Towards Objective Fine-tuning: How LLMs' Prior Knowledge Causes Potential Poor Calibration?

Open in new window