INFO 5720

How LLMs work, their potential and limitations

Course Description

Course information provided by the 2026-2027 Catalog.

Large language models (LLMs) are ubiquitous. This course is about how they work, understood in both a technical and holistic sense: what's going on inside these models, and how do they interact with the rest of the world? More specifically, this course will cover: 1. What a (large) language model actually is and how it compares to other types of ML models. 2. The current LLM ecosystem. 3. Prompting, "thinking", in-context learning, finetuning. 4. How LLMs are evaluated. 5. How to interpret what LLMs are doing.

Prerequisites background equivalent to: INFO 1100 or CS 1110 or CS 1112.

Last 4 Terms Offered (None)

Learning Outcomes

• Describe the current ecosystem of neural network language models.
• Evaluate capabilities of models through qualitative and quantitative methods.
• Apply language models to new tasks through finetuning and in-context learning.
• Contrast contemporary neural language models with existing probabilistic methods.
• Analyze the parameters and activation behaviors of networks.
• Document the results of experiments and evaluations in writing with tabular and graphical supporting material.

View Enrollment Information