Apps · PROJECT NOTES
DABstep
Multi-file data-analysis tasks combining payment records, JSON business data and documentation, with a public code-executing agent baseline and answer scorer.
What can you do with it?
Evaluate data joins, aggregation, business-rule interpretation and precise result reporting across heterogeneous files; local scoring is available when reference answers are provided.
What should you know before using it?
The public default file has 450 questions with empty answer fields; the 10 answered dev questions share six of those IDs. Scoring checks final answers, not notebook validity or GUI completion. Harbor reconstructs default labels from scored public submissions, so its scores are not established as equivalent to official hidden-answer evaluation. The current Space closes new manual validation requests while keeping unvalidated submissions open.
What is verified here?
Runtime: The author baseline documents Python 3.10, pinned smolagents dependencies, a model endpoint and seven shared context files. It executes Python for analysis; source review did not verify sandbox isolation. The Space includes the scorer and leaderboard service, not an official locally complete hidden-answer test set.
License note: Dataset and Space LICENSE files declare CC BY 4.0. Space setup.py instead lists an MIT classifier; package-code licensing metadata is inconsistent and should not be presented as uniformly resolved. Upstream license ↗
This entry was reviewed against original project material. We have not independently run the software or reproduced its results. AI-development claims are author-reported; check upstream for current versions, setup requirements, and licenses.
- Resource type
- Dataset + benchmark
- Provenance
- Development provenance not established
- Reviewed
- Platforms
- Consult upstream documentation