Skip to content

Dataset · ddrg

Multilingual corpus of European Union legal documents

ddrg/super_eurlex

1 to 10 million legal acts from the EUR-Lex portal in six European languages across three formats.

What's inside

Collected from the EUR-Lex website, the corpus contains 1 to 10 million full legal documents in English, German, Czech, Bulgarian, Danish, and Greek, organized by language and legal sector. Each document is presented as cleaned text, raw and cleaned HTML (preserving tables), along with associated sector metadata.

What you can do with it

  • Training and fine-tuning language models and embeddings for European law
  • Classifying legal documents by sector and topic
  • Parsing and extracting data from complex legal HTML tables
Tasks
Text classification, Fill mask
Languages
Bulgarian, Czech, Danish, German, Greek, English, Spanish, Estonian and 4 more
Size
1M–10M rows
Downloads this month
3.9K
Likes
4

How to load it

from datasets import load_dataset

ds = load_dataset("ddrg/super_eurlex")