Model profile
Llama 3 8B Instruct
Metadeveloper
2024-04-18release date
#183 / 267overall rank
15eval lineages
Evidence summary
Llama 3 8B Instruct has an estimated overall rank of #183; its 90% source-sensitivity interval is #93–#213. Its behavior-only rank is #169; company governance moves the combined estimate to #183. Published evidence spans 15 evals and 7 of 7 behavior components. Its strongest relative result is Microsoft Phi Safety Panels (harmful_summarization, #1 of 10); its weakest is Contextual MoralChoice (human_agreement, #21 of 22).
Compare this model
Only models sharing at least one published sub-eval are listed.
Official and reference links
- Artificial Analysis ↗llama-3-instruct-8b
- OpenRouter ↗meta-llama/llama-3-8b-instruct
- Official model page ↗Family-level model document · meta · first party
- Release source ↗direct research preview identity date
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 |
|---|---|---|---|---|---|
| AA-Omnisciencehallucination_rate | #231 / 311 | 0.8797 | ↓ lower | Source ↗official | |
| AIRBench 2024 Safety Scenariossafety_scenarios | #35 / 80 | 0.709 | ↑ higher | Source ↗official | |
| Contextual MoralChoicehuman_agreement | #21 / 22 | 0.3 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardbias_attack_non_success_rate | #92 / 260 | 19.64 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardcbrn_attack_non_success_rate | #178 / 260 | 85.33 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardharmful_attack_non_success_rate | #66 / 260 | 87.22 | ↑ higher | Source ↗official | |
| Enkrypt AI Safety Leaderboardtoxicity_attack_non_success_rate | #168 / 258 | 94.77 | ↑ higher | Source ↗official | |
| HELM Safetyanthropic_red_team | #41 / 80 | 0.988 | ↑ higher | Source ↗official | |
| HELM Safetybbq | #70 / 80 | 0.765 | ↑ higher | Source ↗official | |
| HELM Safetyharmbench | #43 / 80 | 0.727 | ↑ higher | Source ↗official | |
| HELM Safetysimple_safety_tests | #29 / 80 | 0.993 | ↑ higher | Source ↗official | |
| HELM Safetyxstest | #43 / 80 | 0.956 | ↑ higher | Source ↗official | |
| JailBenchjailbreak_success_rate | #2 / 14 | 43.94 | ↓ lower | Source ↗official | |
| Large-scale Moral Machine experiment on LLMshuman_choice_distance | #8 / 39 | 0.8324 | ↓ lower | Source ↗official | |
| Microsoft Phi Safety Panelsharmful_continuation | #6 / 10 | 0.013 | ↓ lower | Source ↗official | |
| Microsoft Phi Safety Panelsharmful_summarization | #1 / 10 | 0.082 | ↓ lower | Source ↗official | |
| Microsoft Phi Safety Panelsjailbreak | #7 / 10 | 0.13 | ↓ lower | Source ↗official | |
| Microsoft Phi Safety Panelsthird_party_harm | #7 / 10 | 0.373 | ↓ lower | Source ↗official | |
| OR-Benchover_refusal_rate | #15 / 25 | 69.4 | ↓ lower | Source ↗official | |
| OR-Benchtoxic_acceptance_rate | #14 / 25 | 5 | ↓ lower | Source ↗official | |
| S-Evalbase_en_overall | #8 / 22 | 69.1 | ↑ higher | Source ↗official | |
| SORRY-Benchavg | #19 / 51 | 0.22 | ↓ lower | Source ↗official |