Princeton University at CogSci 2026

Princeton University researchers presented 16 AI-related papers at the 48th Annual Meeting of the Cognitive Science Society (CogSci 2026) in Rio de Janeiro, Brazil, covering topics from opinion diffusion in human-AI societies to multilingual generalization in LLMs. Notable work includes an oral presentation on automated discovery of psychological representations using language model agents by Daniel Lima Braga, Raja Marjieh, Akshay Kumar Jagadish, and colleagues, and a paper on rational teachers 'lying' to bounded students by Huang Ham, Dilip Arumugam, Carlos Correa, Bonan Zhao, Tom Griffiths, and Natalia Vélez.

The 48th Annual Meeting of the Cognitive Science Society, or CogSci 2026 https://cognitivesciencesociety.org/cogsci-2026/ , was held this week in Rio de Janeiro, Brazil, bringing together researchers in artificial intelligence, linguistics, psychology, neuroscience, and philosophy. Below is a roundup of AI-related work from Princeton researchers showcased at the conference. The full conference program is available here https://cognitivesciencesociety.org/wp-content/uploads/2026/07/Podium-CogSci-2026-program-v106.pdf . Oral Presentation: An Experimental Method to Study Opinion Diffusion in Human-AI Hybrid Societies Authors: Léna Gaubert, Rémi Devaux, Elif Çelen, Raja Marjieh , Diana Mangalagiu, Antoine Jardin, Nori Jacoby Links: Paper https://arxiv.org/pdf/2605.09197 Oral Presentation: Automated Discovery of Psychological Representations using Language Model Agents Authors: Daniel Lima Braga , Raja Marjieh , Akshay Kumar Jagadish · Princeton University, Ilia Sucholutsky, Nori Jacoby, Brenden M. Lake , Tom Griffiths Oral Presentation: Human-like learning in reasoning models: behavioral and neuroimaging evidence Authors: Csaba Botos, Sreejan Kumar , Austin Andrews, Laurence Hunt, Christopher Summerfield, Rui Ponte Costa, Josh Tenenbaum, Marcelo Gomes Mattar, Momchil Tomov Oral Presentation: Rational Teachers Should ‘Lie’ to Bounded Students Authors: Huang Ham , Dilip Arumugam , Carlos Correa , Bonan Zhao , Tom Griffiths , Natalia Vélez Link: Paper https://zhaobn.github.io/assets/pdfs/ham25lie.pdf Oral Presentation: Similarity All The Way Up: Multilingual Generalization in LLMs Relies on Language-Level Similarity Structures Authors: Supantho Rakshit , Adele Goldberg , Henry Conklin Links: Project Website https://adele.scholar.princeton.edu/publications/similarity-all-way-multilingual-generalization-llms-relies-language-level-similarity Poster : Interpretational alignment: How agents learn from physical guidance depends on how they interpret it Authors: Zhuolun Zhong, Ben Prystawski, Sarah Wu, Bella Fascendini , Sepehr Saeedpour, Joseph Austerweil Poster : Identifying Concepts Used by Human-Like Neural Network Chess Engines Authors: Issac Li , Evan Russek , Ionatan Kuperwajs , Tom Griffiths Poster : Signatures of discrete action symbols emerge in a task-optimized neural model Authors: Carlos Correa , Lucas Tian, Yue Liu, Sreejan Kumar , Daniel J. Hanuska, Kedar Garzón Gupta, Xiao-Jing Wang, Winrich A. Freiwald, Josh Tenenbaum, Marcelo Gomes Mattar Poster : An Episodic Memory Model Can Account for Reinforcement Learning in Humans Authors: Javier Masis , Younes Strittmatter , Jon Cohen Poster : Comparing LLM and Human Responses to Human- and AI-labeled Partners During Naturalistic Conversation Authors: Rachel Metzgar , Michael Graziano Poster : Online library learning in human visual puzzle solving Authors: Pinzhe Zhao, Emanuele Sansone, Marta Kryven, Bonan Zhao Links: Paper https://arxiv.org/pdf/2603.23244 Poster : Representational Similarity and Context Inference as a Shared Computational Account for False Memories in Humans Authors: Mohan Gupta , Caleb Kha-Uong , Jordan Taylor , Jon Cohen Poster : A Resource-Rational Analysis of Forgetting in Continual Learning Authors: Shirley Yu , Alexander Ku , Tom Griffiths Poster : Recovering Event Probabilities from Large Language Model Embeddings via Axiomatic Constraints Authors: Jian-Qiao Zhu, Haijiang Yan, Tom Griffiths Poster : Modeling individual differences in learning and memory using large language models Authors: Akshay Kumar Jagadish , Milena Rmus, Marvin Mathony, Eric Schulz Poster : Evolutionary Theory Makes Predictions About Cognition Beyond Rational Optimization Author: Cameron Turner Poster : Meta-Learning Captures Human-Like Geometric Sensitivity Authors: Max Gupta , Declan Campbell , Tom Griffiths Poster : Modeling Context-Sensitive Effects of Inequity on Reinforcement Learning Authors: Huang Ham , Mu-Chen Wang, Joseph Kable, Adrianna Jenkins Poster : Human-AI Synergy Supports Collective Creative Search Authors: Chenyi Li, Raja Marjieh , Haoyu Hu, Mark Steyvers, Katherine Collins, Ilia Sucholutsky, Nori Jacoby Links: Paper https://arxiv.org/pdf/2602.10001 Poster : Steering Risk Preferences in Large Language Models by Aligning Behavioral and Neural Representations Authors: Jian-Qiao Zhu, Haijiang Yan, Tom Griffiths Poster : Serendipity by Design: Evaluating the Impact of Cross-domain Mappings on Human and LLM Creativity Authors: Qiawen Liu, Marina Dubova, Henry Conklin , Takumi Harada, Tom Griffiths Link: Paper https://arxiv.org/html/2603.19087 Poste r: Pipeline for Systematic Curation of Human Experiments Authors: Younes Strittmatter , Tom Griffiths , Akshay Kumar Jagadish Poster : Joint Modeling of Choices and Response Times in Multi-stage Decisions via Likelihood Approximation Authors: Jialin Li, Carlos Correa , Doris Yu, Prakhar Godara, Marcelo Gomes Mattar Poster : A Rational Analysis of the Effects of Sycophantic AI Authors: Rafael Batista , Tom Griffiths Links: Paper https://arxiv.org/pdf/2602.14270 Poster : Are they human? Detecting large language models by probing human memory constraints Authors: Simon Schug , Brenden M. Lake Links: Paper https://arxiv.org/pdf/2604.00016 Poster : People Intuitively Schedule Tasks to Improve Collective Efficiency Authors: Elizabeth Mieczkowski , Cheryl Li , Natalia Vélez , Tom Griffiths Poster : Do Large Language Models Resolve Fairness-Efficiency Trade-offs Like People? Authors: Cheryl Li , Elizabeth Mieczkowski , Veronica Valera , Natalia Vélez , Tom Griffiths Poster : When Topology Matters: Perturbative Analysis of Nonlinear Social Learning on Networks Authors: Burak Aydin , Raja Marjieh , Tom Griffiths Poster : Parallelograms Strike Back: LLMs Generate Better Analogies than People Authors: Qiawen Liu, Raja Marjieh , Jian-Qiao Zhu, Adele Goldberg , Tom Griffiths Links: Paper https://arxiv.org/pdf/2603.19066 Poster : How The Bias-Variance Tradeoff Shapes Human Strategy Selection Authors: Maximilian Maier, Falk Lieder, Tom Griffiths Links: Paper https://osf.io/preprints/psyarxiv/cegxd v2