Do Language Models Exhibit Human-like Structural Priming Effects?