Sun, Y., Zhao, Z., Wei, Y., Zhang, Y., & Gong, C. (2026). Well Begun, Half Done: Reinforcement Learning with Prefix Optimization for LLM Reasoning. Proceedings of the AAAI Conference on Artificial Intelligence, 40(39), 33144–33152. https://doi.org/10.1609/aaai.v40i39.40598