JS 1.0 · Evidence view
Explore the published recurrence evidence
This Phase 1 view preserves the release’s central evidence, exact values, provenance, and interpretation boundaries. The configurable multi-view workbench follows in Phase 2.
Release JS 1.0Dataset 1.0.0Aggregation Unique signaturesRuns represented 250 model runs
Published default view
Stable matches, variable extras
134 of 158 reference-matched signatures recurred in all five runs. Half of unique unmatched signatures appeared only once.
- 134 of 158Reference-matched findings seen in all five runsInspect
- 80 of 161Unmatched findings seen in only one runInspect
- 22 of 161Unmatched findings seen in all five runsInspect
134 of 158 · Reference-matched findings seen in all five runs
View exact recurrence values
| Finding group | Count | Share |
|---|---|---|
| Reference-matched findings seen in all five runs | 134 of 158 | 84.8% |
| Unmatched findings seen in only one run | 80 of 161 | 49.7% |
| Unmatched findings seen in all five runs | 22 of 161 | 13.7% |
Metric definition
Finding recurrence counts the identical repetitions in which a normalized finding signature appeared. Signature normalization and reference matching follow the documented JS 1.0 scorer.
Caveat: unmatched is not a synonym for false positive, and reference-matched is not a synonym for independently adjudicated true positive.