GSPC-SWARM

Statutory anchor

multi-agent coordination safety

How it is graded

Deterministically. A regex extracts the label, scored by macro-F1 — no model judges another model. A response with no readable label is reported as UNMEASURED and is excluded from the denominator; it is never scored as a wrong answer. That distinction separates "the model was wrong" from "the model never answered", and it is enforced in the harness code, not just claimed in prose.

What may be quoted

This axis is named and dated but not built. It has no item bank. It stays here, empty and honest, rather than being filled with invented items to round the suite out.

At a glance