Despite the remarkable performance of large language models (LLMs), such as generative pre-trained Transformers (GPTs), across various tasks, they often perpetuate social biases and stereotypes ...