Joining My Group: Information for Prospective Students and Collaborators

I am interested in working with motivated students and collaborators who want to engage with research in interpretability, training data attribution, and large language model evaluation. If you are considering reaching out, please read this first.

What I am looking for

I am looking for students who:

• Are interested in the science of LLMs: interpretability, analysis, and evaluation. These are broad topics and include different domains and questions I find interesting. I find other topics interesting as well, but the science of LLMs is the core.
• Are willing to work hard: read, write, code, and learn a lot.
• Can commit full-time to work in my lab (or convince me why you can work elsewhere at the same time).

Before you contact me

Familiarize yourself with my work:

• Visit my website and skim the papers on my selected publications page, or the most recent ones.
• Carefully read ~three papers that you find the most interesting.
• Think about what questions or directions appeal to you and why.

Don't worry about finding the perfect research topic. But you should get familiar with the kinds of questions I'm interested in, and be able to explain why you find them interesting. Alternatively, tell me about your own direction.

What to include in your email

Send me an email with the following:

• Your CV
• Your availability: how many days a week you will spend working on research (I prefer students who can commit full-time)
• A brief statement about why you are interested in my research (or working with me)

What happens next

If your background and interests align with what I am looking for, I will get back to you to schedule some time to meet — I can't answer all emails — sorry!

p.s. I do not appreciate fully LLM-generated emails. The students in my group are there to learn, and I view my role in academia as training humans (mostly), so I expect communication to be with the aforementioned humans, not bots.