I am a research scientist at Anthropic, as well as a final-year PhD candidate at Mila - Quebec AI Institute, where I am advised by Marc G. Bellemare and Doina Precup. My core research interests are in reinforcement learning and decision-making, especially as applied to large language models.
Before graduate school, I worked as a software engineer at Google in San Francisco. Prior to that, I spent four ideal years at Brown University, where I studied math and computer science, among other things. There, I was fortunate to get my start in research by working with David Abel and Michael Littman.
Abstractive Red-Teaming of Language Model Character
Nate Rahn*, Allison Qi*, Avery Griffin, Jonathan Michala, Henry Sleight, Erik Jones
arXiv blog post
Controlling Large Language Model Agents with Entropic Activation Steering
Nate Rahn, Pierluca D’Oro, Marc G. Bellemare
ICML 2024 Workshop on Mechanistic Interpretability
arXiv
Policy Optimization in a Noisy Neighborhood: On Return Landscapes in Continuous Control
Nate Rahn*, Pierluca D’Oro*, Harley Wiltzer, Pierre-Luc Bacon, Marc G. Bellemare
NeurIPS 2023
arXiv
Value Preserving State-Action Abstractions
David Abel, Nate U. Rahn, Khimya Khetarpal, Dilip Arumugam, Doina Precup, Michael Littman
AISTATS 2020
paper