MMLU-by-task - Logical Fallacies — leaderboard
Metric: Accuracy (%). Source: huggingface.co. 1257 models tracked.
Top models
| # | Model | Score |
|---|---|---|
| 1 | Llama 2 70B Chat | 81.6 |
| 2 | Llama 2 70B Base | 80.37 |
| 3 | falcon-180B | 80.37 |
| 4 | StableBeluga2 | 79.75 |
| 5 | Mistral-7B-v0.1 | 79.14 |
| 6 | LLaMA-65B | 77.91 |
| 7 | Llama 2 70B Chat (HF) | 76.07 |
| 8 | Llama 2 70B Chat GPTQ | 73.62 |
| 9 | vicuna-33B-v1.3 | 71.78 |
| 10 | vicuna-13B-v1.5 | 69.33 |
| 11 | internlm-20B Chat | 68.71 |
| 12 | LLaMA-30B | 68.1 |
| 13 | MythoMax-L2-13B | 67.48 |
| 14 | OpenHermes-13B | 67.48 |
| 15 | trurl-2-13B-academic | 67.48 |
Interactive version: theaggregate.ai/benchmark?slug=mmlu-by-task-logical-fallacies · How the rankings work · Data refreshed daily, snapshot 2026-07-22.