Evolving Knowledge Distillation with Large Language Models and Active Learning