Behavioral estimates of conceptual structure are robust across tasks in humans but not large language models.

Siddharth Suresh, Kushin Mukherjee, Lisa Padua, Timothy T. Rogers

Browse the full CogSci paper archive.