Question: AI Safety Evaluation Frameworks and Red-Teaming Protocols

HELM safety benchmark coverage metric

In audit

1/5 publication-ready

2 verified (technical)

Findings metrics

—

total

—

verified

—

pending

—

rejected