← Models

Model profile

Opt 66B

2022-06-23release date
2eval lineages

Evidence summary

Published evidence spans 2 evals and 1 of 7 behavior components. Its strongest relative result is ChineseSafe (score, #9 of 22); its weakest is HELM Classic RealToxicityPrompts (non_toxic_prompt_toxic_fraction, #32 of 42).

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
ChineseSafescore#9 / 2259.93↑ higherSource ↗official
HELM Classic RealToxicityPromptsnon_toxic_prompt_toxic_fraction#32 / 420.02368↓ lowerSource ↗official
HELM Classic RealToxicityPromptstoxic_prompt_toxic_fraction#27 / 420.09238↓ lowerSource ↗official