New asynchronous RL method Li2SAO targets efficient post-training of large language models
A paper introduces Li2SAO, an asynchronous reinforcement learning framework for LLM post-training that uses group-relative actor-critic bootstrapping with a light single-rollout design to improve training efficiency.
Read the full story at arXiv cs.CL ↗
Timeline · 1 report
- 2026-11-02 05:00 · arXiv cs.CL
Li$^{2}$SAO: Lifting Light Singl-Rollout Asynchronous RL Optimization via Group-Relative Actor-Critic Bootstrapping