Lindia Tjuatja
/'lɪndia 'tʃuatʃa/ • 蔡玲丽 (càilínglì) • she/her
Hi! I’m Lindia, a final-year PhD at CMU’s Language Technologies Institute advised by Graham Neubig. I am also a frequent visitor at NYU with Tal Linzen. I work at the intersection of linguistics, cognitive science, and NLP. In Fall 2027, I will be an assistant professor at UT Austin Linguistics, joining the amazing computational linguistics faculty and broader NLP community.
My research largely centers around understanding how models of language learn structure in text (linguistic or otherwise), and how this may allow them to exhibit more complex behaviors. These interests manifest in work across evaluation and interpretability of large language models, especially with respect to when, where, and how they may be similar or different to humans. From an applications angle, I also work on how our understanding of these models can aid in creating tools to better serve understudied and endangered languages.
In my (precious) free time, I enjoy painting (oil and gouache are my mediums of choice), playing the viola, hanging out with my cat, running, and reading!
news
| Sep 26, 2025 | Took a walk down the street and gave a talk at the University of Pittsburgh’s Linguistics Colloquium on Linguistic Judgments of Language Models (slides can be found here). |
|---|---|
| Jul 30, 2025 | Presented “BehaviorBox: Automated Discovery of Fine-Grained Performance Differences Between Language Models” at ACL in Vienna! |
| Jan 22, 2025 | “What Goes Into a LM Acceptability Judgment? Rethinking the Impact of Frequency and Length” has been accepted to NAACL 2025! See y’all in Albuquerque! |
| Nov 11, 2024 | Presented two papers at EMNLP in Miami: “Do LLMs Exhibit Human-like Response Biases? A Case Study in Survey Design” and “GlossLM: A Massively Multilingual Corpus and Pretrained Model for Interlinear Glossed Text”! |
| May 16, 2023 | My first PhD paper was accepted to a conference! Keep an eye out for “Syntax and Semantics Meet in the Middle: Probing the Syntax-Semantics Interface of LMs Through Agentivity” at *SEM, co-located with ACL 2023 in Toronto. |