Sparse Autoencoders Reveal Interpretable and Steerable Features in VLA Models

2026-03-19 · Aiden Swann, Lachlain McGranahan, Hugo Buurmeijer, Monroe Kennedy, Mac Schwager arxiv

Vision-Language-Action (VLA) models have emerged as a promising approach for general-purpose robot manipulation. However, little research has mechanistically explored when and why they generalize across objects, scenes, and instructions. To probe internal representations, we train Sparse Autoencoders (SAEs) on the VLA's hidden-layer activations. SAEs learn sparse dictionaries over model activations, often revealing features that correspond to interpretable directions in the model's representation space. We identify SAE features corresponding to motion primitives and semantic concepts, including features that are general across episodes and causally steerable. We propose a metric to categorize features as general transferable primitives or episode-specific memorizations, offering a promising glimpse towards VLA generalization. We validate these findings through steering experiments on both the LIBERO simulation benchmark and on real-world DROID hardware. We find that amplifying general and semantic features induces behaviors consistent with their meanings, whereas ablating them destroys model performance. Furthermore, we demonstrate steering as a way to control behavior in unpromptable directions. Together, these results provide mechanistic evidence that VLAs can learn reusable internal features linking perception, language, and action across tasks and scenes. Our project page is located at https://drvla.github.io

📄 PDF Abstract BibTeX arXiv:2603.19183

Code (0)

등록된 구현이 없습니다.

Tasks

Robot Manipulation

Interpretable and Steerable Concept Bottleneck Sparse Autoencoders

2025-12-11 · Akshay Kulkarni, Tsui-Wei Weng, Vivek Narayanaswamy, Shusen Liu 외 arxiv

Sparse autoencoders (SAEs) promise a unified approach for mechanistic interpretability, concept discovery, and model steering in LLMs and LVLMs. However, realizing this potential requires learned features to be both inte…

Image Generation

Mechanistic Interpretability of Antibody Language Models Using SAEs

2025-12-05 · Rebonto Haque, Oliver M. Turnbull, Anisha Parsan, Nithin Parsan 외 arxiv

Sparse autoencoders (SAEs) are a mechanistic interpretability technique that have been used to provide insight into learned concepts within large protein language models. Here, we employ TopK and Ordered SAEs to investig…

Mechanistic Interpretability of EEG Foundation Models via Sparse Autoencoders

2026-05-13 · William Lehn-Schiøler, Magnus Ruud Kjær, Rahul Thapa, Magnus Guldberg Pedersen 외 arxiv

EEG foundation models achieve state-of-the-art clinical performance, yet the internal computations driving their predictions remain opaque: a barrier to clinical trust. We apply TopK Sparse Autoencoders (SAEs) across thr…

Aligning Sentence Embeddings to Human Concepts via Sparse Autoencoders

2026-06-19 · Wonseok Shin, Songkuk Kim arxiv

Dense sentence embeddings are fundamental to modern Retrieval-Augmented Generation (RAG) systems but suffer from a lack of interpretability due to feature superposition. This opacity hinders the alignment of retrieval pr…

Information Retrieval

Stable and Steerable Sparse Autoencoders with Weight Regularization

2026-03-04 · Piotr Jedryszek, Oliver M. Crook arxiv

Sparse autoencoders (SAEs) are widely used to extract human-interpretable features from neural network activations, but their learned features can vary substantially across random seeds and training choices. To improve s…