I am a second-year PhD student in the UT Austin Computational Linguistics Group. I am fortunate to be advised by Kyle Mahowald. I work on interpretability, with a focus on what the study of language model internals can tell us more broadly about language and cognition.
Previously, I graduated from Cornell University Summa Cum Laude, with degrees in Computer Science, Mathematics, and Computational Cognitive Science (the latter through the College Scholar Program). During my undergraduate studies, I was fortunate enough to work with Dr. Morten H. Christiansen and Dr. Eugene Vinitsky utilizing Multi-Agent Reinforcement Learning to study idiolects in human language. I have also spent time at NASA’s Jet Propulsion Laboratories working on Retrieval-Augmented Generation (RAG) systems.
News
- Jul 2026Causal Drawbridges: Characterizing Gradient Blocking of Syntactic Islands in Transformer LMs accepted to COLM 2026. See you in San Francisco 🇺🇸!
- Apr 2026France or Spain or Germany or France: A Neural Account of Non-Redundant Redundant Disjunctions accepted to CogSci 2026. See you in Rio 🇧🇷!
- Nov 2025🏆 Causal Interventions Reveal Shared Structure Across English Filler–Gap Constructions awarded Outstanding Paper at EMNLP 2025!
- Aug 2025Causal Interventions Reveal Shared Structure Across English Filler–Gap Constructions accepted to EMNLP 2025 as Oral Spotlight. See you in Suzhou 🇨🇳!
- Apr 2025Reinforcement learning produces efficient case-marking systems accepted to CogSci 2025
- Oct 2024Models Can and Should Embrace the Communicative Nature of Human-Generated Math accepted to 4th MATH-AI Workshop at NeurIPS'24. See you in Vancouver!
- Aug 2024Started PhD at UT Austin
- May 2024Graduated Cornell!
- May 2024Defended thesis, awarded Summa Cum Laude Honors
Education
- PhD, Computational Linguistics (Minor: Computer Science), University of Texas at Austin, 2024–2029
- BA, Computer Science, Mathematics, and Computational Cognitive Science (Summa Cum Laude), Cornell University, 2020–2024
Selected Papers
-
🏆 Sasha Boguraev, Christopher Potts, Kyle Mahowald. Causal Interventions Reveal Shared Structure Across English Filler–Gap Constructions EMNLP 2025 Outstanding Paper Award, Oral Spotlight.
-
Sasha Boguraev, Kyle Mahowald. Causal Drawbridges: Characterizing Gradient Blocking of Syntactic Islands in Transformer LMs COLM 2026.
-
Sasha Boguraev, Qing Yao, Kyle Mahowald. France or Spain or Germany or France: A Neural Account of Non-Redundant Redundant Disjunctions CogSci 2026.
-
Sasha Boguraev, Katrin Erk, Kyle Mahowald, James Shearer, Steve Wechsler. Reinforcement learning produces efficient case-marking systems CogSci 2025.
In the Media
I was named one of the Cornell Arts and Sciences Extraordinary Seniors, and had a brief profile written on my time at Cornell.