Multiple-choice science question answering on MMLU Chemistry Larger Dist. Shift 37, 36
44.79ACCMAP
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| MAPBackbone=Llama2-7B, PEFT Method=LoRA2024.06 | 44.79 | 17.55 | 1.55 | |
| BLoBBackbone=Llama2-7B, PEFT Method=LoRA, Number of samples (N)=102024.06 | 44.79 | 9.77 | 1.38 | |
| LAPBackbone=Llama2-7B, PEFT Method=LoRA2024.06 | 44.1 | 10.76 | 1.35 | |
| ENSBackbone=Llama2-7B, PEFT Method=LoRA2024.06 | 43.75 | 12.04 | 1.28 | |
| MCDBackbone=Llama2-7B, PEFT Method=LoRA2024.06 | 42.71 | 17.9 | 1.59 | |
| BBBBackbone=Llama2-7B, PEFT Method=LoRA2024.06 | 42.36 | 15.67 | 1.49 | |
| BLoBBackbone=Llama2-7B, PEFT Method=LoRA, Number of samples (N)=02024.06 | 41.67 | 15.05 | 1.44 | |
| MLEBackbone=Llama2-7B, PEFT Method=LoRA2024.06 | 40.62 | 21.72 | 1.56 | |
| BLoBBackbone=Llama2-7B, PEFT Method=LoRA, Number of samples (N)=52024.06 | 40.62 | 14.33 | 1.39 |