Fork me on GitHub

About this tool:

This tool measures how much bias exists in popular word embeddings which are used in various domains extensively and is based on this paper.
First you should select two target word groups 1 and 2. Then you should select two attribute word groups A and B. Then we will calculate the difference between the two sets of target words in terms of their relative similarity to the two sets of attribute words. This is measured using a permutation test and if the p value is close to zero, it means that target word similarities to attribute words are different.
You can add words to groups by typing them in related text box and pressing Enter. Remove words from groups by clicking on them.
You can use the following examples for start:

Sample Word Groups:

Target Words Group 1:

Target Words Group 2:

Attribute Words Group A:

Attribute Words Group B: