Elizabeth Schaefer '26: A “random class” led her to large language models
With a double major in Computer Science and Humanities, Elizabeth Schaefer is using large language models (LLMs) to tackle a wide range of problems, such as how to detect and reduce gender biases in medical training datasets. In a rare opportunity for an undergraduate researcher, she presented this work to the Association for Computational Linguistics (ACL) in Vienna last year.
“I think a lot of my interest in medicine and AI for medicine is about the opportunity to have an actual impact in the world,” she said.
Beyond research, Elizabeth is deeply involved in the Yale community. She serves as president of Morse College, her residential college, and has been a head teaching assistant for the introductory computer science sequence. She also works in the Yale Admissions Office, where she leads the graphic design team. We spoke to Elizabeth about interdisciplinary research, making the most of her time at Yale, and how AI can help surgeons.
Why did you choose to attend Yale?
I was deciding between Yale and a similar peer, maybe a more engineering-esque school. But I thought Yale Engineering was so amazing because I think what engineering really is, and computer science too, is having a deep understanding of the real world and then solving it. And Yale's interdisciplinary nature would allow me to study a wide range of things instead of just being stuck in engineering school. I'm so happy I made that decision—I love it here so much. I actually work for Yale Admissions, so I spend every year trying to convince more kids to choose Yale. I think it comes down to interdisciplinary nature, the communities you can build here, and I love the residential college system.
Did you expect to major in both computer science and English?
No, I was thinking just CS; double majoring really didn't occur to me. But the humanities major was very flexible and allowed me to work really well with my CS curriculum. And I realized, I think junior year, that I had taken so many literature classes just for fun that I had enough to make up the major itself. All I had left to do was a thesis. So I added the humanities major pretty late, but that was a good feeling.
You based your thesis on research you did in the lab of Prof. Arman Cohan.
We focused a lot on explainable chain of thought [an AI prompt engineering technique] for large language models in surgical contexts. Working with my collaborators, we would test the large language models—the frontier ones that are currently ahead of the game. We would try to test them on surgical vignettes. And instead of just seeing if they got the answer correct—there's a lot of benchmarks that will do that—we actually had the surgeons look at how the model was able to explain its answer and see how comparable that would be to true surgeon thought.
More Details
Published Date
May 12, 2026
