主页
← Awesome Robotics Manipulation Index
Awesome Robotics Manipulation · full_paper
TOPReward: Token Probabilities as Hidden Zero-Shot Rewards for Robotics
arXiv · 2026-02-22 · High-Level Structured Planning / Multimodal Reasoning / Reward Reasoning
任务规划
多模态推理
强化学习
操作
新标签打开 AlphaXiv
新标签打开 Paper
新标签打开 Code
AlphaXiv 中文论文页面(可滚动查看)