It OCR's them too
Do the search
[full legal name of service provider site:copyright.gov]
All of those documents are *scans* inside PDFs *not* text inside PDFs. Note that the names are often badly OCR'd but the regular text is very close to perfect, typical of OCR.