LegalPT aggregates the maximum amount of publicly available legal data in Portuguese, drawing from varied sources including legislation, jurisprudence, legal articles, and government documents.
This version is deduplicated using MinHash algorithm and Locality Sensitive Hashing, following the approach of Lee et al. (2022).
The raw version is also available here.
Dataset is composed by six corpora:
Ulysses-Tesemõ, MultiLegalPile (PT)… See the full description on the dataset page:
https://huggingface.co/datasets/eduagarcia/LegalPT_dedup.
