How ELMo introduced contextual word representations before the transformer era.
embeddings
How Word2Vec learns dense vector representations that capture word meaning.
An accessible explanation of vector embeddings, similarity search, and why embeddings are useful.