53,000+ multiple choice questions focused on identifying the relevant legal holding of a cited case. The dataset targets legal reasoning and natural language understanding within the domain of judicial decisions.
Use Cases
- Train a legal NLP model to identify the correct holding of a case using the multiple-choice options
- Evaluate the legal reasoning capabilities of LLMs on questions regarding cited cases
- Benchmark models on their ability to interpret the relevant holding from a set of multiple-choice answers
Strengths
- 53,000+ multiple choice questions
- Focuses on identifying the 'holding' of a cited case
- Multiple-choice format for automated evaluation