---
dataset_info:
  features:
  - name: text
    dtype: string
  - name: summary
    dtype: string
  splits:
  - name: train
    num_bytes: 1483542519.7144387
    num_examples: 55993
  - name: validation
    num_bytes: 185439503.0714796
    num_examples: 6999
  - name: test
    num_bytes: 185465998.21408162
    num_examples: 7000
  download_size: 904301529
  dataset_size: 1854448020.9999998
configs:
- config_name: default
  data_files:
  - split: train
    path: data/train-*
  - split: validation
    path: data/validation-*
  - split: test
    path: data/test-*
task_categories:
- summarization
language:
- en
size_categories:
- 10K<n<100K
---

# LegalSumm: A Comprehensive Legal Document Summarization Dataset

LegalSumm is a merged dataset combining three prominent legal document collections (GovReport, BillSum, and CaseSumm) to create a comprehensive resource for training and evaluating legal text summarisation models.

## Dataset Composition

**Total samples:** 69,992 document-summary pairs

**Sources:**
- GovReport: 19,466 U.S. government reports (27.8%)
- BillSum: 23,455 U.S. Congressional bills (33.5%)
- CaseSumm: 27,071 U.S. Supreme Court opinions (38.7%)

## Features

- **text**: Original legal document (full text)
- **summary**: Human-written summary of the document

## Preprocessing

Documents have been standardized to ensure consistent formatting across different legal document types.  
All samples have been randomly shuffled and split into training (80%), validation (10%), and test (10%) sets.