Mistral 7B V0.1
Código abiertopor Mistral AI · Publicado el 2023-09-20
41.6
puntuación promedio
N/A
Precio de entrada
N/A
Precio de salida
N/A
Ventana de contexto
text-generation
Tipo
Tested on 16 benchmarks with 41.6% average. Top scores: TriviaQA (75.2%), HellaSwag (74.7%), OpenBookQA (73.1%).
Puntuaciones de benchmark
| Benchmark | Categoría | Puntuación | Bar |
|---|---|---|---|
| TriviaQA | knowledge | 75.2 | |
| HellaSwag | knowledge | 74.7 | |
| OpenBookQA | knowledge | 73.1 | |
| ARC AI2 | knowledge | 71.5 | |
| PIQA | knowledge | 66.0 | |
| GSM8K | math | 54.4 | |
| Winogrande | knowledge | 50.6 | |
| MMLU | knowledge | 50.0 | |
| BBH | reasoning | 41.5 | |
| IFEval | language | 23.9 | |
| MMLU-PRO | knowledge | 22.4 | |
| BBH (HuggingFace) | general | 22.0 | |
| ANLI | knowledge | 20.6 | |
| MUSR | reasoning | 10.7 | |
| GPQA | knowledge | 5.6 | |
| MATH Level 5 | math | 3.0 |
