We will calculate your course grade based on these components:

Component Weight Description
Participation 10% Show up to class and contribute to discussion.
Reading Responses 10% Short written reflections on assigned papers, submitted before class.
Paper Presentations 15% Present research papers and lead class discussion. Expect no more than 2 presentations per student in this course.
Red Teaming Homework 15% Hands-on assignments where you will attempt to jailbreak or bypass safety mechanisms in LLMs, document your methodology, and analyze the results.
Course Project 50% Conduct original research in AI security, culminating in a conference-style paper. Individual or pairs; larger groups require proportionately more work. Grade based on milestone artifacts and final report.