Add leaderboard database README

README.md

@@ -0,0 +1,38 @@
+---
+pretty_name: HAKARI-Bench Leaderboard Database
+tags:
+- information-retrieval
+- retrieval
+- benchmark
+- duckdb
+- leaderboard
+- evaluation
+---
+
+# HAKARI-Bench Leaderboard Database
+
+This dataset hosts the DuckDB database used by the
+[HAKARI-Bench leaderboard](https://huggingface.co/spaces/hakari-bench/leaderboard).
+It is derived from the raw benchmark result artifacts in
+[`hakari-bench/results`](https://huggingface.co/datasets/hakari-bench/results)
+and is packaged for leaderboard, viewer, notebook, and SQL use.
+
+The database is produced by the HAKARI-Bench implementation in
+[`hakari-bench/hakari-bench`](https://github.com/hakari-bench/hakari-bench).
+Because the benchmark code, schema, and build workflow evolve over time, this
+dataset card intentionally points to the canonical documentation instead of
+duplicating implementation details here.
+
+## File
+
+- `duckdb/hakari_bench.duckdb`: current leaderboard DuckDB database.
+
+## Links
+
+- [Raw result artifacts](https://huggingface.co/datasets/hakari-bench/results)
+- [Leaderboard Space](https://huggingface.co/spaces/hakari-bench/leaderboard)
+- [HAKARI-Bench GitHub repository](https://github.com/hakari-bench/hakari-bench)
+- [DuckDB schema and leaderboard query guide](https://github.com/hakari-bench/hakari-bench/blob/main/docs/duckdb_schema.md)
+- [Evaluation runbook](https://github.com/hakari-bench/hakari-bench/blob/main/docs/evaluation_runbook.md)
+- [New model evaluation and official result submission workflow](https://github.com/hakari-bench/hakari-bench/blob/main/docs/new_model_results_workflow.md)
+- [Contributing result files](https://github.com/hakari-bench/hakari-bench/blob/main/docs/contributing_results.md)