← Back to papers

A foundation model to predict and capture human cognition

★ ★ ★ ★ ☆

Paper Summary

Paperzilla title
Centaur: The AI Psychologist That Predicts Your Choices (But Doesn't Know Why)

Researchers developed "Centaur," an AI model trained on a massive dataset of human choices across 160 psychology experiments. Centaur accurately predicts human behavior in new tasks and even shows alignment with brain activity, suggesting a step towards a unified theory of cognition.

Explain Like I'm Five

Scientists built a smart computer program that learned how people make decisions by watching them in many games. Now, it can guess what choices you might make, even in new games!

Possible Conflicts of Interest

None identified

Identified Limitations

Limited generalizability
The generalizability of the model needs to be further validated on a larger variety of tasks and populations.
Lack of interpretability
Although the model captures human behavior across many experiments, the reasons for this predictive ability remains unclear.
Dataset bias
The current version of the dataset has a bias towards specific cognitive domains and demographic groups.

Rating Explanation

This study presents a novel and promising approach to building a domain-general model of human cognition. Although the model exhibits several limitations, it achieves a remarkable level of performance across various benchmarks. Its potential for future research in cognitive science is substantial.

Good to know

This is the Starter analysis. Paperzilla Pro fact-checks every citation, researches author backgrounds and funding sources, and uses advanced AI reasoning for more thorough insights.

Explore Pro →

Topic Hierarchy

Domain: Social Sciences
Field: Psychology

File Information

Original Title: A foundation model to predict and capture human cognition
Uploaded: July 08, 2025 at 11:35 AM
Privacy: Public