In this segment of the LLM course, we’ll break down the transformers' architecture and explain the mechanics behind encoders and decoders, embeddings, multi-headed attention, and the significance of a feed-forward layer. You’ll learn the advantages of transformers over RNNs.

We’ll examine GPT models closely and begin our practical part of the LLM tutorial. We’ll connect to OpenAI’s API and implement a simple chatbot with a personality: a poetic chatbot. I’ll also show you how to use LangChain to work with your own custom data, feeding information from the 365 web pages to our model.

The Hugging Face package is an open-source package, which allows us an alternative way to interact with LLMs. We’ll learn about pre-trained and customized tokenizers and how to integrate Hugging Face into Pytorch and Tensorflow deep learning workflows.

This section of our Intro to Large Language Models course will explore BERT's architecture and contrast it with GPT models. It will delve into the workings of question-answering systems both theoretically and practically and examine variations of BERT—including the optimized RoBERTa and the smaller lightweight version DistilBERT.

In the final Intro to Large Language Models course section, we’ll look under the hood of XLNET (a novel LLM), that uses permutations of data sets to train a model. We’ll also compare XLNet and our previously discussed models, BERT and GPT.