FlashREINFORCE: Critic-Free Single-Rollout Asynchronous RL for Agentic Language Models

arXiv:2611.18627 · cs.CL, cs.LG · Submitted 2026-09-19 · Read on arXiv

cs.CL, cs.LG

Submitted: 2026-09-19

Updated: 2026-11-18

Related papers