Stronger Together: On-Policy Reinforcement Learning for Collaborative LLMs Paper โข 2510.11062 โข Published Oct 13 โข 28