Investigating the Role of Prompting and External Tools in Hallucination Rates of Large Language Models