Editing the Goal: Per-Token SAE Steering and Its Depth Limits in pi 0.5
cs.AI
Submitted: 2026-10-31
Updated: 2026-10-31
Comments: Accepted to the Science of Physical AI Safety (SPAIS) Workshop at CoRL 2026. 6 pages, 2 figures
Code: https://github.com/Physical-Intelligence/openpi
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Sparse Autoencoders Reveal Interpretable and Steerable Features in VLA Models
- Observing and Controlling Features in Vision-Language-Action Models
- VLAs are Confined yet Capable of Generalizing to Novel Instructions
- PaliGemma: A versatile 3B VLM for transfer
- LIBERO-PRO: Towards Robust and Fair Evaluation of Vision-Language-Action Models Beyond Memorization
Related papers
- Multi-Objective Structured Pruning of LLMs for Latency and Model Size Optimization
- MAVEN-T: Reinforced Heterogeneous Distillation for Real-Time Multi-Agent Trajectory Prediction
- Model Discovery Agent: LLM-assisted Bayesian experiment design for data-efficient discovery of mechanistic world models
- Embeddings for Preferences, Not Semantics
- Can We Trust Item Response Theory for AI Evaluation?
- Diagnosing Task Insensitivity in Language Agents