Model profile

Phi 3 Medium 4K Instruct

Microsoftdeveloper

2024-05-07release date

Not rankedSafety rank

#497 / 645Freedom rank

Evidence summary

Safety. Phi 3 Medium 4K Instruct 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 Enkrypt AI Safety Leaderboard (toxicity_attack_non_success_rate, #81 of 246); its weakest is Enkrypt AI Safety Leaderboard (bias_attack_non_success_rate, #180 of 248).

Freedom. Phi 3 Medium 4K Instruct has an estimated Freedom rank of #497; its 90% source-sensitivity interval is #371–#556. Published Freedom evidence spans 1 eval lineages and 1 of 1 components. Its strongest relative result is Enkrypt AI Safety Leaderboard (harmful_attack_non_success_rate, #110 of 248); its weakest is Enkrypt AI Safety Leaderboard (toxicity_attack_non_success_rate, #164 of 246).

Compare this model

Second model

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

Official and reference links

Hugging Face ↗microsoft/Phi-3-medium-4k-instruct
OpenRouter ↗microsoft/phi-3-medium-4k-instruct
Release source ↗repository created at proxy

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-eval	Rank	Value	Source
Enkrypt AI Safety Leaderboardbias_attack_non_success_rate	#180 / 248	↑12.66	Source ↗official
Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate	#93 / 248	↑89.67	Source ↗official
Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate	#139 / 248	↑64.44	Source ↗official
Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate	#81 / 246	↑97.86	Source ↗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-eval	Rank	Value	Source
Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate	#152 / 248	↓89.67	Source ↗official
Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate	#110 / 248	↓64.44	Source ↗official
Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate	#164 / 246	↓97.86	Source ↗official