Values by metric family

Each value as the document printed it. Select a value for its source, its checks and its history.

KeyVerifiedUnverifiedDisputedCorrected read off a figure or stated in wordsA value opens its source and history.

M1 · Evaluation awareness 2 values

M1 Evaluation awareness: values about Muse Spark 1.1
Evaluation and metricValueDocument
From its own documentMuse Spark 1.1 Evaluation Report · 9 Jul 2026
Verbalized eval awarenessShare of samples verbalizing evaluationCondition: Agentic Misalignment scenariosMuse Spark 1.1 Evaluation Report9 Jul 2026
Verbalized eval awarenessShare of samples verbalizing evaluationCondition: dangerous-capability evalsMuse Spark 1.1 Evaluation Report9 Jul 2026

M3 · Sabotage and sandbagging 1 value

M3 Sabotage and sandbagging: values about Muse Spark 1.1
Evaluation and metricValueDocument
From its own documentMuse Spark 1.1 Evaluation Report · 9 Jul 2026
SHADE-ArenaOverall success under CoT monitorMuse Spark 1.1 Evaluation Report9 Jul 2026

M4 · Misalignment audits 1 value

M4 Misalignment audits: values about Muse Spark 1.1
Evaluation and metricValueDocument
From its own documentMuse Spark 1.1 Evaluation Report · 9 Jul 2026
Agentic MisalignmentHarmful action rateCondition: without system mitigationsMuse Spark 1.1 Evaluation Report9 Jul 2026

M5 · Honesty and hallucination 1 value

M5 Honesty and hallucination: values about Muse Spark 1.1
Evaluation and metricValueDocument
From its own documentMuse Spark 1.1 Evaluation Report · 9 Jul 2026
MASKHonesty rateCondition: without system mitigationsMuse Spark 1.1 Evaluation Report9 Jul 2026

M6 · Sycophancy 1 value

M6 Sycophancy: values about Muse Spark 1.1
Evaluation and metricValueDocument
From its own documentMuse Spark 1.1 Evaluation Report · 9 Jul 2026
Internal sycophancySycophancy rateCondition: without system mitigationsMuse Spark 1.1 Evaluation Report9 Jul 2026

M8 · Jailbreak robustness 1 value

M8 Jailbreak robustness: values about Muse Spark 1.1
Evaluation and metricValueDocument
From its own documentMuse Spark 1.1 Evaluation Report · 9 Jul 2026
StrongREJECTAttack success rateCondition: without system mitigationsMuse Spark 1.1 Evaluation Report9 Jul 2026

M9 · Prompt injection 1 value

M9 Prompt injection: values about Muse Spark 1.1
Evaluation and metricValueDocument
From its own documentMuse Spark 1.1 Evaluation Report · 9 Jul 2026
AgentDojoPrompt-injection ASR pass@1Condition: without system mitigationsMuse Spark 1.1 Evaluation Report9 Jul 2026

M10 · Dangerous capabilities and risk determinations 2 values

M10 Dangerous capabilities and risk determinations: values about Muse Spark 1.1
Evaluation and metricValueDocument
From its own documentMuse Spark 1.1 Evaluation Report · 9 Jul 2026
CybenchPass@1Condition: without system mitigationsMuse Spark 1.1 Evaluation Report9 Jul 2026
Risk determinationRisk tier, cyberCondition: pre-mitigationMuse Spark 1.1 Evaluation Report9 Jul 2026

Risk determinations

The developer's formal decisions about Muse Spark 1.1 under its framework, as printed. Levels from different frameworks do not map onto one another.

Risk determinations about Muse Spark 1.1
DomainLevel as printedFramework and document
CyberHigh cannot be ruled out Advanced AI Scaling Framework Muse Spark 1.1 Evaluation Report · 9 Jul 2026Advanced AI Scaling Framework Muse Spark 1.1 Evaluation Report · 9 Jul 2026