Our research at a glance

We are a new lab sitting at the intersection of cognitive science, artificial intelligence, and linguistics. The goal of our research is to understand how language works in minds and machines. What representations and algorithms support language learning, comprehension, and production? How does language interact with other forms of knowledge and reasoning? And how can our understanding of human cognition help us develop or interpret artificial intelligence systems? We approach these questions through a highly interdisciplinary approach, combining methods such as computational modeling, behavioral experiments, and mechanistic interpretability.

Learn more about our research through our recent Publications.

Current research areas

Cognitive interpretability of AI

How can we develop a scientific framework for understanding the cognitive capabilities and limitations of AI systems, especially large language models?

Show sample publications

Hu, J., & Frank, M. (2024). Auxiliary task demands mask the capabilities of smaller language models. COLM. https://openreview.net/forum?id=U5BUzSn4tD

Hu, J., Mahowald, K., Lupyan, G., Ivanova, A., & Levy, R. (2024). Language models align with human judgments on key grammatical constructions. PNAS. https://doi.org/10.1073/pnas.2400917121

Hu, J., & Levy, R. (2023). Prompting is not a substitute for probability measurements in large language models. EMNLP. https://doi.org/10.18653/v1/2023.emnlp-main.306

Pragmatic communication

How do people reason about other agents to convey meaning beyond what is literally said? How does communication relate to social intelligence more broadly?

Show sample publications

Tsvilodub, P., Klumpp, J.-F., Mohammadpour, A., Hu, J., & Franke, M. (2026). On Emergent Social World Models - Evidence for Functional Integration of Theory of Mind and Pragmatic Reasoning in Language Models. ACL. https://arxiv.org/abs/2602.10298

Hu, J., Floyd, S., Jouravlev, O., Fedorenko, E., & Gibson, E. (2023). A fine-grained comparison of pragmatic language understanding in humans and language models. ACL. https://aclanthology.org/2023.acl-long.230

Hu, J., Levy, R., Degen, J., & Schuster, S. (2023). Expectations over Unspoken Alternatives Predict Pragmatic Inferences. TACL. https://doi.org/10.1162/tacl_a_00579

Zaslavsky, N., Hu, J., & Levy, R. P. (2021). A Rate-Distortion view of human pragmatic reasoning. SCiL. https://aclanthology.org/2021.scil-1.32

Learning from language

What forms of knowledge can be learned through experience with language? What kinds of input data, learning objectives, and inductive biases support the acquisition of linguistic knowledge?

Show sample publications

Hu, J., Wilcox, E. G., Song, S., Mahowald, K., & Levy, R. P. (2026). What Can String Probability Tell Us About Grammaticality? TACL. https://doi.org/10.1162/TACL.a.611

Ivanova, A. A., Sathe, A., Lipkin, B., Kumar, U. U., Radkani, S., Clark, T. H., Kauf, C., Hu, J., Pramod, R. T., Grand, G., Paulun, V. C., Ryskina, M., Akyürek, E., Wilcox, E. G., Rashid, N., Choshen, L., Levy, R., Fedorenko, E., Tenenbaum, J., & Andreas, J. (2025). Elements of World Knowledge (EWoK): A Cognition-Inspired Framework for Evaluating Basic World Knowledge in Language Models. TACL. https://doi.org/10.1162/TACL.a.38

Hu, J., Gauthier, J., Qian, P., Wilcox, E., & Levy, R. (2020). A Systematic Assessment of Syntactic Generalization in Neural Language Models. ACL. https://doi.org/10.18653/v1/2020.acl-main.158

AI metacognition

What would it mean for an artificial model to regulate, monitor, or report on its internal states? How can we design evaluations that rigorously measure these abilities?

Show sample publications

Song, S., Hu, J., & Mahowald, K. (2025). Language Models Fail to Introspect About Their Knowledge of Language. COLM. https://openreview.net/forum?id=AivRDOFi5H

Song, S., Lederman, H., Hu, J., & Mahowald, K. (2025). Privileged Self-Access Matters for Introspection in AI. arXiv. https://arxiv.org/abs/2508.14802