Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

LawBench

Benchmarks

Task NameDataset NameSOTA ResultTrend
Legal Knowledge and Reasoning BenchmarkLawBench
Memorization Score69.19
30
Legal EvaluationLawBench
Accuracy62
26
Legal Question AnsweringLawBench revised (test)
ROUGE-144.68
17
Legal ReasoningLawBench
Micro-F170
10
Reading ComprehensionLawBench 2-5
RC-F197.59
8
Legal Charge PredictionLawBench 8 (test)
F1 Score69.51
6
Hallucination DetectionLawBench Open-ended tasks
Average Entropy1.68
5
Legal ReasoningLawBench
Top-1 Accuracy70.1
4
Autonomous LLM Fine-tuningLawBench
Hybrid Score51.2
4
Legal ConsultationLawBench Revised
Legal Soundness4.14
3
Showing 10 of 10 rows