hearsay
- Task Description: Classify if a particular piece of evidence qualifies as hearsay.
- Task Type: Binary classification
- Document Type: fact pattern
- Number of Samples: 99
- Input Length Range: 20-69 tokens
- Evaluation Metrics: accuracy (maximize), balanced_accuracy (maximize), f1_macro (maximize), f1_micro (maximize), valid_predictions_ratio (maximize)
- Tags: evidence law, rule application
- Paper: LegalBench: A Collaboratively Built Benchmark for Measuring Legal Reasoning in Large Language Models
- Dataset Download: https://hazyresearch.stanford.edu/legalbench/
19 submissions
| Rank | Model | accuracy | balanced_accuracy | f1_macro | f1_micro | valid_predictions_ratio | Date | Results |
|---|---|---|---|---|---|---|---|---|
| 1 | claude-opus-4-1-20250805 | 0.915 | 0.919 | 0.914 | 0.915 | 1.000 | 2025-08-05 | View |
| 2 | claude-opus-4-20250514 | 0.904 | 0.907 | 0.903 | 0.904 | 1.000 | 2025-07-25 | View |
| 3 | grok-3-mini | 0.883 | 0.885 | 0.882 | 0.883 | 1.000 | 2025-07-29 | View |
| 4 | claude-sonnet-4-20250514 | 0.862 | 0.871 | 0.860 | 0.862 | 0.926 | 2025-07-25 | View |
| 5 | grok-4-0709 | 0.851 | 0.838 | 0.844 | 0.851 | 1.000 | 2025-07-30 | View |
| 6 | o4-mini | 0.830 | 0.810 | 0.818 | 0.830 | 1.000 | 2025-07-08 | View |
| 7 | o3 | 0.798 | 0.777 | 0.782 | 0.798 | 1.000 | 2025-07-08 | View |
| 8 | gpt-5-2025-08-07 | 0.787 | 0.762 | 0.767 | 0.787 | 1.000 | 2025-08-08 | View |
| 9 | openai/gpt-oss-120b | 0.777 | 0.763 | 0.767 | 0.777 | 1.000 | 2025-08-05 | View |
| 10 | deepseek-ai/DeepSeek-V3 | 0.755 | 0.733 | 0.737 | 0.755 | 1.000 | 2025-07-10 | View |
| 11 | claude-3-5-haiku-20241022 | 0.745 | 0.768 | 0.743 | 0.745 | 1.000 | 2025-08-01 | View |
| 12 | deepseek-ai/DeepSeek-R1 | 0.740 | 0.724 | 0.727 | 0.740 | 0.802 | 2025-07-11 | View |
| 13 | gpt-4o-mini | 0.734 | 0.723 | 0.725 | 0.734 | 1.000 | 2025-07-02 | View |
| 14 | meta-llama/Meta-Llama-3.1-70B-Instruct-Turbo | 0.734 | 0.703 | 0.702 | 0.734 | 1.000 | 2025-07-25 | View |
| 15 | o3-mini | 0.723 | 0.699 | 0.700 | 0.723 | 1.000 | 2025-07-09 | View |
| 16 | meta-llama/Meta-Llama-3.1-405B-Instruct-Turbo | 0.723 | 0.691 | 0.687 | 0.723 | 1.000 | 2025-07-30 | View |
| 17 | google/gemma-2-27b-it | 0.713 | 0.709 | 0.709 | 0.713 | 1.000 | 2025-07-24 | View |
| 18 | claude-3-haiku-20240307 | 0.691 | 0.707 | 0.691 | 0.691 | 1.000 | 2025-07-25 | View |
| 19 | gpt-4.1-nano | 0.617 | 0.636 | 0.615 | 0.617 | 1.000 | 2025-07-03 | View |