Knowledge and Commonsense Reasoning on MMLU
89.1AccuracyGPT-4o
Evaluation Results
| Method | Links | |
|---|---|---|
| GPT-4oshots=52024.10 | 89.1 | |
| GPT-4Vshots=52024.10 | 86.4 | |
| Gemini-1.5 Proshots=52024.10 | 85.9 | |
| Gemini-1.5 Flashshots=52024.10 | 78.9 | |
| ARIAshots=52024.10 | 73.3 | |
| Llama3.2-11Bshots=52024.10 | 69.4 | |
| Pixtral-12Bshots=52024.10 | 69.2 | |
| Engram-40BShots=5-shot2026.01 | 60.6 | |
| Engram-27BShots=5-shot2026.01 | 60.4 | |
| IAP-mvBackbone=LLaMA-2-13B-Chat2024.09 | 58.25 | |
| Self-discBackbone=LLaMA-2-13B-Chat2024.09 | 58.07 | |
| IAP-ssBackbone=LLaMA-2-13B-Chat2024.09 | 57.72 | |
| MoE-27BShots=5-shot2026.01 | 57.4 | |
| Prompt #8Backbone=LLaMA-2-13B-Chat, Strategy=Zero-shot CoT2024.09 | 57.37 | |
| Prompt #7Backbone=LLaMA-2-13B-Chat, Strategy=Zero-shot CoT2024.09 | 55.61 | |
| Dense-4BShots=5-shot2026.01 | 48.6 | |
| Prompt #4Backbone=LLaMA-2-13B-Chat, Strategy=Zero-shot CoT2024.09 | 47.89 | |
| Prompt #2Backbone=LLaMA-2-13B-Chat, Strategy=Zero-shot CoT2024.09 | 42.81 | |
| Prompt #3Backbone=LLaMA-2-13B-Chat, Strategy=Zero-shot CoT2024.09 | 41.23 | |
| Prompt #5Backbone=LLaMA-2-13B-Chat, Strategy=Zero-shot CoT2024.09 | 39.3 | |
| Prompt #1Backbone=LLaMA-2-13B-Chat, Strategy=Zero-shot CoT2024.09 | 37.54 | |
| Prompt #9Backbone=LLaMA-2-13B-Chat, Strategy=Zero-shot CoT2024.09 | 28.77 | |
| Prompt #6Backbone=LLaMA-2-13B-Chat, Strategy=Zero-shot CoT2024.09 | 26.14 |