LATIDIA · Ciberseguridad

CascadeLedger: Auditoría contrafactual de la integridad de la información y los recursos en cascadas de LLM

arXiv: 2610.20668v1Tipo de anuncio: nuevo Resumen: Presentamos CascadeLedger, un marco de auditoría de ejecución contrafactual para cascadas de modelos de lenguaje presupuestados. El protocolo aplica la semántica de llamada antes de leer, vincula el éxito

La noticia

arXiv:2610.20668v1 Announce Type: new Abstract: We introduce CascadeLedger, a counterfactual execution-audit framework for budgeted language-model cascades. The protocol enforces call-before-read semantics, binds successful reads to service and cache identities, records availability and component-wise budget transitions, and freezes the event trace before any target or task scorer is joined. The central object is aggregate-equivalent trace validity: two executions can share a final answer, declared resource projection, and service histogram while differing in whether their routing decisions used only legitimately observable information. We complement single-trace invariants with counterfactual prefix non-interference (CPNI), which couples a visible prefix and policy randomness while changing only a declared hidden family. The counterfactual influence frontier (CIF) localizes