13971/) (2023)
- [Emerging Properties in Self-Supervised Vision Transformers](https://mlsanity.com/paper/2104.14294/) (2021)
- [BLIP-2: Bootstrapping Language-Image Pre-training with Frozen Image Encoders and Large Language Models](https://mlsanity.com/paper/2301.12597/) (2023)
- [Improved Baselines with Visual Instruction Tuning](https://mlsanity.com/paper/2310.03744/) (2023)
- [Learning to Prompt for Vision-Language Models](https://mlsanity.com/paper/2109.01134/) (2021)
- [Fine-Grained Visual Classification of Aircraft](https://mlsanity.com/paper/1306.5151/) (2013)
- [EuroSAT: A Novel Dataset and Deep Learning Benchmark for Land Use and Land Cover Classification](https://mlsanity.com/paper/1709.00029/) (2017)
- [Conditional Prompt Learning for Vision-Language Models](https://mlsanity.com/paper/2203.05557/) (2022)
- [Sparse Autoencoders Find Highly Interpretable Features in Language Models](https://mlsanity.com/paper/2309.08600/) (2023)
- [MaPLe: Multi-modal Prompt Learning](https://mlsanity.com/paper/2210.03117/) (2022)
- [Gemma Scope: Open Sparse Autoencoders Everywhere All At Once on Gemma 2](https://mlsanity.com/paper/2408.05147/) (2024)
- [Exploring Visual Prompts for Adapting Large-Scale Models](https://mlsanity.com/paper/2203.17274/) (2022)
- [Self-regulating Prompts: Foundational Model Adaptation without Forgetting](https://mlsanity.com/paper/2307.06948/) (2023)
- [Improving Dictionary Learning with Gated Sparse Autoencoders](https://mlsanity.com/paper/2404.16014/) (2024)
- [Transformer visualization via dictionary learning: contextualized embedding as a linear superposition of transformer factors](https://mlsanity.com/paper/2103.15949/) (2021)
- [Discover-then-Name: Task-Agnostic Concept Bottlenecks via Automated Concept Discovery](https://mlsanity.com/paper/2407.14499/) (2024)

## Built on by (most cited first, 30 of 68)

- [Sparse Autoencoders Learn Monosemantic Features in Vision-Language Models](https://mlsanity.com/paper/2504.02821/) (2025)
- [Steering CLIP's vision transformer with sparse autoencoders](https://mlsanity.com/paper/2504.08729/) (2025)
- [A Survey on Mechanistic Interpretability for Multi-Modal Foundation Models](https://mlsanity.com/paper/2502.17516/) (2025)
- [Interpretable and Testable Vision Features via Sparse Autoencoders](https://mlsanity.com/paper/2502.06755/) (2025)
- [Prisma: An Open Source Toolkit for Mechanistic Interpretability in Vision and Video](https://mlsanity.com/paper/2504.19475/) (2025)
- [From superposition to sparse codes: interpretable representations in neural networks](https://mlsanity.com/paper/2503.01824/) (2025)
- [On the Limits of Sparse Autoencoders: A Theoretical Framework and Reweighted Remedy](https://mlsanity.com/paper/2506.15963/) (2025)
- [DNA: Uncovering Universal Latent Forgery Knowledge](https://mlsanity.com/paper/2601.22515/) (2026)
- [VL-SAE: Interpreting and Enhancing Vision-Language Alignment with a Unified Concept Set](https://mlsanity.com/paper/2510.21323/) (2025)
- [CytoSAE: Interpretable Cell Embeddings for Hematology](https://mlsanity.com/paper/2507.12464/) (2025)
- [Adversarial Attacks Leverage Interference Between Features in Superposition](https://mlsanity.com/paper/2510.11709/) (2025)
- [AudioSAE: Towards Understanding of Audio-Processing Models with Sparse AutoEncoders](https://mlsanity.com/paper/2602.05027/) (2026)
- [From What to How: Attributing CLIP's Latent Components Reveals Unexpected Semantic Reliance](https://mlsanity.com/paper/2505.20229/) (2025)
- [Explaining Similarity in Vision-Language Encoders with Weighted Banzhaf Interactions](https://mlsanity.com/paper/2508.05430/) (2025)
- [Causal Interpretation of Sparse Autoencoder Features in Vision](https://mlsanity.com/paper/2509.00749/) (2025)
- [Specialization after Generalization: Towards Understanding Test-Time Training in Foundation Models](https://mlsanity.com/paper/2509.24510/) (2025)
- [ConceptScope: Characterizing Dataset Bias via Disentangled Visual Concepts](https://mlsanity.com/paper/2510.26186/) (2025)
- [SAVE: Sparse Autoencoder-Driven Visual Information Enhancement for Mitigating Object Hallucination](https://mlsanity.com/paper/2512.07730/) (2025)
- [Probing Cross-modal Information Hubs in Audio-Visual LLMs](https://mlsanity.com/paper/2605.10815/) (2026)
- [Symbolic Rule Extraction from Attention-Guided Sparse Representations in Vision Transformers](https://mlsanity.com/paper/2505.06745/) (2025)
- [Latent Anomaly Knowledge Excavation: Unveiling Sparse Sensitive Neurons in Vision-Language Models](https://mlsanity.com/paper/2604.07802/) (2026)
- [Inside the Visual Mind: Neuroscience-Motivated Concept Circuits for Interpreting and Steering Vision Transformers](https://mlsanity.com/paper/2606.06664/) (2026)
- [Measuring the (Un)Faithfulness of Concept-Based Explanations](https://mlsanity.com/paper/2504.10833/) (2025)
- [Beyond Interpretability: When, Why, and How Sparse Autoencoders Enable Label-Free Visual Steering](https://mlsanity.com/paper/2506.01247/) (2025)
- [Mammo-SAE: Interpreting Breast Cancer Concept Learning with Sparse Autoencoders](https://mlsanity.com/paper/2507.15227/) (2025)
- [Sparse CLIP: Co-Optimizing Interpretability and Performance in Contrastive Learning](https://mlsanity.com/paper/2601.20075/) (2026)
- [Language Models Can Explain Visual Features via Steering](https://mlsanity.com/paper/2603.22593/) (2026)
- [Improving Sparse Autoencoder with Dynamic Attention](https://mlsanity.com/paper/2604.14925/) (2026)
- [Sparse Autoencoders enable Robust and Interpretable Fine-tuning of CLIP models](https://mlsanity.com/paper/2605.15961/) (2026)
- [Whisper Hallucination Detection and Mitigation via Hidden Representation Steering and Sparse AutoEncoders](https://mlsanity.com/paper/2606.07473/) (2026)

## Abstract (arXiv)

Adapting foundation models for specific purposes has become a standard approach to build machine learning systems for downstream applications. Yet, it is an open question which mechanisms take place during adaptation. Here we develop a new Sparse Autoencoder (SAE) for the CLIP vision transformer, named PatchSAE, to extract interpretable concepts at granular levels (e.g., shape, color, or semantics of an object) and their patch-wise spatial attributions. We explore how these concepts influence the model output in downstream image classification tasks and investigate how recent state-of-the-art prompt-based adaptation techniques change the association of model inputs to these concepts. While activations of concepts slightly change between adapted and non-adapted models, we find that the majority of gains on common adaptation tasks can be explained with the existing concepts already present in the non-adapted foundation model. This work provides a concrete framework to train and use SAEs for Vision Transformers and provides insights into explaining adaptation mechanisms.