Rohan Paul· @rohanpaul_ai · X·· 3 天前AI 评分62
AI 导读
商汤发布 SenseNova-RoboRSI，其核心在于体验内化，利用 GPT-6-Astra 通过上下文学习从示范中提取视觉、状态、动作和工具等信息。商汤将该具身经验蒸馏为开源的 RoboRSI-27B-Preview 模型，实现了主动感知、空间推理、规划与反馈修正，且任务成本约为其五十分之一。
正文 · 原文
The key idea behind SenseNova-RoboRSI is experience internalization.
GPT-6-Astra uses ICL to extract useful information from demonstrations, including visual examples, robot states, action trajectories, tools, and recovery strategies.
SenseTime distills this embodied experience into the open-source RoboRSI-27B-Preview, which demonstrates active perception, spatial reasoning, planning, and feedback-driven revision at roughly 1/50 the task cost.
来源：Rohan Paul · x.com