NeurIPS 2026 Evaluations & Datasets Hosting Guidelines

Please note: These guidelines will be updated on a rolling basis until April 5, 2026 (OpenReview portal opens for submission). Authors should revisit this page regularly to avoid misunderstandings or missing important updates.

This webpage serves as an instruction guide for authors making a submission to the Evaluations and Datasets Track. Authors of datasets will be required to make their dataset available along with Croissant machine-readable metadata in order to streamline the review process and meet industry standards for high quality documentation, reproducibility, and accessibility. We are recommending use of preferred hosting platforms - Dataverse, Kaggle, Hugging Face, and OpenML - which will make compliance simple for authors and reviewers. 

The dataset hosting process as part of submitting to the Evaluations and Datasets Track involves:

1. Choosing among 4 options to host your dataset: Harvard Dataverse, Kaggle, Hugging Face, and OpenML
2. Using platform tooling to download the automatically generated Croissant file
3. Complete the Croissant file with Responsible AI (RAI) metadata. We aim to provide additional tooling for this
4. Including a URL to your dataset and uploading the generated Croissant file in OpenReview
5. If your submission is accepted: making your dataset public by the camera ready deadline

Choose a Preferred Hosting Platform

Harvard Dataverse, Kaggle, Hugging Face, and OpenML platforms are the preferred hosting platforms for datasets. These platforms automatically generate a Croissant file and allow us to perform programmatic metadata verification and dataset assessment, which will streamline and standardize the review process. When you make your dataset accessible via one of these platforms, making a submission will be as simple as providing a URL to the dataset and uploading a generated Croissant file.

The table below outlines key platform features to help authors choose where to host their dataset. Authors may make their dataset accessible via more than one platform at any time.

Authors are responsible for reviewing and complying with the Terms of Service of the platform(s) they choose to use.

Please note that these platforms populate only the “core” Croissant fields and you need to ensure that you add the Responsible AI fields (“rai” fields). This can be done manually (see below). We aim to provide tools to simplify the process before the opening of the submission platform and will update these guidelines accordingly.

Alternatives

Self-hosting your Dataset and Other Data Storage Platforms

If you choose NOT to release your dataset via one of these preferred platforms, you can self-host the data or use other platforms, but you will still be required to make your dataset accessible to reviewers via URL (e.g., to a GitHub repo, public cloud bucket, Zenodo, etc.) and manually generate and upload a Croissant file representation of your dataset as part of the OpenReview submission process.

Generating a Croissant File

This Python tutorial demonstrates how to generate a Croissant file. Find more documentation on the Croissant GitHub repository.

You can also try the Croissant editor or run it locally (GitHub).

Please note that these tutorials and editors focus on the “core” Croissant fields and you need to ensure that you add the Responsible AI fields (“rai” fields). In the python tutorial, this is as simple as ensuring that the following fields are filled in during the creation of the “metadata” (where relevant): 

conforms_to=["http://mlcommons.org/croissant/RAI/1.0"], data_collection, data_collection_type, data_collection_missing_data, data_collection_raw_data, data_collection_timeframe, data_imputation_protocol, data_preprocessing_protocol, data_manipulation_protocol, data_annotation_protocol, data_annotation_platform, data_annotation_analysis, annotations_per_item, annotator_demographics, machine_annotation_tools, data_biases, data_use_cases, data_limitations, data_social_impact, personal_sensitive_information, data_release_maintenance_plan.

How to Publish on Preferred Hosting Platforms

This section provides specific guidance and documentation on how to make your dataset and its Croissant metadata file accessible via the preferred hosting platforms: Harvard Dataverse, Kaggle, Hugging Face, and OpenML.

What to Include in Your Submission

When you submit to Open Review, you will be required to provide:

1. A URL to your dataset accessible to reviewers
2. A Croissant metadata file
3. A confirmation that you verified the validity of your Croissant file.

Please use this online tool to verify your Croissant file. This tool will check whether it is valid, sufficiently complete, and whether the data can be automatically accessed and loaded (the latter may not be possible for all datasets). The same information will be made available to reviewers. 

After submissions are closed, if your Croissant file is invalid or if your data is not accessible, your submission may be desk-rejected. Otherwise, reviewers will commence review of your submission paper, dataset, and metadata. 

If your dataset is accepted, you will be required to make it public by the camera ready deadline. Failure to do so may result in removal from the conference and proceedings.

FAQ

Please check the Evaluations and Datasets Track FAQs