I am a Ph.D. candidate in Natural Language Processing (NLP) at The Hebrew University of Jerusalem, advised by Prof. Gabriel Stanovsky. I am currently a research intern at Google Research and previously interned at IBM Research and Allen Institute for AI (AI2).

My research focuses on evaluating large language models (LLMs): what we can reliably conclude from an evaluation, and what it tells us about how models compare to humans. One line of my work aims to make evaluation itself more trustworthy and efficient, by accounting for prompt sensitivity and stochasticity (ReliableEval, PromptSuite) and by extending Item Response Theory to multilingual evaluation. Another line takes a cognitive perspective, comparing humans and LLMs on phenomena such as the framing effect and in-context learning. Most recently, I have been studying LLM agents, asking whether they can infer a model of their environment through interaction.

I’m also a dogs lover, pop music fan, and a former professional basketball player.