Improving Scientific Hypothesis Generation with Knowledge Grounded Large Language Models