Running BERT and GPT2 on Gershman and Tenenbaum's Phrase similarity data - View it on GitHub