Commonsense Question Answering on MCEval CSQA 8K (test)
84.6AccuracyMagn-Probe
Evaluation Results
| Method | Links | |
|---|---|---|
| Magn-ProbeModel=Llama2-70B2024.12 | 84.6 | |
| ActModel=Qwen2.5-7B2024.12 | 81.6 | |
| Magn-ProbeModel=Qwen2.5-7B2024.12 | 81.4 | |
| LM-ProbModel=Qwen2.5-7B2024.12 | 76.8 | |
| LM-ProbModel=Llama2-70B2024.12 | 74.3 | |
| Magn-ProbeModel=Llama3.1-8B2024.12 | 74 | |
| ActModel=Llama3.1-8B2024.12 | 73.3 | |
| Magn-ProbeModel=Llama3.2-3B2024.12 | 69.1 | |
| ActModel=Llama3.2-3B2024.12 | 68.1 | |
| LM-ProbModel=Llama3.1-8B2024.12 | 67.4 | |
| Magn-ProbeModel=Llama2-7B2024.12 | 63.9 | |
| ActModel=Llama2-7B2024.12 | 61.3 | |
| LM-ProbModel=Llama2-7B2024.12 | 61 | |
| LM-ProbModel=Llama3.2-3B2024.12 | 55.1 |