citation_prediction_open
- Task Description: Given a legal statement, predict the name of the case which best supports the statement.
- Task Type: Open generation
- Document Type: judicial text
- Number of Samples: 55
- Input Length Range: 20-95 tokens
- Evaluation Metrics: accuracy (maximize)
- Tags: judicial text, open generation, legal knowledge
- Paper: LegalBench: A Collaboratively Built Benchmark for Measuring Legal Reasoning in Large Language Models
- Dataset Download: https://hazyresearch.stanford.edu/legalbench/
7 submissions
| Rank | Model | accuracy | Date | Results |
|---|---|---|---|---|
| 1 | claude-3-5-haiku-20241022 | 0.113 | 2025-08-01 | View |
| 2 | gpt-4o-mini | 0.075 | 2025-07-02 | View |
| 3 | gpt-4.1-nano | 0.038 | 2025-07-03 | View |
| 4 | google/gemma-2-27b-it | 0.019 | 2025-07-24 | View |
| 5 | claude-3-haiku-20240307 | 0.019 | 2025-07-25 | View |
| 6 | meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo | 0.019 | 2025-07-24 | View |
| 7 | meta-llama/Meta-Llama-3.1-405B-Instruct-Turbo | 0.019 | 2025-07-28 | View |