Hak Hyun Kim

I am an M.S. student in Computer Science at Dartmouth College, advised by Soroush Vosoughi and Peter Chin. Before Dartmouth, I received my B.S. in Electronic and Electrical Engineering from Sungkyunkwan University, where I was advised by Hayoung Oh.

My research focuses on AI safety. I combine controlled experiments and mathematical analysis to study how internal representations shape behavior and to evaluate the reliability of AI systems.

My research spans mechanistic interpretability, training dynamics, multi-agent systems, AI systems, and applied machine learning. (Click a research question in Publications to see the answer.)

My goal is to design AI models that circumvent the need for post-hoc reverse-engineering and post-training alignment interventions. My research investigates how to construct inherently interpretable architectures with legible world models, and how pretraining paradigms dictate the internal computations that naturally drive aligned behavior.