Introduction to Large Language Models with Applications in Text Classification

This workshop offers an introduction to the core concepts underlying Large Language Models (LLMs) like BERT and GPT, as well as the various tools available to users interested in employing these models for practical tasks. As the motivating example, we will use LLMs to classify text answers to questions on a real-world survey. We will cover the following topics:

1. What are LLMs? – We will start with a brief overview of the development of LLMs, what they are, and the tasks for which they are used. While we may “peek under the hood” of these models, the focus will be on broad concepts rather than technical details.
2. What tools do we use? – We will show how to use the PyTorch and Hugging Face packages in python to fine-tune and evaluate LLMs. We will also cover how to use OpenAI’s API to classify text with ChatGPT.

By the end of the workshop, participants will have a foundational understanding of LLMs and the basic tools we use to implement them in practice. This is a workshop for practioners, by a practioner, and so the focus will be on high-level concepts and applications rather than technical details

If you wish to follow along with coding examples during the workshop, bring a laptop with python installed. If you’re new to python, anaconda (https://www.anaconda.com/download) is a good option which bundles python with some commonly used data science packages. Coding examples will be provided to participants for use during and after the workshop.