WorkFactory: Turning Professional Intents into Executable Work for Scalable Agent Training
READ THE ORIGINAL SOURCEOpen original sourcehuggingface.co · Opens in a new tabWorkFactory is a framework for constructing coherent, executable professional tasks to train working agents. It starts from professional intents, builds an information support graph to guide task instantiation, and uses probe executions to find support failures and guide repairs. The authors construct 112,686 tasks across 612 occupations and fine-tune Qwen models at 4B, 27B, and 35B-A3B scales into the WorkPilot family. All three improve over base models on four professional-work benchmarks, with WorkPilot-35B-A3B rising from 879 to 1565 Elo on GDPval-AA v2.1.
One primary source; other links are context, not independent corroboration.
- Primary sourceHugging Face Daily Papers
- Discovery linkHugging Face
- Linked in sourceView arXiv page
- Linked in sourceView PDF
The paper page is linked. Peer review and findings have not been independently checked by AINewsIQ.
Open full article view →- Original paper+45
- Fetched source page+30
- No editor link review+0
- Original non-discussion link+5
This measures the stored source trail and editorial steps. It does not certify accuracy, peer review, or independent corroboration. Popularity adds no points.
View full scoring →Read the scoring rubric →What the source says
WorkFactory is a framework for constructing coherent, executable professional tasks to train working agents. It starts from professional intents, builds an information support graph to guide task instantiation, and uses probe executions to find support failures and guide repairs. The authors construct 112,686 tasks across 612 occupations and fine-tune Qwen models at 4B, 27B, and 35B-A3B scales into the WorkPilot family. All three improve over base models on four professional-work benchmarks, with WorkPilot-35B-A3B rising from 879 to 1565 Elo on GDPval-AA v2.1.
A READER'S EVIDENCE CHECKHow to read this signalSix questions about claims, sources, risks, and open checks
This is a signal-level guide until claim-level evidence is available. “Not recorded” means the evidence has not been assessed here; it does not mean a result is false. How we check AI news →
- 01Who is making each claim?
The claimant and whether each claim was measured or asserted are not recorded in this public brief.
- 02What was actually tested?
Not recorded per claim. Check the source for sample size, tasks, model and benchmark versions, and reproducibility.
- 03Where are the sources?
Original link: huggingface.co. Brief basis: fetched source page. Primary page date and quote locations are not established in this panel.
- 04What could mislead?
Source claims not independently checked; scaling analyses show model-dependent diminishing returns, and the benchmark results are reported by the authors.
- 05What is at risk if wrong?
Not assessed per claim. Consider safety, privacy, cost, and adoption decisions before relying on the result.
- 06What would change the view?
No claim-specific check is recorded. An inspectable primary method or independent replication could change the assessment.
Key findings from the stored brief
- Constructs 112,686 professional tasks spanning 612 occupations using information support graphs and probe-guided repair.
- WorkPilot models at 4B, 27B, and 35B-A3B scales improve over their base models across four professional-work benchmarks.
- WorkPilot-35B-A3B improves from 879 to 1565 Elo on GDPval-AA v2.1.
- Planning and repair reduce probe-diagnosed task environment defects and improve downstream performance in controlled analyses.
Source claims not independently checked; scaling analyses show model-dependent diminishing returns, and the benchmark results are reported by the authors.
This is a source-based abstract, not independent validation of the source's claims.
Evidence & sources
These links show where the signal came from and what the source page linked to. A linked page is context, not independent corroboration.
The paper page is linked. Peer review and findings have not been independently checked by AINewsIQ.
Timeline & updates
- Signal dated by source or discovery feed
This timestamp places the signal in the radar. It may be the discovery post date rather than the primary page's first publication date.
- Shared brief stored
AINewsIQ prepared one abstract from fetched source page.
A publication revision history is not stored for this signal. New engagement snapshots do not mean the source itself changed.
Why this may matter
Reliable expert demonstrations for professional work are a bottleneck for training agents. A method that generates executable, information-complete tasks at scale could make supervised training of general-purpose agents more practical for real-world professional use.
Limits and open questions
Source claims not independently checked; scaling analyses show model-dependent diminishing returns, and the benchmark results are reported by the authors.
Organization-specific financial, security, and operational effects require the organization's own context and review.
Discussion
Comments are reviewed before publication. Discuss the evidence, not the person.
No published comments yet.
Reader discussion opens with accounts. You can read the original source and published comments now.