← Models

Model profile

InternLM2 20B SFT

InternLMdeveloper
2024-01-11release date
Not rankedSafety rank
#172 / 662Freedom rank

Evidence summary

Safety. InternLM2 20B SFT does not meet the evidence gate for a Safety rank. Published Safety evidence spans 1 eval lineages and 3 of 7 components. Its strongest relative result is SafeDialBench (legality, #7 of 18); its weakest is SafeDialBench (fairness, #18 of 18).

Freedom. InternLM2 20B SFT has an estimated Freedom rank of #172; its 90% source-sensitivity interval is #46–#495. Published Freedom evidence spans 1 eval lineages and 1 of 1 components. Its strongest relative result is SafeDialBench (fairness, #1 of 18); its weakest is SafeDialBench (legality, #12 of 18).

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
SafeDialBenchaggression#17 / 18↑6.957Source ↗official
SafeDialBenchethics#11 / 18↑7.547Source ↗official
SafeDialBenchfairness#18 / 18↑7.1Source ↗official
SafeDialBenchlegality#7 / 18↑7.977Source ↗official
SafeDialBenchmorality#11 / 18↑7.213Source ↗official
SafeDialBenchprivacy#7 / 18↑7.597Source ↗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
SafeDialBenchaggression#2 / 18↓6.957Source ↗official
SafeDialBenchethics#8 / 18↓7.547Source ↗official
SafeDialBenchfairness#1 / 18↓7.1Source ↗official
SafeDialBenchlegality#12 / 18↓7.977Source ↗official
SafeDialBenchmorality#8 / 18↓7.213Source ↗official
SafeDialBenchprivacy#12 / 18↓7.597Source ↗official