BenchmarkLawExamQA Civil
LawExamQA_Civil
A legal question answering evaluation dataset for the civil law domain, used to assess the accuracy of AI models on legal reasoning and examination questions. Exact match is used as the primary evaluation metric.
Timeline (last 90 days)
Sep 24
On the LawExamQA_Civil dataset, LEGO achieved an exact match accuracy of 40.53%, surpassing the RAG and CoT baseline methods evaluated.
Unverified50%