Common-sense reasoning on Winogrande Small (In-Distribution)
78.82AccuracyENS
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| ENSBackbone=Llama3.1-8B, Fine-tuning method=LoRA2026.07 | 78.82 | 14.72 | 0.75 | |
| MCDBackbone=Llama3.1-8B, Fine-tuning method=LoRA, Monte Carlo samples (M)=102026.07 | 78.03 | 16.13 | 0.83 | |
| MLEBackbone=Llama3.1-8B, Fine-tuning method=LoRA2026.07 | 77.87 | 17.02 | 0.88 | |
| DALorRABackbone=Llama3.1-8B, Fine-tuning method=LoRA, Monte Carlo samples (M)=102026.07 | 77.43 | 7.81 | 0.58 | |
| BLoBBackbone=Llama3.1-8B, Fine-tuning method=LoRA, Monte Carlo samples (M)=102026.07 | 77.34 | 8.84 | 0.58 | |
| C-LoRABackbone=Llama3.1-8B, Fine-tuning method=LoRA, Monte Carlo samples (M)=102026.07 | 77.26 | 16.84 | 0.99 | |
| MAPBackbone=Llama3.1-8B, Fine-tuning method=LoRA2026.07 | 76.9 | 18.71 | 0.99 | |
| LABackbone=Llama3.1-8B, Fine-tuning method=LoRA2026.07 | 76.05 | 4.18 | 0.56 | |
| TFBBackbone=Llama3.1-8B, Fine-tuning method=LoRA, Monte Carlo samples (M)=102026.07 | 74.65 | 8.23 | 0.59 |