Understanding Tokens and Cost in Large Language Models