Hello!

I am an assistant professor of Linguistics (and by courtesy Computer Science) at the University of British Columbia. Before that, I was a Kempner Fellow at Harvard, and before that I did my PhD in Computer Science at the Stanford NLP group, advised by Dan Jurafsky.

I am looking to recruit students. If you are applying for a PhD in NLP or computational linguistics, apply to UBC linguistics or CS and mention that you'd like to work with me!

I work on understanding and defining the capabilities of large language models in relation to the human language system.

I am especially interested in pursuing an interdisciplinary research program, combining computational empirical machine learning methods with theories of human language. My principal interests include: how language models learn and use generalizable grammatical abstractions, the interaction between structure and meaning representations in high-dimensional vector spaces, and using multilingual settings to test the limits of abstraction in language models.

Talks and News

Joining the Lab

I'm looking for PhD and masters students to join the lab! You can work with me if you apply through Linguistics or through CS. The UBC NLP Group is growing, with many new faculty members and exciting research directions, and we'd love to have you join! I am generally looking for students who want to deeply understand how lanugage models work, who have a scientific interest in human language, and who have some experience working with machine learning. We are a very open-minded linguistics department, and a fun, interdisciplinary NLP group with research to suit many tastes.

You can apply for a PhD in Linguistics or CS, or for a masters (ling masters, CS masters). You can also apply to both departments. There are different experience requirements for the programs, and also slightly more funding sources available to support students through Linguistics, I'm happy to advise on which one is a better fit.

I tend to not reply to open-ended general emails about joining the group, sorry. If you have a specific question about joining the lab or working together, please do feel free to email me! (It's great if you can include a short sentence about why you think our research interests match, but don't worry too much about crafting this)

Papers

Some representative papers

Injecting structural hints: Using language models to study inductive biases in language learning - Isabel Papadimitriou and Dan Jurafsky, Findings of EMNLP 2023

Mission: Impossible Language Models - Julie Kalini, Isabel Papadimitriou , Richard Futrell, Kyle Mahowald, and Christopher Potts, ACL 2024 Best Paper 🏆

Deep Subjecthood: Higher-Order Grammatical Features in Multilingual BERT - Isabel Papadimitriou , Ethan A. Chi, Richard Futrell, and Kyle Mahowald, EACL 2021 (oral presentation) [pdf] [code]

All Papers

Vocabulary embeddings organize linguistic structure early in language model training - Isabel Papadimitriou and Jacob Prince (preprint)

Investigating the interaction of linguistic and mathematical reasoning in language models using multilingual number puzzles - Antara Raaghavi Bhattacharya, Isabel Papadimitriou, Kathryn Davidson, and David Alvarez-Melis EMNLP 2025

Interpreting the Linear Structure of Vision-language Model Embedding Spaces - Isabel Papadimitriou*, Huangyuan (Chloe) Su*, Thomas Fel*, Sham Kakade, and Stephanie Gil, COLM 2025 spotlight presentation 🔦

Using Shapley interactions to understand how models use structure - Divyansh Singhvi, Diganta Misra, Andrej Erkelens, Raghav Jain, Isabel Papadimitriou*, and Naomi Saphra*, ACL 2025

Archetypal SAE: Adaptive and Stable Dictionary Learning for Concept Extraction in Large Vision Models - Thomas Fel, Ekdeep Singh Lubana, Jacob S. Prince, Matthew Kowal, Victor Boutin, Isabel Papadimitriou, Binxu Wang, Martin Wattenberg, Demba Ba, and Talia Konkle, ICML 2025

Mission: Impossible Language Models - Julie Kalini, Isabel Papadimitriou , Richard Futrell, Kyle Mahowald, and Christopher Potts, ACL 2024 Best Paper 🏆

Injecting structural hints: Using language models to study inductive biases in language learning - Isabel Papadimitriou and Dan Jurafsky, Findings of EMNLP 2023

Oolong: Investigating What Makes Crosslingual Transfer Hard with Controlled Studies - Zhengxuan Wu*, Alex Tamkin*, Isabel Papadimitriou*, EMNLP 2023 [pdf]

Separating the Wheat from the Chaff with BREAD: An open-source benchmark and metrics to detect redundancy in text - Isaac Caswell, Lisa Wang, Isabel Papadimitirou, GEM Workshop 2023

Multilingual BERT has an accent: Evaluating English influences on fluency in multilingual models - Isabel Papadimitriou* , Kezia Lopez*, and Dan Jurafsky, Findings of EACL 2023, SIGTYP 2023 [slides]

The Greek possessive modal eho as a special agentive modality - Isabel Papadimitriou and Cleo Condoravdi, LSA 2023 (poster) [abstract]

When classifying grammatical role, BERT doesn't care about word order... except when it matters - Isabel Papadimitriou , Richard Futrell, and Kyle Mahowald, ACL 2022 (oral presentation) [pdf] [code]

Language, Section 2.1 - Isabel Papadimitriou and Christopher D. Manning

In On the Opportunities and Risks of Foundation Models (full list of co-authors)

Deep Subjecthood: Higher-Order Grammatical Features in Multilingual BERT - Isabel Papadimitriou , Ethan A. Chi, Richard Futrell, and Kyle Mahowald, EACL 2021 (oral presentation) [pdf] [code]

Learning Music Helps You Read: Using transfer to study linguistic structure in language models - Isabel Papadimitriou and Dan Jurafsky, EMNLP 2020 (oral presentation) [pdf] [code]

Teaching

I was the TA for CS324H, History of Natural Language Processing, taught by Dan Jurafsky and Chris Manning in Winter 2024

I was a TA for CS224N, Natural Language Processing with Deep Learning, taught by Chris Manning in Winter 2023

I was the TA for the Independent Study in Machine Translation seminar taught by Noah Goodman in Winter 2020

At Berkeley I TAed CS70 Discrete Math and Probability Fall 2015, taught by Satish Rao and Jean Walrand

Other things

I did my undergraduate at Berkeley, where I studied Computer Science and in History. My history thesis was based on research at the archives of the League For Democracy in Greece, a London-based solidarity organisation supporting the left in the Greek Civil War. It received the Kirk Underhill Prize.

I often want a latex template of just a plain document, but where I can use ACL natbib as if I was writing a paper. If you do too, look here!

Did you ever want the word frequency counts of different words in the C4 corpus? I counted and uploaded them here.