Rules only
FloorMaintainer
The zero-cost rules-only baseline, run by the benchmark harness.
- Vendor
- OpenDisclosure
- Provider
- None
- Region
- Not applicable
- Version
- pilot 2026-10-14
- Composition
- OpenDisclosure's rule engine without a model
F2 score
16.3%95% CI 10.2%–22.1%Leak recall
13.5%95% CI 8.4%–18.5%Over-redaction
4.1%95% CI 2.2%–7.1%Output precision
64.4%95% CI 54.5%–70.7%Cost per 1,000 pages
EUR 0.0095% CI not applicable · measured list-price costRecall by bucket
Universal identifiers44.6%95% CI 30.4%–57.0%
Names7.3%95% CI 4.7%–11.2%
Contextual personal data5.2%95% CI 1.4%–12.3%
Breakdowns
Groups below the reporting floor are not shown.
| Language | Documents | F2 | Recall | Over-redaction |
|---|---|---|---|---|
| de | 24 | 17.9% | 14.9% | 11.3% |
| en | 24 | 30.1% | 25.8% | 7.8% |
| es | 22 | 5.8% | 4.7% | 0.7% |
| fr | 20 | 30.0% | 25.6% | 6.5% |
| it | 23 | 7.6% | 6.2% | 6.6% |
| nl | 22 | 12.3% | 10.1% | 3.7% |
Provenance
- Dataset version and split
- 1.0.0-r3+labels.20261006.1 · tuning
- Scoring
- 1.3
- Harness version
- pilot-2026-10-15
- Run date
- 2026-10-14
- Documents
- 135
- Pages
- 892
- Repeats
- 1
- Price date
- 2026-09-29
- Spend
- EUR 0.00
- Credits
- None · paid by OpenDisclosure
Cite this
@misc{eu-redaction-rules-only-no-roster,
author = {OpenDisclosure},
title = {EU Redaction Benchmark: Rules only},
year = {2026},
url = {https://opendisclosure.eu/benchmark/systems/rules-only/no-roster/},
note = {r3-tuning-pilot-2026-10-15; tuning split; scoring 1.3}
}Download JSON