Restoring Calibration for Aligned Large Language Models: A Calibration-Aware Fine-Tuning Approach | Read Paper on Bytez