Data license clarification for Najd Benchmark: redistribution and evaluation

#2
by ma7dev - opened

Hello,

I am Mazen Alotaibi, writing on behalf of Najd Research (https://huggingface.co/najdresearch and https://github.com/najdresearch). We are documenting the provenance and permissions of our public Arabic/Saudi evaluation dataset, Najd Benchmark: https://huggingface.co/datasets/najdresearch/najd-benchmark.

The selected rows identified below are already included in our historical release. We could not establish applicable redistribution terms from the source card and license paths we checked. Could you point us to the applicable data license, or confirm in writing whether you are authorized to grant permission for us to:

  1. Retain and publicly redistribute these selected questions, answers and labels with source attribution through Hugging Face and reproducible collection scripts.
  2. Normalize their format and use them to evaluate models and publish benchmark results for research.

Please state the required citation/notices, any restrictions or third-party material exclusions, and whether the permission covers the historical publication as well as future use. If you are only hosting the data and cannot grant these rights, please direct us to the rights holder. We will preserve the terms you provide with the source provenance.

Thank you,
Mazen Alotaibi β€” Najd Research
https://najdresearch.com

Affected data

The source identifier above selects the exact affected rows through each case's provenance.sourceId. A public license link covering these data is sufficient if one already exists. If rights belong to another creator, please direct us to them.

Sign up or log in to comment