Evaluating Creative Short Story Generation in Humans and Large Language Models