I'm a CS student at UPenn. I work on making intelligent systems smarter and, more broadly, understanding and aligning them. I see these as fundamentally connected rather than in tension. It's an ambitious research direction, which is why I emphasize building good tools. My open-source libraries are now used across several research labs.
I hope my work can help expand human possibilities and productivity for long-term prosperity. Outside of research, I row for my college varsity team and served in the Marine Corps.
Papers(See more)
2024
We succeed in systematically controlling language model behavior with programmables rules like "if input is about xxx, then refuse."
2024
We argue that language-only models lack understanding of the physical manifestation of language, as demonstrated through a series of tasks called the H-Test.
2023
We present a synthetic instruction-response generation framework designed to mimic the sequential and orderly nature of human learning.
2021
We show that combining handcrafted linguistic features with transformers can create the state-of-the-art readability classification model.