Long context understanding on InfiniteBench En.MC
83.4AccuracyLlama 3 405B
Evaluation Results
| Method | Links | |
|---|---|---|
| Llama 3 405Bzero-shot=true2024.07 | 83.4 | |
| GPT-4ozero-shot=true2024.07 | 82.5 | |
| Llama 3 70Bzero-shot=true2024.07 | 78.2 | |
| GPT-4 (0125)zero-shot=true2024.07 | 72.1 | |
| Llama 3 8Bzero-shot=true2024.07 | 65.1 |