Versions

One version is on record: the copy we retrieved on 26 Sep 2026. We know of no other.

  1. 26 Sep 2026

    Date retrieved

    Copy retrieved 26 Sep 2026

    Date from
    the date we retrieved it; the copy states no version date
    Copy
    Copy retrieved on 26 Sep 2026
    Values
    10 values recorded from this version

No version has a file hash or an archived snapshot yet. From dataset v0.2 each retrieved version carries both (Methodology §7).

Revisions

No revisions are recorded for this document. We know of only one version of it.

Values

Every value we recorded from this document, grouped by metric family and ordered by where the document prints it. Location is the section, table or page as the document numbers it. None has been blind-verified yet.

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 in the Grok 4 Fast Model Card
ModelEvaluationConditionValueLocationChecked
Grok 4 Fast (non-reasoning)MASKDishonesty ratenon-reasoningTable 2Unverified
Grok 4 Fast (reasoning)MASKDishonesty ratereasoningTable 2Unverified

M6 Sycophancy

1 value

M6 Sycophancy: values in the Grok 4 Fast Model Card
ModelEvaluationConditionValueLocationChecked
Grok 4 Fast (reasoning)SycophancySycophancy ratereasoningTable 2Unverified

M7 Harmful compliance and over-refusal

1 value

M7 Harmful compliance and over-refusal: values in the Grok 4 Fast Model Card
ModelEvaluationConditionValueLocationChecked
Grok 4 Fast (reasoning)AgentHarmAnswer rate on harmful agentic tasksreasoning; no attackTable 1Unverified

M8 Jailbreak robustness

2 values

M8 Jailbreak robustness: values in the Grok 4 Fast Model Card
ModelEvaluationConditionValueLocationChecked
Grok 4 Fast (non-reasoning)JailbreaksAnswer rate on should-refuse promptsnon-reasoning; system prompt jailbreakTable 1Unverified
Grok 4 Fast (reasoning)JailbreaksAnswer rate on should-refuse promptsreasoning; system prompt jailbreakTable 1Unverified

M9 Prompt injection

1 value

M9 Prompt injection: values in the Grok 4 Fast Model Card
ModelEvaluationConditionValueLocationChecked
Grok 4 Fast (non-reasoning)AgentDojoPrompt-injection attack success ratenon-reasoning; with system prompt mitigationTable 1Unverified

M10 Dangerous capabilities and risk determinations

3 values

M10 Dangerous capabilities and risk determinations: values in the Grok 4 Fast Model Card
ModelEvaluationConditionValueLocationChecked
Grok 4 Fast (reasoning)BioLP-BenchAccuracyreasoning; without safeguardsTable 3Unverified
Grok 4 Fast (reasoning)CybenchUnguided success ratereasoning; without safeguardsTable 3Unverified
Grok 4 Fast (reasoning)VCTAccuracyreasoning; without safeguardsTable 3Unverified

Extraction coverage

What we read of this document, and where each value was read.

Our note We have not written a coverage note for this document yet.

All 10 values were read in the document itself.

All 10 values were extracted for version 0 of the dataset through a web reader, which did not always reach the later sections of long PDFs (Methodology §3).