AI Trust Glossary · Canonical Definition
BM Score Dimension - 15%
Anomaly Rate
One of five BM Score dimensions. Measures the frequency of unexpected or deviant behaviors relative to an agent's established baseline performance.
Explanation
An anomaly is any output or action outside the agent's normal operating pattern - not necessarily wrong, but unexpected. High anomaly rates indicate unpredictability. The measure is anomaly count divided by total actions; zero anomalies is suspicious and may indicate measurement error, since real systems have natural variance.
Why it matters
Anomalies are early warning signals. An agent whose anomaly rate rises before a major failure typically showed subtle anomalies weeks earlier. Tracking this dimension catches deterioration before it becomes a crisis.
How Borealis uses it
Reported in telemetry as anomalySummary: { totalActions, anomalyCount }. The scoring engine applies the 15% weight. Layer 2 statistical detection flags agents whose anomaly patterns look artificially uniform - a sign of telemetry gaming.