Yiqiao Yin. "Use large language model to enhance reasoning of another large language model through reward updated GRPO." Scientific Reports (2026). https://doi.org/10.1038/s41598-026-39296-8