TY  - RPRT
TI  - Latent Activation Editing: Inference-Time Refinement of Learned Policies for Safer Multirobot Navigation
AU  - Satyajeet Das
AU  - Darren Chiu
AU  - Zhehui Huang
AU  - Lars Lindemann
AU  - Gaurav S. Sukhatme
PY  - 2026
UR  - https://arxiv.org/abs/2509.20623
ID  - 2509.20623
ER  -