Domaines
Statistical physics
Type of internship
Théorique, numérique Description
Our aim is to evaluate how well large language models can predict human judgments of harmonious color pairings and palette completion. We will benchmark several models against existing human survey data using different color encodings (e.g., HEX, HSL, CIELab). Ultimately, the goal is to quantify agreement, analyze failure cases, and inform both cognitive science and practical color recommendation tools.
Contact
Michael Benzaquen