Gabriel Stanovsky
Gabriel Stanovsky
Associate Prof. at HUJI
I develop AI models that help experts tackle real-world research questions across domains such as law, medicine, archaeology, and beyond. I also head HUJI's BSc. program in AI. Outside work, I’m a wannabe movie buff, concert goer, book lover, and radio enthusiast.
People
Uri Berger
Language Acquisition, specifically in realistic settings- e.g. using interaction or multi-modality.
Coadvised with Omri Abend and Lea Frermann.
Eliya Habba
NLP in legal domain.
Coadvised with Renana Keydar.
Itay Itzhak
Language models interpretability, biases and evaluation.
Coadvised with Yonatan Belinkov.
Gili Lior
Exploring aspects at the intersection of human cognition and natural language processing.
Adi Gabay
Epistemic and logical reasoning in LLMs.
Coadvised with Liat Peterfreund.
Hoshen Lugasi
Timing and social intelligence in LLM agents.
Ido Oron
Musical Zeitgeist detection in LLM embeddings.
Coadvised with Yoel Greenberg.
Shahar Spencer
Explainable AI for Ancient Language Processing.
Coadvised with Shai Gordin.
Recent Papers
From Feelings to Metrics: Understanding and Formalizing How Users Vibe-Test LLMs
2026 · COLM
Growing Pains: Extensible and Efficient LLM Benchmarking Via Fixed Parameter Calibration
2026 · COLM
Proceedings of the Fifth Workshop on Generation, Evaluation and Metrics (GEM)
2026 · ACL Workshop
ScheMatiQ: From Research Question to Structured Data through Interactive Schema Discovery
2026 · ACL Demo
Leveraging Digitized Newspapers to Collect Summarization Data in Low-Resource Languages
2026 · EACL Findings
Courses & Seminars
Introduction to Machine Learning
2022- · Undergraduate · 600+ students
Advanced Natural Language Processing
2023- · Graduate · 60 students
A Seminar on Causality
2026 · Graduate · 20 students
On Language Models and Consciousness: A Seminar on NLP and Philosophy
2022 · Graduate · 20 students
Recent Talks
ScheMatiQ: From Research Question to Structured Data
2026, Ai2, Seattle, Washington
On the Brittleness of Evaluation in NLP
2025, University of Illinois Urbana-Champaign, USA
Beyond Benchmarks: On The False Promise of AI Regulation
2025, The Annual President Chaim Herzog Conference
Multilingual Zero-Shot Transfer in Low-Resource Settings
2023, Keynote at the Workshop on Ancient Language Processing, Bulgaria
Contact & CV
For collaboration, advising, or speaking invitations, email me at gabriel.stanovsky@mail.huji.ac.il.