Question Answering on TriviaQA (AUROC, AUARC, ECE)
87AUROCSGPU
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| SGPULLM Architecture=Llama-3.1-8B2025.12 | 87 | 86.7 | 13.9 | |
| KLE-MaternLLM Architecture=Llama-3.1-8B2025.12 | 86 | 85.1 | 7.8 | |
| KLE-HeatLLM Architecture=Llama-3.1-8B2025.12 | 84.4 | 84.7 | 10.5 | |
| UMPIRELLM Architecture=Llama-3.1-8B2025.12 | 83.9 | 82.1 | 31 | |
| Cos EigenscoreLLM Architecture=Llama-3.1-8B2025.12 | 83.3 | 78.6 | 8 | |
| DSELLM Architecture=Llama-3.1-8B2025.12 | 82.8 | 83.8 | 28.1 | |
| SELLM Architecture=Llama-3.1-8B2025.12 | 82.1 | 83.6 | 25.5 | |
| RegularLLM Architecture=Llama-3.1-8B2025.12 | 75.2 | 81.9 | 6 | |
| Cov EigenscoreLLM Architecture=Llama-3.1-8B2025.12 | 72.4 | 74 | 24 |