Dr. Yanai Elazar
Interpretability, Data Attribution, Generative AI, Machine Learning, Natural Language Processing
CV
Dr. Elazar is an expert in generative models and machine learning. He completed his PhD at Bar-Ilan University and went on to postdoctoral positions at the Allen Institute for AI (AI2) and the University of Washington. Throughout his career he has received prestigious awards, including a Google Fellowship for outstanding doctoral students and a Rothschild Fellowship for young researchers. He is an active figure in the international AI community, organizing leading meetings and workshops (such as BlackBox NLP) and collaborating with researchers from top institutions across the US, Canada, and Europe.
Research
Explainability and Reliability of Language Models (NLP)
Generative models such as large language models achieve impressive results across a wide range of domains, but can they be trusted? By design, generative models based on machine learning algorithms operate as a black box, and we do not understand how they work, when, or why. For complex and high-stakes tasks, the ability to understand, inspect, and control generative models becomes essential.
Dr. Yanai Elazar develops methods that "open" this black box and enable us to understand model behavior. His research includes building tools to identify features and trace the sources from which a model has learned, with the goal of making systems more reliable and transparent. His work enables the analysis of the processes that lead to model behaviors such as trustworthiness, fairness, efficiency, and knowledge.
Key Research Areas:
Interpretability: Uncovering the internal mechanisms behind model decision-making.
Evaluation of Generative Models: Developing methodologies to assess model reliability and capabilities.
Data Attribution: Analyzing the influence of training data on model output.
Research Approach:
Experimental - Analysis of generative models and development of tools to evaluate their behaviors.
Career Prospects:
Industry: AI Researchers at leading technology companies.
Academia: Researchers and faculty members in the fields of AI and Machine Learning.
Publications
Last Updated Date : 30/07/2026
