Kind of, w2v is cosine similarity driven and there embeddings often go up to 300 dimensions.