REVIEW 11 cited by
Mechanistic interpretability for steering vision-language-action models
Not yet reviewed by Pith; the record is open.
This paper has not been read by Pith yet. Machine review is queued; the pith claim, tier, and objections will appear here once it completes.
SPECIMEN: schema-true, not a live event
T0 review · schema-true
One-sentence machine reading of the paper's core claim.
pith:XXXXXXXX · record.json · timestamp
Mechanistic interpretability for steering vision-language-action models
read the original abstract
Vision-Language-Action (VLA) models are a promising path to realizing generalist embodied agents that can quickly adapt to new tasks, modalities, and environments. However, methods for interpreting and steering VLAs fall far short of classical robotics pipelines, which are grounded in explicit models of kinematics, dynamics, and control. This lack of mechanistic insight is a central challenge for deploying learned policies in real-world robotics, where robustness and explainability are critical. Motivated by advances in mechanistic interpretability for large language models, we introduce the first framework for interpreting and steering VLAs via their internal representations, enabling direct intervention in model behavior at inference time. We project feedforward activations within transformer layers onto the token embedding basis, identifying sparse semantic directions - such as speed and direction - that are causally linked to action selection. Leveraging these findings, we introduce a general-purpose activation steering method that modulates behavior in real time, without fine-tuning, reward signals, or environment interaction. We evaluate this method on two recent open-source VLAs, Pi0 and OpenVLA, and demonstrate zero-shot behavioral control in simulation (LIBERO) and on a physical robot (UR5). This work demonstrates that interpretable components of embodied VLAs can be systematically harnessed for control - establishing a new paradigm for transparent and steerable foundation models in robotics.
Forward citations
Cited by 11 Pith papers
-
Present but Not Remembered: Auditing How Frozen VLAs Encode, Deploy, and Steer Visual History
VLAs linearly encode past frames as near-redundant copies of the present and deploy them only as architecture-conditional fallback or standing use, so steerability tracks deployment regime not encoding.
-
Event-Grounded Sparse Autoencoders for Vision-Language-Action Policies
Event-grounded SAE analysis in VLA policies produces stronger causal effects on robot behavior than standard methods by anchoring features to clustered end-effector keyframes across simulations and real-robot tests.
-
Local Linearity of LLMs Enables Activation Steering via Model-Based Linear Optimal Control
Local linearity of LLM layers enables LQR-based closed-loop activation steering with theoretical tracking guarantees.
-
What Do They See? Interpreting Complex Road Scenarios Through the Eyes of Vision-Language-Action Models for Safe and Trustworthy Autonomous Vehicle Learning
Erasing objects from front-camera images shows Alpamayo 1's trajectories depend most on large vehicles, pedestrians, and traffic lights, but attributions are seed-unstable and some effects reach the output without tou...
-
Inference-Time Robot Behavior Steering through Physically-Aware Reconfiguration of Task-Structure
ReStruct steers robot policies at inference time by reconfiguring task structure with neural automata and synchronous products, claiming up to 25% gains over VLA models in success and preference adherence.
-
Learning What to Say to Your VLA: Mostly Harmless Vision Language Action Model Steering
A search-and-distill framework with conformalized improvement head produces a language feedback policy that boosts frozen VLA performance by 24.7% in simulation and 65% on hardware while guaranteeing harmlessness on p...
-
Interpreting Video Representations with Spatio-Temporal Sparse Autoencoders
Spatio-temporal contrastive SAEs recover temporal coherence lost by hard TopK, improve action probes by +3.9% and retrieval by up to 2.8× R@1, and expose a monosemanticity metric artifact.
-
Open-Loop Planning, Closed-Loop Verification: Speculative Verification for VLA
SV-VLA uses infrequent heavy VLA planning of action chunks plus a lightweight closed-loop verifier to achieve both efficiency and robustness in dynamic robot control.
-
Behavior Uncloning: Distilling Mode Redirection into Policy Weights without Inference-Time Steering
MoRE improves robot policy success rates by 44 percentage points by distilling mode redirection into weights, matching filtered retraining performance without inference overhead.
-
Act on What You See: Unlocking Safe Social Navigation in Vision-Language-Action Models
SALSA aligns social features and adds future-risk signals in VLA models to cut near-collisions by 86.4% and raise social accuracy from 53% to 93% on SCAND and real robots.
-
Early Warning Signals for OpenVLA Failure under Visual Distribution Shift
OpenVLA layer-16 activations allow a logistic probe to predict failure within 15 steps under occlusion (AUROC 0.972) better than baselines, with some transfer to camera jitter.
discussion (0)
Sign in with ORCID, Apple, or X to comment. Anyone can read and Pith papers without signing in.