A base-sized English embedding model from the E5 family, optimized for text retrieval and semantic similarity tasks.