Model profile
Llama 3.1 Nemotron Instruct 70B
Evidence summary
Safety. Llama 3.1 Nemotron Instruct 70B does not meet the evidence gate for a Safety rank. Published Safety evidence spans 2 eval lineages and 4 of 7 components. Its strongest relative result is Enkrypt AI Safety Leaderboard (bias_attack_non_success_rate, #16 of 270); its weakest is Enkrypt AI Safety Leaderboard (harmful_attack_non_success_rate, #142 of 270).
Freedom. Llama 3.1 Nemotron Instruct 70B has an estimated Freedom rank of #589; its 90% source-sensitivity interval is #417–#612. 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, #128 of 270); its weakest is Enkrypt AI Safety Leaderboard (toxicity_attack_non_success_rate, #230 of 268).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗llama-3-1-nemotron-instruct-70b
- Official model catalog ↗Family-level model document · nvidia · first party
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 |
|---|---|---|---|---|
| AA-Omnisciencehallucination_rate | #142 / 358 | ↓0.7126 | Source ↗official | |
| Enkrypt AI Safety Leaderboardbias_attack_non_success_rate | #16 / 270 | ↑54.26 | Source ↗official | |
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #67 / 270 | ↑90.83 | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #142 / 270 | ↑72.22 | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #38 / 268 | ↑99.05 | 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 |
|---|---|---|---|---|
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #200 / 270 | ↓90.83 | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #128 / 270 | ↓72.22 | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #230 / 268 | ↓99.05 | Source ↗official |