LawBench

LawBench is an evaluation benchmark designed to assess the comprehensive performance of Large Language Models (LLMs) in highly specialized legal domains. It consists of 20 legal tasks and 10,000 evaluation questions, covering various natural language processing (NLP) tasks across multiple legal subfields. LawBench is targeted at researchers and developers who want to benchmark the legal reasoning and knowledge of LLMs. It provides performance comparisons and downloadable results.