← Models

Model profile

GPT Neox 20B

eleutheraideveloper
2022-04-07release date
Not rankedSafety rank
#350 / 645Freedom rank

Evidence summary

Safety. GPT Neox 20B does not meet the evidence gate for a Safety rank. Published Safety evidence spans 1 eval lineages and 1 of 7 components. Its strongest relative result is HELM Classic RealToxicityPrompts (toxic_prompt_toxic_fraction, #15 of 40); its weakest is HELM Classic RealToxicityPrompts (non_toxic_prompt_toxic_fraction, #26 of 40).

Freedom. GPT Neox 20B has an estimated Freedom rank of #350; its 90% source-sensitivity interval is #199–#475. Published Freedom evidence spans 1 eval lineages and 1 of 1 components. Its strongest relative result is HELM Classic RealToxicityPrompts (non_toxic_prompt_toxic_fraction, #15 of 40); its weakest is HELM Classic RealToxicityPrompts (toxic_prompt_toxic_fraction, #26 of 40).

Compare this model

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

Official and reference links

Safety evals

Rank and direction are specific to the Safety portfolio. Black marks the observed result; the grey dot marks the value implied by the Safety rank. Values stay on each source’s native scale.

Eval / sub-evalRankValueDistributionSource
HELM Classic RealToxicityPromptsnon_toxic_prompt_toxic_fraction#26 / 400.02188Source ↗official
HELM Classic RealToxicityPromptstoxic_prompt_toxic_fraction#15 / 400.07803Source ↗official

Freedom evals

Rank and direction are specific to the Freedom portfolio. Black marks the observed result; the grey dot marks the value implied by the Freedom rank. Values stay on each source’s native scale.

Eval / sub-evalRankValueDistributionSource
HELM Classic RealToxicityPromptsnon_toxic_prompt_toxic_fraction#15 / 400.02188Source ↗official
HELM Classic RealToxicityPromptstoxic_prompt_toxic_fraction#26 / 400.07803Source ↗official