Are Large Language Models More Empathetic than Humans?