When do Word Embeddings Accurately Reflect Surveys on our Beliefs About People?

• https://www.aclweb.org/anthology/2020.acl-main.405/

A somewhat similar idea to Kozlowski2019geometry, but it’s about bias about people, to study whether the preconceptions and stereotypes that we have about certain groups of people (e.g. “thugs”, “judge”, etc.) are reflected in Word embedding.

Maybe what they measure is closer to Implicit association rather than “belief”?