CS 224V

Project Info

Each project will be in a group of 2 students.

A core part of CS224V will be working on a project throughout the quarter. We are very excited to announce that we have mentors and advisors across many disciplines who have signed on to help you with these projects. These includes various leaders across Stanford as well as with external partners.

This is an exciting time for LLM-based projects because for the first time, it is possible to create a usable NLP project within a quarter. There are numerous meaningful applications that you can create and there are more tools you can build that enable new applications. The sky is the limit!

This document serves as an inspiration for students on their choice of course projects. It is a collection of potential project ideas. Some of these will lead to AI papers, or papers in other disciplines, or public services to be hosted on our World Wide Knowledge website. We are looking for your input on new projects as well. Many of these projects can be worked on by multiple groups, possibly using different corpora.

Project Proposals

[Due Wed 10/15 3:00pm] (Google Form)

Please answer all the questions in this document: Project Proposal Form

Project Proposal Presentations

[Wed 10/15, Mon 10/20, 3:00pm-4:20pm, in class]

The goal of these proposal presentations is to ensure that everybody gets feedback on their projects, and everybody benefits from learning what everybody is doing.

Because of the size of the class, if you are working on a mentor-proposed project, or if you have already presented your project on Oct 8, you do not need to participate in proposal presentations.

Presentation is mandatory for everybody else. Please sign up here for a slot to present.

Course Project Final Deliverables

Each project group is required to submit the following components for the course final project.

1. Poster Session (December 3)

We will hold a poster session on December 3, following the CS224N poster session format. Submit via Gradescope on December 2nd.

3. Demo Video Showcase (December 3)

Record a virtual presentation (1 min) demonstrating your prototype. Submit via Gradescope on December 2nd.

• LLM Agent Related Projects: Show an end-to-end demo of your working system.
• Model Training Projects: Use slides to explain your methods, results, and qualitative analysis.

If you are unsure whether a working system demo is required, please make a private Ed post.

2. Final Report (December 9)

The report should be at least 6 pages long (excluding references and appendix), formatted using the ACL submission template, and must follow ACL style guidelines (layout, references, supplementary materials, etc.).

Required Sections:

• Abstract – 2–3 paragraphs summarizing the paper, including key results
• Introduction – Motivation, main idea, and overall contributions
• Related Work
• Core Ideas / Methodology
• Experimental Results
• Insights and Discussion
• Future Work
• Conclusion
• Ethical Considerations – Explicitly discuss any potential ethical issues (e.g., data use, societal impact, or deployment implications). Refer to the ACL Ethics Guidelines for examples.
• Authorship Statemen – Briefly describe each member’s contributions (also required for single-author projects)
• References
• Appendix – Include examples, prompts, or supplementary materials

4. Code Submission (December 9)

Submit your complete codebase as a .zip file containing:

• A README with environment setup instructions and step-by-step guidance to reproduce results or run your prototype.
• All necessary files for running your system. Code that fails to compile or execute will negatively impact your grade.

Submission Details

All three components (Final Report, Demo Video, and Code) must be submitted via Gradescope by December 9.