← Models

Model profile

Spark3.5

2024-01-30release date
1eval lineages

Evidence summary

Published evidence spans 1 evals and 3 of 7 behavior components. Its strongest relative result is CRiskEval (deception_willingness, #7 of 17); its weakest is CRiskEval (improvement_intent, #14 of 17).

Compare this model

Only models sharing at least one published sub-eval are listed.

Official and reference links

Published eval results

Rank is within that sub-eval. Black marks the observed result; the grey dot marks the value implied by the global rank. Values stay on each source’s native scale.

Eval / sub-evalRankValueDistributionBetterSource
CRiskEvaldeception_willingness#7 / 1721.74↓ lowerSource ↗official
CRiskEvaldesire_for_resource#12 / 1738.75↓ lowerSource ↗official
CRiskEvalharmful_goal#13 / 1742.16↓ lowerSource ↗official
CRiskEvalimprovement_intent#14 / 1754.3↓ lowerSource ↗official
CRiskEvalmalicious_coordination#11 / 1715.78↓ lowerSource ↗official
CRiskEvalself_preservation#10 / 1737.19↓ lowerSource ↗official
CRiskEvalsituational_awareness#9 / 1764.56↓ lowerSource ↗official