Multiple-choice science question answering on OBQA In-Distribution 64
82.73AccuracyBLoB
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| BLoBBackbone=Llama2-7B, PEFT Method=LoRA, Number of samples (N)=02024.06 | 82.73 | 8.36 | 0.56 | |
| LAPBackbone=Llama2-7B, PEFT Method=LoRA2024.06 | 82.12 | 8.7 | 0.52 | |
| BBBBackbone=Llama2-7B, PEFT Method=LoRA2024.06 | 82.06 | 11.38 | 0.66 | |
| BLoBBackbone=Llama2-7B, PEFT Method=LoRA, Number of samples (N)=52024.06 | 81.79 | 3.4 | 0.53 | |
| MCDBackbone=Llama2-7B, PEFT Method=LoRA2024.06 | 81.72 | 14.45 | 1.06 | |
| MLEBackbone=Llama2-7B, PEFT Method=LoRA2024.06 | 81.52 | 12.55 | 0.73 | |
| BLoBBackbone=Llama2-7B, PEFT Method=LoRA, Number of samples (N)=102024.06 | 81.52 | 3.77 | 0.5 | |
| MAPBackbone=Llama2-7B, PEFT Method=LoRA2024.06 | 81.38 | 15.34 | 1.06 | |
| ENSBackbone=Llama2-7B, PEFT Method=LoRA2024.06 | 81.38 | 13.26 | 0.75 |