LLMs have been hot ever since ChatGPT was released to the general public. While most of the applied AI folks know how to use the APIs, there is perhaps a lack o

LLMs have been hot ever since ChatGPT was released to the general public. While most of the applied AI folks know how to use the APIs, there is perhaps a lack of mathematical understanding of ‘why’ it works. Here is a comprehensive deconstruction of LLMs in general and Transformers in particular from a mathematical perspective.