TEXT
COMMERCIAL
LABELED
Contributors type a question they needed answered that week, in whichever language they think in. Paste is blocked in the form, so every entry was keyed in with its typos and local phrasing. Each question carries two things prompt logs lack: why the person needed to know, and two or three criteria a useful answer would have to cover. Reviewers reject textbook questions, copied text and anything with fewer than two criteria; ten per contributor at most keeps the global author pool wide. LMSYS-Chat-1M cannot be redistributed and WildChat has no rubrics; here every row is a prompt with a human rubric attached, under a commercial license.
Volume
50,000+
Questions
Contributors
10,000+
Regions
100+ countries
Updated
What’s in the collection and how it’s delivered.
Data modality
Text
Annotated
Labeled (reason and answer criteria from the asker)
Exclusivity
Non-exclusive rights
Capture tool
Typed in the Kled app (no pasting)
Use cases
Assistant evaluation sets with human rubrics, preference and reward-model data, query understanding
Format
Plain text
Resolution
Plain text, UTF-8
Total volume
50,000+ questions (approx. 250 MB)
Regions
100+ countries
Purity level
99%+ passed QA review
Where this data comes from and how it’s licensed.
Every file comes straight from the contributor who provided it and is rights-cleared before it’s listed, so your license traces back to the source.
Contributor
Provides the data and uploads it straight to Kled.
Kled
Clears the rights before anything is listed.
Your team
Licenses it under the terms below.
License terms
Non-exclusive commercial training license.
Consent
Each contributor opted in to this use when they uploaded.
Specs reflect the current collection and may change as new data is added. Final terms are set in your license agreement.
