Biases in Edge Language Models: Detection, Analysis, and Mitigation