Uncertainty Quantification on SciQ (test)
74.5AUROCSENTSAR
Evaluation Results
| Method | Links | |
|---|---|---|
| SENTSARLLM Model=Vicuna-13b, Number of generations=5, Rouge-L threshold=0.52023.07 | 74.5 | |
| SARLLM Model=Vicuna-13b, Number of generations=5, Rouge-L threshold=0.52023.07 | 74.1 | |
| SENTSARLLM Model=LLaMA-2-13b-chat, Number of generations=5, Rouge-L threshold=0.52023.07 | 73.7 | |
| SARLLM Model=LLaMA-2-13b-chat, Number of generations=5, Rouge-L threshold=0.52023.07 | 72.5 | |
| SENTSARLLM Model=WizardLM-13b, Number of generations=5, Rouge-L threshold=0.52023.07 | 71.9 | |
| PELLM Model=LLaMA-2-13b-chat, Number of generations=5, Rouge-L threshold=0.52023.07 | 71.8 | |
| TOKENSARLLM Model=LLaMA-2-13b-chat, Number of generations=5, Rouge-L threshold=0.52023.07 | 71.8 | |
| SENTSARLLM Model=Vicuna-33b, Number of generations=5, Rouge-L threshold=0.52023.07 | 71.7 | |
| SARLLM Model=Vicuna-33b, Number of generations=5, Rouge-L threshold=0.52023.07 | 71 | |
| PELLM Model=Vicuna-13b, Number of generations=5, Rouge-L threshold=0.52023.07 | 70.8 | |
| SARLLM Model=WizardLM-13b, Number of generations=5, Rouge-L threshold=0.52023.07 | 70.7 | |
| TOKENSARLLM Model=Vicuna-13b, Number of generations=5, Rouge-L threshold=0.52023.07 | 70.6 | |
| SELLM Model=LLaMA-2-13b-chat, Number of generations=5, Rouge-L threshold=0.52023.07 | 69.2 | |
| LN-PELLM Model=LLaMA-2-13b-chat, Number of generations=5, Rouge-L threshold=0.52023.07 | 68.8 | |
| TOKENSARLLM Model=WizardLM-13b, Number of generations=5, Rouge-L threshold=0.52023.07 | 68.1 | |
| PELLM Model=WizardLM-13b, Number of generations=5, Rouge-L threshold=0.52023.07 | 67.7 | |
| SELLM Model=Vicuna-13b, Number of generations=5, Rouge-L threshold=0.52023.07 | 67.5 | |
| SELLM Model=Vicuna-33b, Number of generations=5, Rouge-L threshold=0.52023.07 | 67.4 | |
| LN-PELLM Model=Vicuna-13b, Number of generations=5, Rouge-L threshold=0.52023.07 | 66.8 | |
| LN-PELLM Model=Vicuna-33b, Number of generations=5, Rouge-L threshold=0.52023.07 | 66.8 | |
| PELLM Model=Vicuna-33b, Number of generations=5, Rouge-L threshold=0.52023.07 | 66.5 | |
| TOKENSARLLM Model=Vicuna-33b, Number of generations=5, Rouge-L threshold=0.52023.07 | 66.5 | |
| SELLM Model=WizardLM-13b, Number of generations=5, Rouge-L threshold=0.52023.07 | 64.9 | |
| LN-PELLM Model=WizardLM-13b, Number of generations=5, Rouge-L threshold=0.52023.07 | 63.8 | |
| LSLLM Model=Vicuna-13b, Number of generations=5, Rouge-L threshold=0.52023.07 | 58.9 | |
| LSLLM Model=Vicuna-33b, Number of generations=5, Rouge-L threshold=0.52023.07 | 58.4 | |
| LSLLM Model=LLaMA-2-13b-chat, Number of generations=5, Rouge-L threshold=0.52023.07 | 57.8 | |
| LSLLM Model=WizardLM-13b, Number of generations=5, Rouge-L threshold=0.52023.07 | 57.4 |