---
license: mit
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
dataset_info:
  features:
  - name: uuid
    dtype: string
  - name: title
    dtype: string
  - name: parties
    dtype: string
  - name: start_date
    dtype: string
  - name: end_date
    dtype: string
  - name: full_text
    dtype: string
  - name: confidentiality
    dtype: string
  - name: privacy
    dtype: string
  - name: tdm
    dtype: string
  - name: pricing
    dtype: string
  - name: content
    dtype: string
  - name: term
    dtype: string
  splits:
  - name: train
    num_bytes: 19795644
    num_examples: 789
  download_size: 7506387
  dataset_size: 19795644
language:
- en
---

This is a dataset for experiments related to legal document processing.  It contains 801 legal contracts from the [California Digital Libary](https://cdlib.org/services/collections/licensed/resources/redacted-license-agreements/) and [SPARC](https://sparcopen.org/our-work/big-deal-knowledge-base/contracts-library/).
The PDF files were converted to markdown using Docling. The markdown headers are used to identify sections of interest.