Model See Model Do: Speech-Driven Facial Animation with Style Control