Researchers from the University of Auckland Introduced ChatLogic: Enhancing Multi-Step Reasoning in Large Language Models with Over 50% Accuracy Improvement in Complex Tasks