Implementing An Automated Socratic Method to Reduce Hallucinations in Large Language Models | Synapse