Apps · PROJECT NOTES
OSWorld-Human
Human reference trajectories and efficiency scoring for OSWorld tasks, measuring how many agent actions a successful workflow requires.
What can you do with it?
Compare agent action efficiency against annotated human trajectories using the released weighted-efficiency scorer.
What should you know before using it?
The authors explicitly say not to use these data for training because the trajectories expose OSWorld evaluation solutions. This adds reference data and a scorer, not new desktop images. Report the OSWorld task version, maximum steps and single-action versus grouped-action convention. No explicit repository license was located.
What is verified here?
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