Top-k Ranking on SummEval
88Top-k RecallGlobal-uncertainty
Evaluation Results
| Method | Links | |
|---|---|---|
| Global-uncertaintyBudget (B)=320, Judge=Llama-3.1-8B, Bias handling condition=bias-aware2026.07 | 88 | |
| Top-k-awareBudget (B)=200, Judge=Llama-3.1-8B, Bias handling condition=bias-aware2026.07 | 86 | |
| Top-k-awareBudget (B)=320, Judge=Llama-3.1-8B, Bias handling condition=bias-aware2026.07 | 85 | |
| Top-k-awareBudget (B)=120, Judge=Llama-3.1-8B, Bias handling condition=bias-aware2026.07 | 80 | |
| Round-robinBudget (B)=320, Judge=Llama-3.1-8B, Bias handling condition=bias-aware2026.07 | 76 | |
| Global-uncertaintyBudget (B)=200, Judge=Llama-3.1-8B, Bias handling condition=bias-aware2026.07 | 76 | |
| Round-robinBudget (B)=200, Judge=Llama-3.1-8B, Bias handling condition=bias-aware2026.07 | 72 | |
| Global-uncertaintyBudget (B)=120, Judge=Llama-3.1-8B, Bias handling condition=bias-aware2026.07 | 70 | |
| Top-k-awareBudget (B)=60, Judge=Llama-3.1-8B, Bias handling condition=bias-aware2026.07 | 68 | |
| Round-robinBudget (B)=120, Judge=Llama-3.1-8B, Bias handling condition=bias-aware2026.07 | 61 | |
| Global-uncertaintyBudget (B)=60, Judge=Llama-3.1-8B, Bias handling condition=bias-aware2026.07 | 54 | |
| Round-robinBudget (B)=60, Judge=Llama-3.1-8B, Bias handling condition=bias-aware2026.07 | 53 |