A Comparative Approach to Assessing Linguistic Creativity of Large Language Models and Humans