Model profile
Phi 3 Mini 4K Instruct
Evidence summary
Safety. Phi 3 Mini 4K Instruct has an estimated Safety rank of #156; its 90% source-sensitivity interval is #37–#310. Its behavior-only rank is #155; company governance moves the combined estimate to #156. Published Safety evidence spans 4 eval lineages and 3 of 7 components. Its strongest relative result is Open LLM Safety Index (jailbreakbench_safety_rate, #1 of 21); its weakest is Enkrypt AI Safety Leaderboard (bias_attack_non_success_rate, #180 of 270).
Freedom. Phi 3 Mini 4K Instruct has an estimated Freedom rank of #495; its 90% source-sensitivity interval is #181–#608. Published Freedom evidence spans 4 eval lineages and 1 of 1 components. Its strongest relative result is PandaBench JBB direct-request panel (safety_rate, #14 of 46); its weakest is Open LLM Safety Index (jailbreakbench_safety_rate, #20 of 21).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Hugging Face ↗microsoft/Phi-3-mini-4k-instruct
- Official model page ↗Exact model document · Reviewed official Hugging Face owner · official repository
- Release source ↗direct research preview identity date
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 | Distribution | Source |
|---|---|---|---|---|
| DSPSafeBenchscore | #2 / 12 | ↑78.62 | Source ↗official | |
| Enkrypt AI Safety Leaderboardbias_attack_non_success_rate | #180 / 270 | ↑13.18 | Source ↗official | |
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #99 / 270 | ↑89.5 | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #127 / 270 | ↑75.56 | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #153 / 268 | ↑95.77 | Source ↗official | |
| Open LLM Safety Indexjailbreakbench_safety_rate | #1 / 21 | ↑0.8 | Source ↗official | |
| Open LLM Safety Indexstrongreject_safety_rate | #7 / 21 | ↑0.5333 | Source ↗official | |
| PandaBench JBB direct-request panelsafety_rate | #28 / 46 | ↑0.97 | 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 | Distribution | Source |
|---|---|---|---|---|
| DSPSafeBenchscore | #11 / 12 | ↓78.62 | Source ↗official | |
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #170 / 270 | ↓89.5 | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #144 / 270 | ↓75.56 | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #116 / 268 | ↓95.77 | Source ↗official | |
| Open LLM Safety Indexjailbreakbench_safety_rate | #20 / 21 | ↓0.8 | Source ↗official | |
| Open LLM Safety Indexstrongreject_safety_rate | #15 / 21 | ↓0.5333 | Source ↗official | |
| PandaBench JBB direct-request panelsafety_rate | #14 / 46 | ↓0.97 | Source ↗official |
Values evaluations
Descriptive values and political-framing results are separate from safety/ethics ranks. Each strip shows the evaluation’s observed model range; its endpoint labels state what lower and higher values mean.