Anat Kleiman

Kempner Graduate Fellow
PhD Student in Computer Science

Preferred Pronouns:

She/Her

Contact Information

Areas I Research:

About

Anat Kleiman is a PhD candidate in Computer Science at Harvard University and a Graduate Fellow at the Kempner Institute. Her research focuses on building reliable and adaptable large language models, with particular interest in how models acquire, retain, and selectively forget. Furthermore, she is also drawn to the rise of emergent capabilities in LLMs.

Research Focus

Kleiman studies reliable, adaptable, and efficient large language models, with a focus on continual learning, model merging, machine unlearning, and hallucination mitigation. Her broader interests include understanding how generative models generalize beyond their training data and developing methods that improve how models retain and reason over information.