Document Embeddings Using Sentence Transformers

Hello, I am working with SPECTER, a BERT model that generates document embeddings. When I am using SentenceTransformers to load in the model, and when I do .encode on a text such as a PDF file, I generate an embedding for the file. However, I do not know if Sentence Transformers and SPECTER are reading the document as a document with context and not just as one sentence. Can SentenceTransformers (with the SPECTER model) work on document-level embeddings or is it only for individual sentences?