Epstein Files OCR: Page-Level Text from Early Document Release

Epstein Files OCR — Datasets 1–8 (Early Release) contains page-level OCR output in Markdown format from a public release of documents related to the Jeffrey Epstein case. The dataset is designed for question answering, information retrieval, and text classification tasks. It was created by the author 'ishumilin' and last updated on March 17, 2026.

Use Cases

• Question answering based on the OCR text content of legal documents.
• Text retrieval for finding specific information within the scanned pages.
• Text classification to categorize pages by content type, such as flight logs or court documents.

Strengths

• Dataset is explicitly designed for NLP tasks like question answering and text retrieval.
• Source documents are from a public release related to a specific, well-known legal case.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• The dataset is marked as no longer maintained, directing users to a 'Complete OCR Dataset'.

Provenance

Source

Public release of documents related to Jeffrey Epstein / the Epstein case.

Collection Method

Scanned pages converted to text using an automated OCR pipeline.

Freshness

Last updated 2026-03-17 21:13:00; freshness should be verified.