IP-CaRD v1.0: Anonymized Dataset for Intellectual Property Case Retrieval

IP-CaRD v1.0 is an anonymized dataset for intellectual property case retrieval research. This public release provides anonymized case-level metadata, data split information, aggregate annotation statistics, and key reported results associated with the manuscript “Human-Centered Hybrid Intelligence: A Specialized Generalist Model for Intellectual Property Case Retrieval in Judicial Decision Support”. The released JSON file does not redistribute full judicial document texts, case numbers, public-source document names, de-identified factual summaries, query-candidate pair-level labels, case-level triadic annotation text, or expert-level annotation records. These materials are excluded because the underlying judicial materials and annotation records are subject to legal, privacy, and research-use restrictions.