LegalSearchQA: Multiple-Choice Benchmark for Legal Knowledge Retrieval

LegalSearchQA is a benchmark designed to evaluate retrieval and reasoning over current U.S. and international law. Each question requires interpreting recent legal sources like statutes and court decisions that may have changed after common LLM training cutoffs. The dataset was created by author 'boqiny' and last updated on March 29, 2026.

Use Cases

• Benchmarking legal knowledge retrieval systems based on the requirement to find recent legal sources.
• Evaluating large language model reasoning on current law based on the multiple-choice question format.
• Testing the temporal robustness of legal AI systems based on questions referencing sources that change over time.

Strengths

• Focuses on current legal sources that may have changed after common LLM training cutoffs.
• Designed for evaluating both retrieval and reasoning, as indicated by the description.
• Last updated on March 29, 2026, suggesting recent maintenance.

Limitations

• Description metadata is limited; actual data quality requires manual inspection after download.
• Column-level documentation is absent; field semantics must be inferred after download.
• Row count and file size are unknown, which may limit suitability assessment.

Provenance

Source

huggingface

Collection Method

Method of gathering is not described in the provided input.

Time Range

Covers recent legal sources; specific temporal range is not provided.

Freshness

Last updated 2026-03-29 20:23:05.

Geography

Covers U.S. and international law.