Our paper Steering Large Reasoning Models towards Concise Reasoning via Flow Matching is accepted by TMLR.