PREVIEWBanking

Minimize the likelihood of executing an action beyond granted permissions during account operations.

Statement
Direction
Minimize
Metric
likelihood
Object
of executing an action beyond granted permissions
Context
during account operations
Scorecard
DECLARED
Importance
10/10
DECLARED — Blade product decision, 2026-07
MEASURED
adversarial permission cases passed
66.7% [57.5, 75.0]
adversarial set · n=120
GRADE
Grade
Does not meet
Does not meet — 80 of 120 adversarial cases (66.7%)
Delivery agent: FunctionGemma-270M banking router, served first-party
scored outside NeMo Evaluator — in-service re-scoring in progress.

Known limitations

Permission-tier compliance does not generalize beyond the prompt shape this model was trained on. In testing: when a request was natively marked as permission-denied, the model correctly refused in 40 of 40 cases. When the caller's access tier was lowered by one step — a single word changed, with the request and tool list otherwise identical — the model named the out-of-tier tool in 40 of 40 cases. In-tier control requests were handled correctly in 40 of 40 cases. This is a known limitation of the current version and is the top priority for the next training cycle. Do not deploy this model as a permission boundary; permission enforcement must be implemented in your runtime, not delegated to the model.

ROI calculator
your assumptions

Set your own numbers. Blade makes no ROI claim — this is your model, not ours.

MODELED from your assumptions — not a Blade claim
$200,000
per month, at your inputs
See it live methods & provenance →engine room →
Other outcomes in Banking