fin2C: A Manually Curated Financial Text2Cypher Benchmark

fin2C is a manually curated financial Text-to-Cypher benchmark containing 1,500 natural language–Cypher pairs grounded in the Bloom graph. The benchmark covers five query categories of increasing complexity, ranging from node-level retrieval and multi-hop graph traversals to aggregation, ranking, filtering, and advanced Cypher constructs. This Zenodo record contains: fin2C.zip: the complete fin2C benchmark; MTQ-Bloom.zip: a deduplicated version of the original Mind The Query dataset for the Bloom graph; llama3_lora_bloomt2c.zip: the LoRA adapter weights for the fine-tuned LLaMA-3.1-8B model; Bloomt2c.dump: a Neo4j dump of the Bloom graph used for creating fin2C and for running experiments.