Publish Dataset

← Back to admin

Package this project's annotations into a well-documented, reusable dataset and publish it to the HuggingFace Hub, deposit it on Zenodo for a citable DOI, or download a self-contained archive. A rich README (dataset card) with label distributions and inter-annotator agreement is generated automatically — the same numbers you get from Paper Mode.

1 · Target

2 · Preprocessing

Anonymize annotators (A1, A2 …) & strip internal fields Include raw per-annotator split Include aggregated gold split

Aggregation

Min annotators / item

Include spans split Include source items split Include survey / phase responses (may contain PII) Scrub emails / phone numbers from item text Bundle referenced media files

Train/val/test split

Data format

3 · Metadata

Title

Description

License

Version

Authors (one per line: Name; Affiliation; ORCID)

Keywords (comma-separated)

Citation (BibTeX)

Preview

Click “Preview README” to render the dataset card and split counts.

1 · Target

2 · Preprocessing

3 · Metadata

Preview

Result