VLA-Forget: Vision-Language-Action Unlearning for Embodied Foundation Models

作者：Ravi Ranjan · 单位：Florida International University · 会议/期刊：arXiv · 日期：2026-04-05 · 来源：Bottlenecks / Safety