CSRP: Chain-of-Thought Reasoning for Chinese Text Correction via Reinforcement Learning with Efficiency-Aware Rewards.

Wei Tian, Yuhao Zhou, Man Lan

Browse the full ACL paper archive.