This is a language model that can be used to obtain document embeddings suitable for downstream tasks like semantic search and clustering.