Addressing Topic Granularity and Hallucination in Large Language Models for Topic Modelling