https://scripties.uba.uva.nl/search?id=715755 Implementation of the Positive Operator Association Test that identifies implicit biases in words embeddings. Based on the IAT and WEAT.
First use: Possibly change directories in run_poat.py and download word embeddings from: https://nlp.stanford.edu/projects/glove/ (This project used the 'Common Crawl' embeddings with 300 dimensions)
WEAT: https://science.sciencemag.org/content/356/6334/183 IAT: https://psycnet.apa.org/fulltext/1998-02892-004.html