They aimed to provide a practical guide for predicting the performance of large language models (LLMs) using smaller, ...