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.

M5 · Honesty and hallucination 2 values

M5 Honesty and hallucination: values about o1
Evaluation and metricValueDocument
From other documents
PersonQAHallucination rateCondition: no browsingOpenAI o3 and o4-mini System Card16 Apr 2025First reportedOpenAI o3 and o4-mini System Card16 Apr 2025First reported
SimpleQAHallucination rateCondition: no browsingOpenAI o3 and o4-mini System Card16 Apr 2025First reportedOpenAI o3 and o4-mini System Card16 Apr 2025First reported

M7 · Harmful compliance and over-refusal 2 values

M7 Harmful compliance and over-refusal: values about o1
Evaluation and metricValueDocument
From other documents
Challenging refusal evalNot unsafe (aggregate)OpenAI o3 and o4-mini System Card16 Apr 2025First reportedOpenAI o3 and o4-mini System Card16 Apr 2025First reported
Standard refusal evalNot overrefuse (aggregate)OpenAI o3 and o4-mini System Card16 Apr 2025First reportedOpenAI o3 and o4-mini System Card16 Apr 2025First reported

M8 · Jailbreak robustness 2 values

M8 Jailbreak robustness: values about o1
Evaluation and metricValueDocument
From other documents
Human sourced jailbreaksnot_unsafeOpenAI o3 and o4-mini System Card16 Apr 2025First reportedOpenAI o3 and o4-mini System Card16 Apr 2025First reported
StrongRejectnot_unsafe (aggregate)OpenAI o3 and o4-mini System Card16 Apr 2025First reportedOpenAI o3 and o4-mini System Card16 Apr 2025First reported

M9 · Prompt injection 1 value

M9 Prompt injection: values about o1
Evaluation and metricValueDocument
From other documents
Instruction hierarchySystem<>user conflictOpenAI o3 and o4-mini System Card16 Apr 2025First reportedOpenAI o3 and o4-mini System Card16 Apr 2025First reported

M10 · Dangerous capabilities and risk determinations 1 value

M10 Dangerous capabilities and risk determinations: values about o1
Evaluation and metricValueDocument
From other documents
PaperBenchReplication scoreOpenAI o3 and o4-mini System Card16 Apr 2025First reportedOpenAI o3 and o4-mini System Card16 Apr 2025First reported