The word count extractor is an example of an extractor in Stardog BITES framework. BITES works on files ingested by the BITES API, not on SPARQL query results. Although we do support calling named entity recognition as a SPARQL service, this is not possible with other BITES extractors.
If you have already ingested a document with BITES using the word count extractor, you can query the extracted RDF which is contained in a named graph for each document. The following query will return all BITES documents and their word count: