Athulith Paraselli

Athulith Paraselli

He/Him

Master’s student and researcher studying mechanistic interpretability, multimodal models, and AI cognition.

Hello! I am a second-year Master’s student in Computer Science at Brown University, working with Ellie Pavlick and Thomas Serre. I am broadly interested in cognition and causal abstractions of AI systems. My research uses interpretability methods to uncover the mechanisms AI systems use to solve tasks, with the goal of establishing stronger guarantees about when they can be trusted alongside understanding fundamental principles of intelligence.

Before Brown, I completed my B.S. in Mathematics–Computer Science at UC San Diego, where I had the opportunity to work with Dr. George Sugihara, Dr. Xi Li, and Dr. Zhuowen Tu on projects spanning ecological forecasting, multimodal models, and robust machine learning. I also served as one of the inaugural project leads for Eta Kappa Nu’s (HKN) projects program, where I led the development of a data pipeline and predictive model for real-time forecasting of bioluminescent algal blooms off the coast of La Jolla!

Selected Publications

Slow to See, Slow to Suppress: Understanding the Effects of Modality in Context-Memory Conflicts
Athulith Paraselli, Etha Tianze Hua, Ellie Pavlick
Findings of EMNLP 2026
[Paper] [Project page] [Code]

Updates