T3LLM The implementation for our paper titled Chain-of-thought Reviewing and Correction for Time Series Question Answering.