Abstract
Vision-language-action (VLA) and world-action models (WAMs) often degrade under out-of-distribution task variations despite retaining partial task capability. To recover such capability, we propose RoboIRS, an inference-time internal representation steering method that uses successful and failed rollouts to train linear classifiers, select outcome-relevant intervention locations, and derive task-specific steering directions without updating policy parameters. On 15 simulation tasks with a frozen $π0.5$ policy, RoboIRS improves the average success rate from 44.4% to 66.2%, outperforming alternative inference-time intervention baselines while adding little inference time. We further validate RoboIRS on real-robot manipulation using the same $π0.5$ policy and demonstrate its applicability to a world-action model Cosmos Policy, where the average success rate improves from 35.4% to 55.4%. These results show that directly steering internal robot-policy representations can improve the performance of robot policies at inference time. Project website is available at https://rollingoat.github.io/roboirs/.
Keywords
Subject
Publication details
- Journal
- Not available
- Open access
- Green open access
Cite this article
APA 7
Lei, J., Liu, C., Li, D., Zhang, Z., Liang, X., She, Y., Fan, Z., & Zheng, M. (2026). RoboIRS: Inference-Time Internal Representation Steering for Generalist Robot Policies. https://omanscience.com/en/articles/roboirs-inference-time-internal-representation-steering-for-generalist-robot-policies
MLA 9
Lei, Jiuzhou, et al. "RoboIRS: Inference-Time Internal Representation Steering for Generalist Robot Policies." https://omanscience.com/en/articles/roboirs-inference-time-internal-representation-steering-for-generalist-robot-policies.
Chicago (author–date)
Lei, Jiuzhou, Chang Liu, Dayou Li, Zhiyuan Zhang, Xiao Liang, Yu She, Zhiwen Fan, and Minghui Zheng. 2026. "RoboIRS: Inference-Time Internal Representation Steering for Generalist Robot Policies." https://omanscience.com/en/articles/roboirs-inference-time-internal-representation-steering-for-generalist-robot-policies.
Harvard
Lei, J., Liu, C., Li, D., Zhang, Z., Liang, X., She, Y., Fan, Z. and Zheng, M. (2026) 'RoboIRS: Inference-Time Internal Representation Steering for Generalist Robot Policies', Available at: https://omanscience.com/en/articles/roboirs-inference-time-internal-representation-steering-for-generalist-robot-policies.
Vancouver
Lei J, Liu C, Li D, Zhang Z, Liang X, She Y, et al. RoboIRS: Inference-Time Internal Representation Steering for Generalist Robot Policies. https://omanscience.com/en/articles/roboirs-inference-time-internal-representation-steering-for-generalist-robot-policies
IEEE
J. Lei, C. Liu, D. Li, Z. Zhang, X. Liang, Y. She, Z. Fan, and M. Zheng, "RoboIRS: Inference-Time Internal Representation Steering for Generalist Robot Policies," https://omanscience.com/en/articles/roboirs-inference-time-internal-representation-steering-for-generalist-robot-policies.