Evaluations
The evaluation set is your golden record of how this workflow should behave. It grows from every reviewer correction and gates every change.
Evaluation items
38
33 from reviewer corrections
Latest eval pass rate
94.1%▲2.3 pts
over 34 items
Regression runs
2
every candidate is gated by eval
Run history
Each run scores a candidate version against the eval set before promotion.
| Run | Baseline → Candidate | Pass rate | Delta | Items | When |
|---|---|---|---|---|---|
| run_h1 | v1.3v1.4 | 94.1% | ▲2.3 pts | 34 | 5d ago |
| run_h2 | v1.2v1.3 | 91.8% | ▲2.7 pts | 29 | 2w ago |
Coverage by topic
What the eval set tests.
Evaluation set
Golden examples with their expected outcome. Corrections land here automatically.
| Clause | Division | Expected | Tags | Source | Added |
|---|---|---|---|---|---|
| Through-penetration firestop systems shall provide a fire-resistance rating of not less than 2 hours, tested in accordance with ASTM E814. | 07 84 00 · Firestopping | Needs review | fire-ratingdivision-07 | seed | 3w ago |
| Cast-in-place concrete shall attain a minimum 28-day compressive strength of 4,000 psi per ACI 318. | 03 30 00 · Cast-in-Place Concrete | Compliant | concretedivision-03 | seed | 3w ago |
| Aluminum-framed curtain wall shall limit deflection to L/175 of the span under design wind load per ASTM E330. | 08 44 00 · Curtain Wall | Compliant | curtain-walldivision-08 | seed | 3w ago |
| Structural steel welding shall conform to AWS D1.1 by AWS-certified welders. | 05 12 00 · Structural Steel | Needs review | weldingdivision-05 | seed | 2w ago |
| Roof insulation shall provide a minimum thermal resistance of R-30 for the applicable climate zone per ASHRAE 90.1. | 07 22 00 · Roof Insulation | Needs review | insulationdivision-07 | seed | 2w ago |
| Spec clause #100 under Division 05 · compliance requirement against governing standard. | Division 05 | Compliant | weldingdivision-05 | correction | 4w ago |
| Spec clause #101 under Division 07 · compliance requirement against governing standard. | Division 07 | Non-compliant | fire-ratingdivision-07 | correction | 3w ago |
| Spec clause #102 under Division 04 · compliance requirement against governing standard. | Division 04 | Non-compliant | masonrydivision-04 | correction | 3w ago |
| Spec clause #103 under Division 08 · compliance requirement against governing standard. | Division 08 | Non-compliant | curtain-walldivision-08 | correction | 3w ago |
| Spec clause #104 under Division 07 · compliance requirement against governing standard. | Division 07 | Non-compliant | fire-ratingdivision-07 | correction | 3w ago |
| Spec clause #105 under Division 07 · compliance requirement against governing standard. | Division 07 | Needs review | insulationdivision-07 | correction | 3w ago |
| Spec clause #106 under Division 07 · compliance requirement against governing standard. | Division 07 | Needs review | insulationdivision-07 | correction | 3w ago |
| Spec clause #107 under Division 07 · compliance requirement against governing standard. | Division 07 | Compliant | insulationdivision-07 | correction | 3w ago |
| Spec clause #108 under Division 05 · compliance requirement against governing standard. | Division 05 | Compliant | weldingdivision-05 | correction | 2w ago |