Skip to content

Dataset · robbd

Cleaned Epstein documents corpus after OCR

robbd/epstein-index

Machine-readable texts from public court and government archives for information retrieval and feature extraction.

What's inside

The corpus includes 10k to 100k cleaned texts obtained via OCR recognition of public documents related to Jeffrey Epstein. It features U.S. Department of Justice publications (EFTA sets 1-12), House Oversight Committee materials, court records, and agency FOIA responses (including the FBI and other services).

What you can do with it

  • Building semantic search and RAG systems over large legal archives
  • Extracting named entities and relations (feature-extraction) from court documents
Tasks
Embeddings, Text retrieval
Size
10K–100K rows
Downloads this month
25.1K
Likes
0

How to load it

from datasets import load_dataset

ds = load_dataset("robbd/epstein-index")