Research code for evaluating category knowledge acquired by neural networks.
A measure how well pairs of learned representations can be correctly judged to belong to the same category
A measure of how abstract learned representations are.
A measure of spread between learned representations that belong to the same category
A measure of how well the similarity structure of learned representations captures the gold category structure
Developed on Ubuntu 18.04 and Python 3.7