Qwen2 7B Instruct
Código abiertopor Alibaba · Publicado el 2024-06-04
50.5
puntuación promedio
N/A
Precio de entrada
N/A
Precio de salida
N/A
Ventana de contexto
text-generation
Tipo
Tested on 25 benchmarks with 50.5% average. Top scores: JSQuAD (89.6%), JCommonsenseQA (89.1%), JNLI (81.3%).
Puntuaciones de benchmark
| Benchmark | Categoría | Puntuación | Bar |
|---|---|---|---|
| JSQuAD | language | 89.6 | |
| JCommonsenseQA | language | 89.1 | |
| JNLI | language | 81.3 | |
| MMMLU — Chinese | language | 61.8 | |
| MMMLU — French | language | 60.8 | |
| MMMLU — Spanish | language | 60.2 | |
| MMMLU — Portuguese | language | 60.1 | |
| MMMLU — Italian | language | 59.0 | |
| MMMLU — German | language | 57.1 | |
| IFEval | language | 56.8 | |
| MMMLU — Japanese | language | 56.6 | |
| JMMLU | language | 56.5 | |
| MMMLU — Indonesian | language | 54.1 | |
| MMMLU — Korean | language | 54.0 | |
| LLM-JP — Overall | language | 51.7 | |
| MMMLU — Arabic | language | 50.7 | |
| MMMLU — Hindi | language | 45.1 | |
| MMMLU — Bengali | language | 43.4 | |
| BBH (HuggingFace) | general | 37.8 | |
| MMMLU — Swahili | language | 34.3 | |
| MMLU-PRO | knowledge | 31.6 | |
| MMMLU — Yoruba | language | 30.2 | |
| MATH Level 5 | math | 27.6 | |
| MUSR | reasoning | 7.4 | |
| GPQA | knowledge | 6.4 |