Meta Llama 3 8B Instruct vs Meta Llama 3 8B
Lado a lado. Cada métrica. Cada benchmark.
| Tipo | Meta Llama 3 8B Instruct | Meta Llama 3 8B |
|---|---|---|
| Provider |  Meta | Meta |
| puntuación promedio | 45.2 | 44.2 |
| Precio de entrada | - | - |
| Precio de salida | - | - |
| Ventana de contexto | - | - |
| Publicado el | 2024-04-17 | 2024-04-17 |
| Código abierto | Open Source | Open Source |
Puntuaciones de benchmark
11 benchmarks · Meta Llama 3 8B Instruct: 6, Meta Llama 3 8B: 5
| Benchmark | Categoría | Meta Llama 3 8B Instruct | Meta Llama 3 8B |
|---|---|---|---|
| BBH (HuggingFace) | general | 28.2 | 48.7 |
| GPQA | knowledge | 1.2 | 19.7 |
| IFEval | language | 74.1 | 16.0 |
| JCommonsenseQA | language | 87.7 | 82.9 |
| JMMLU | language | 46.7 | 44.7 |
| JNLI | language | 61.1 | 60.9 |
| JSQuAD | language | 89.5 | 88.9 |
| LLM-JP — Overall | language | 49.6 | 48.9 |
| MATH Level 5 | math | 8.7 | 18.6 |
| MMLU-PRO | knowledge | 29.6 | 41.2 |
| MUSR | reasoning | 1.6 | 16.0 |