Fetching from the wire…
Security2026-08-30 · source-backed
A 170-artifact, 145-family benchmark scored three scanners on whether they return a decision at all. ModelAudit produced a definitive verdict on 100% of the 135 labeled families, Fickling on 81.5%, ModelScan on 49.6%. But ModelScan reached 100% precision, recall and F1 on the subset it judged. (arXiv 2608.27424) Treat an N/A as unscanned rather than clean, and read a scanner's coverage rate before you trust its published F1 on a supply-chain gate.
Each link below shares sources, entities, or timing with this story.
Shared entity: Treat / Same source domain / Earlier coverage / Downstream implication
Both cover Treat; reported by the same outlet (arxiv.org); earlier Treat coverage from 2026-07-28.
Shared entity: Model / Same source domain / Earlier coverage / Tension
Both cover Model; reported by the same outlet (arxiv.org); earlier Model coverage from 2026-03-04.
Shared entity: Treat / Same source domain / Earlier coverage
Both cover Treat; reported by the same outlet (arxiv.org); earlier Treat coverage from 2026-08-16.
Both cover Treat; reported by the same outlet (arxiv.org); earlier Treat coverage from 2026-08-13.
Shared entity: Model / Same source domain / Earlier coverage
Both cover Model; reported by the same outlet (arxiv.org); earlier Model coverage from 2026-08-11.
Both cover Model; reported by the same outlet (arxiv.org); earlier Model coverage from 2026-08-05.
Both cover Model; reported by the same outlet (arxiv.org); earlier Model coverage from 2026-08-02.
Shared entity: Treat / Same source domain / Earlier coverage
Both cover Treat; reported by the same outlet (arxiv.org); earlier Treat coverage from 2026-07-23.