← AI Techniques Radar
Agents & tool use

UI-Driven Import of Hub Datasets into an Annotation Space for Human Feedback Collection

Use a data-centric annotation UI to import a public dataset from a Hub repository into an annotation workspace, automatically generate an initial configuration from the dataset features, define feedback questions (e.g., labels, ratings, rankings, free-form text), and then collect human feedback through the same interface for downstream fine-tuning or evaluation workflows.

Discovered 2026-11-10 · Sources checked 2026-11-10

This summary was generated with AI assistance.

Why it matters

The approach lowers the barrier to creating high-quality datasets by letting domain experts contribute feedback without writing code. It also enables community or team curation by integrating the annotation workspace with Hub-hosted datasets and sharing it via a public workspace URL when desired.

What changed

Argilla adds a workflow where you can start from a dataset on the Hugging Face Hub by using the UI (an “Import dataset from Hugging Face” step), configure questions/fields in real time, and create an annotated dataset directly from that configuration (with Hub datasets required to be public in the described first version).

Where it helps

  • Import a public Hub dataset into an annotation space for community feedback
  • Upload a new CSV to the Hub, import it into the annotation UI, and start labeling
  • Curate an existing Hub dataset for model fine-tuning or evaluation
  • Improve an existing dataset by collecting additional human feedback
  • Set up an annotation workspace that is open for contributions via OAuth

Evidence & reproduction

Reproduction not evaluated. A linked paper or repository does not establish independent reproduction.

Limitations & open questions

In the described first version, the Hub dataset being imported must be public. Additional configuration is mentioned for restricting annotation to specific collaborators, but no further details are provided here.

Sources