Model profile
Llama 3.2 3B
Evidence summary
Safety. Llama 3.2 3B has an estimated Safety rank of #320; its 90% source-sensitivity interval is #263–#330. Its behavior-only rank is #312; company governance moves the combined estimate to #320. Published Safety evidence spans 3 eval lineages and 4 of 7 components. Its strongest relative result is ThaiSafetyBench (safety_score, #15 of 18); its weakest is KIDBench Implicit Child Cue (implicit_child_cue_total_mean, #13 of 13).
Freedom. Llama 3.2 3B has an estimated Freedom rank of #420; its 90% source-sensitivity interval is #88–#615. Published Freedom evidence spans 2 eval lineages and 1 of 1 components. Its strongest relative result is ThaiSafetyBench (safety_score, #4 of 18); its weakest is VETO Misfired Alignment (misfired_alignment_rate_pct, #23 of 25).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Official model page ↗Family-level model document · meta · first party
- Release source ↗curated outward evidence
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 |
|---|---|---|---|---|
| KIDBench Implicit Child Cueimplicit_child_cue_total_mean | #13 / 13 | ↑3.03 | Source ↗official | |
| ThaiSafetyBenchsafety_score | #15 / 18 | ↑73.92 | Source ↗official | |
| VETO Misfired Alignmentmisfired_alignment_rate_pct | #23 / 25 | ↓11.8 | 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 |
|---|---|---|---|---|
| ThaiSafetyBenchsafety_score | #4 / 18 | ↓73.92 | Source ↗official | |
| VETO Misfired Alignmentmisfired_alignment_rate_pct | #23 / 25 | ↓11.8 | Source ↗official |