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.

M7 · Harmful compliance and over-refusal 1 value

M7 Harmful compliance and over-refusal: values about Kimi K2-Instruct
Evaluation and metricValueDocument
From its own documentKimi K2 technical report · 28 Jul 2025
Promptfoo red-team (no attack)Harmful: pass rate (safe responses)Condition: Basic (no attack)Kimi K2 technical report28 Jul 2025

M8 · Jailbreak robustness 2 values

M8 Jailbreak robustness: values about Kimi K2-Instruct
Evaluation and metricValueDocument
From its own documentKimi K2 technical report · 28 Jul 2025
Promptfoo red-teamHarmful: pass rate (safe responses)Condition: Crescendo multi-turnKimi K2 technical report28 Jul 2025
Promptfoo red-teamHarmful: pass rate (safe responses)Condition: Iterative JailbreakKimi K2 technical report28 Jul 2025

M9 · Prompt injection 1 value

M9 Prompt injection: values about Kimi K2-Instruct
Evaluation and metricValueDocument
From its own documentKimi K2 technical report · 28 Jul 2025
Promptfoo red-team (prompt injection)Harmful: pass rate (safe responses)Condition: Prompt injectionKimi K2 technical report28 Jul 2025