← Models

Model profile

Koala 13B

berkeley ai researchdeveloper
2023-04-03release date
Not rankedSafety rank
#278 / 645Freedom rank

Evidence summary

Safety. Koala 13B 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 HarmBench (dr, #19 of 28); its weakest is HarmBench (dr, #19 of 28).

Freedom. Koala 13B has an estimated Freedom rank of #278; its 90% source-sensitivity interval is #134–#517. Published Freedom evidence spans 1 eval lineages and 1 of 1 components. Its strongest relative result is HarmBench (dr, #10 of 28); its weakest is HarmBench (dr, #10 of 28).

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
HarmBenchdr#19 / 2827.3Source ↗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
HarmBenchdr#10 / 2827.3Source ↗official