Vision-language-action (VLA) models predict chunks of future actions from the current observation, an assumption that fails under partial observability, where decisions depend on information no longer visible. Existing memory-augmented VLAs simultaneously introduce recurrence, retrieval, compression modules, auxiliary objectives, hierarchical memory, or task-specific architectural changes, so the contribution of recurrence itself remains entangled with surrounding machinery. We present a controlled isolation study of recurrence in a strong pretrained VLA backbone. Our formulation augments the transformer with a small set of learnable memory tokens carried across timesteps and updated through self-attention, trained end to end with truncated backpropagation through time, with no auxiliary losses and no architectural changes. We instantiate this as $μ$VLA, a family of OpenVLA-OFT variants parameterized by memory width m, TBPTT length K, and the memory update rule (cross-step gradients or a detached EMA), so that recurrence is the only varying factor. On MIKASA-Robo, $μ$VLA improves average success rate on five training tasks from 0.42 to 0.84 at the strongest setting and reaches 0.23 on held-out tasks with the same memory structure versus 0.07 for the memoryless baseline. On tasks requiring different memory structure, performance remains near baseline. On LIBERO, the strongest recurrent variant achieves 96.2% average success, indicating no regression under full observability. We interpret these results as a calibration of the capability envelope of minimal in-backbone recurrence, identifying the regime in which it is sufficient and the regime where additional memory structure is required. Demos and videos can be found in https://avanturist322.github.io/mu-vla/.

翻译：视觉-语言-动作（VLA）模型根据当前观测预测未来动作序列，但这一假设在部分可观测场景中失效——当决策依赖不再可见的信息时。现有增强记忆的VLA模型同时引入循环机制、检索、压缩模块、辅助目标、分层记忆或任务特定架构改动，导致循环机制本身的贡献与附属机制相互纠缠。本文针对强预训练VLA骨干网络中的循环机制开展受控隔离研究。我们提出的方法通过一组跨时间步携带、经自注意力更新的可学习记忆令牌增强Transformer，采用截断时间反向传播进行端到端训练，不引入辅助损失或架构改动。以此为基础构建$μ$VLA系列——OpenVLA-OFT变体，参数化变量包括记忆宽度m、TBPTT长度K及记忆更新规则（跨步梯度或解耦EMA），使循环机制成为唯一变化因素。在MIKASA-Robo基准上，最强配置的$μ$VLA将五项训练任务的平均成功率从0.42提升至0.84，在需要相同记忆结构的保留任务上达到0.23（无记忆基线为0.07）。对于需要不同记忆结构的任务，性能接近基线。在LIBERO基准上，最强循环变体实现96.2%平均成功率，表明完全可观测场景下无性能退化。我们将这些结果解读为对最小化骨干网络循环能力边界的校准，明确其充分有效的适用区间以及需要额外记忆结构的场景。演示与视频见https://avanturist322.github.io/mu-vla/。