SPO++: Stream-Aligned Policy Optimization for Asynchronous Agentic RL

By Kai Ruan · Paper · cs.AI

Group-relative reinforcement learning waits for sibling rollouts of the same prompt, which is costly for long and variable tool-use trajectories. Single-stream Policy Optimization (SPO) removes this dependency with a persistent prompt-level value estimate, but its recipe whitens

AI Agents · Cs.ai

View original

HomeResourceLoading…