Skip to content

Dataset · common-pile

US court decisions: 6.7M cases

common-pile/caselaw_access_project

A corpus of 6.7M US court cases for training and analyzing legal texts.

What's inside

The dataset contains 6.7M court cases from the Caselaw Access Project and Court Listener. The Caselaw Access Project features nearly 40 million pages of federal and state court decisions and judge opinions over the past 365 years. Court Listener adds over 900k cases collected from 479 courts. Language is English, modality is text.

What you can do with it

  • Train a legal LLM on court texts
  • Fine-tune LoRA for precedent analysis
  • Test summarization and QA on court decisions
Tasks
Text generation
Languages
English
Size
1M–10M rows
Downloads this month
7.3K
Likes
229

Models trained on it

How to load it

from datasets import load_dataset

ds = load_dataset("common-pile/caselaw_access_project")