Model profile
Phi 3 Medium
1eval lineages
Evidence summary
Published evidence spans 1 evals and 2 of 7 behavior components. Its strongest relative result is Microsoft Phi Safety Panels (jailbreak, #3 of 10); its weakest is Microsoft Phi Safety Panels (harmful_summarization, #6 of 10).
Compare this model
Only models sharing at least one published sub-eval are listed.
Published eval results
Rank is within that sub-eval. Black marks the observed result; the grey dot marks the value implied by the global rank. Values stay on each source’s native scale.
| Eval / sub-eval | Rank | Value | Distribution | Better | Source |
|---|---|---|---|---|---|
| Microsoft Phi Safety Panelsharmful_continuation | #5 / 10 | 0.01 | ↓ lower | Source ↗official | |
| Microsoft Phi Safety Panelsharmful_summarization | #6 / 10 | 0.112 | ↓ lower | Source ↗official | |
| Microsoft Phi Safety Panelsjailbreak | #3 / 10 | 0.111 | ↓ lower | Source ↗official | |
| Microsoft Phi Safety Panelsthird_party_harm | #5 / 10 | 0.251 | ↓ lower | Source ↗official |