TY  - RPRT
TI  - Retrieve, Don't Retrain: Extending Vision Language Action Models to New Tasks at Test Time
AU  - Jeongeun Park
AU  - Juhan Park
AU  - Taekyung Kim
AU  - Sungjoon Choi
AU  - Dongyoon Han
AU  - Sangdoo Yun
PY  - 2026
UR  - https://arxiv.org/abs/2606.15631
ID  - 2606.15631
ER  -