This dataset contains 5,082 parsed email threads extracted from OCR'd documents released by the U.S. House Oversight Committee. The emails have been processed using large language models to extract structured information including senders, recipients, timestamps, subjects, and message bodies, with OCR errors corrected and footers removed.
