Uncertainty Quantification For Large Language Models