The Washington law Benchmark (WLB)
Dataset Summary
The Washington Law Benchmark (WLB) is a large-scale, synthetic dataset designed specifically to advance Legal Information Retrieval (IR) and Semantic Search. It bridges the critical "semantic gap" between natural language (how citizens, local governments, and plain-English users describe legal scenarios) and formal statutory legalese (how laws are actually written).
The dataset contains hundreds of thousands of… See the full description on the dataset page: https://huggingface.co/datasets/CSI-lab/RCW_2025_Positive_Query_Pairs.