Sun, Y.; Zhao, Z.; Wei, Y.; Zhang, Y.; Gong, C. Well Begun, Half Done: Reinforcement Learning With Prefix Optimization for LLM Reasoning. AAAI 2026, 40, 33144-33152.