Llama 2-13B
Open Sourcevon Meta · Veroeffentlicht 2024-01-01
42.5
Durchschn. Score
N/A
Eingabepreis
N/A
Ausgabepreis
N/A
Kontextfenster
text
Typ
Tested on 14 benchmarks with 42.5% average. Top scores: TriviaQA (79.6%), LAMBADA (76.5%), HellaSwag (74.3%).
Benchmark-Ergebnisse
| Benchmark | Kategorie | Score | Bar |
|---|---|---|---|
| TriviaQA | knowledge | 79.6 | |
| LAMBADA | knowledge | 76.5 | |
| HellaSwag | knowledge | 74.3 | |
| PIQA | knowledge | 61.6 | |
| ARC AI2 | knowledge | 47.1 | |
| Winogrande | knowledge | 45.6 | |
| BBH | reasoning | 44.3 | |
| OpenBookQA | knowledge | 42.7 | |
| ScienceQA | knowledge | 41.0 | |
| MMLU | knowledge | 40.8 | |
| GSM8K | math | 36.9 | |
| MATH level 5 | math | 3.3 | |
| GPQA diamond | knowledge | 1.8 | |
| CSQA2 | knowledge | 0.1 |