Model profile
Reflection Llama 3.1 70B
Evidence summary
Safety. Reflection Llama 3.1 70B 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, #62 of 268); its weakest is Enkrypt AI Safety Leaderboard (harmful_attack_non_success_rate, #144 of 270).
Freedom. Reflection Llama 3.1 70B has an estimated Freedom rank of #549; its 90% source-sensitivity interval is #402–#616. 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, #125 of 270); its weakest is Enkrypt AI Safety Leaderboard (toxicity_attack_non_success_rate, #207 of 268).
Compare this model
Only models sharing at least one published sub-eval are listed.
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 |
|---|---|---|---|---|
| Enkrypt AI Safety Leaderboardbias_attack_non_success_rate | #106 / 270 | ↑19.12 | Source ↗official | |
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #141 / 270 | ↑87 | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #144 / 270 | ↑71.67 | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #62 / 268 | ↑98.2 | 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 | #126 / 270 | ↓87 | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #125 / 270 | ↓71.67 | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #207 / 268 | ↓98.2 | Source ↗official |