Test fine-grained legal RAG retrieval with highly similar clauses covering five easily confused lease rights.

A structured resource for connecting disputed facts in domestic arbitration with evidentiary support, conflicts, and proof gaps.

Built around key technology commercialization license terms, this dataset supports precise retrieval, structured extraction, and evidence traceability.

Retrieve compensation and benefits clauses with supporting contract evidence to answer employment agreement questions more precisely.

Connect employment clauses with legal authorities and review questions in a structured resource for compliance retrieval and traceable contract review.

Find traceable legal evidence on license terms and termination consequences for contract retrieval and clause analysis.

Train models to interpret patent license scope, commercial terms, and legal obligations through contract-grounded questions and answers.

Evaluate how well models understand commercial litigation documents, from confidentiality and privilege claims to court rulings.

Train and evaluate legal question answering with structured links between case facts, legal rules, reasoning, and conclusions.

Explore the core disputes in patent infringement pleadings through structured claim comparisons, party positions, and defense arguments.

Connect claim amendments to evidence in the original application to train patent QA models that assess support and locate relevant passages.

Train verifiers to trace cross-document evidence in contract packages and identify whether claims are supported or missing critical grounds.

Train models to verify affordable housing policy claims against program rules and paired source evidence.

Connect patent eligibility materials with relevant precedent to evaluate precise legal retrieval in RAG systems.

Test whether models substantiate FTO analysis claims with atomic facts, patent and procedural evidence, and explicit verification labels.

Find evidence for royalties, settlements, audit rights, and other key licensing terms with data designed for traceable legal RAG retrieval.

Train models to interpret trade secret disputes through focused coverage of core legal elements, evidence, and judicial reasoning.

Train models to answer patent portfolio lifecycle questions with evidence grounded in legal status and prosecution records.

Connect planning claims with source-text evidence to enable traceable verification of zoning, land use, and development rules.

Train models to make verifiable patent novelty assessments grounded in single-document prior art and claim-feature mapping.