Add anonymous dataset card

README.md

@@ -1,3 +1,37 @@
----
-license: cc0-1.0
----
+---
+pretty_name: LegalCiteBench
+task_categories:
+- text-generation
+- question-answering
+language:
+- en
+tags:
+- legal-nlp
+- legal-citation
+- benchmark
+- llm-evaluation
+- case-law
+license: cc0-1.0
+---
+
+# LegalCiteBench
+
+LegalCiteBench is a benchmark for evaluating citation reliability in legal language models. It covers closed-book citation retrieval, citation completion, citation error detection, case matching, and case verification and correction.
+
+This repository is anonymized for double-blind review. Public release links and full attribution will be provided after the review period.
+
+## Dataset Files
+
+- `data/cat1_citation_retrieval.jsonl`: citation retrieval
+- `data/cat2_citation_completion.jsonl`: citation completion
+- `data/cat3_citation_error_detection.jsonl`: citation error detection
+- `data/cat4_1_case_matching.jsonl`: case matching
+- `data/cat4_2_case_verification.jsonl`: case verification and correction
+
+## Source
+
+The benchmark is constructed from U.S. judicial opinions from the Case Law Access Project. Reference citations are opinion-derived conservative anchors and should not be interpreted as the only legally relevant authorities for open-ended legal research questions.
+
+## Review Notice
+
+This repository is provided for anonymous review. Please do not attempt to identify the authors during the review period.