Benchmark dataset for performing natural language (NL) to MongoDB Shell (mongosh) code generation.
There is an emerging desire from users for NL query generation.
This benchmarks examines how LLMs generate MongoDB queries and provides proactive guidance for making systems that map NL to MongoDB queries.
Benchmark dataset (flat CSV file, Braintrust evaluation… See the full description on the dataset page:
https://huggingface.co/datasets/mongodb-eai/natural-language-to-mongosh.