Preference Alignment on UF-P-4
62.46Accuracy (%)SPL
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SPLModel=Llama-3.1-8B, beta=3 x 10^-52026.03 | 62.46 | 85.9 | |
| SPLModel=Llama-3.1-8B, beta=3 x 10^-62026.03 | 62.21 | 96.19 | |
| SPLModel=Llama-3.1-8B2026.03 | 62.21 | — | |
| SPLModel=Llama-3.1-8B, beta=3 x 10^-72026.03 | 61.92 | 93.75 | |
| SPLModel=Llama-3.2-3B, beta=3 x 10^-52026.03 | 61.62 | 77.15 | |
| SPLModel=Llama-3.2-3B, beta=3 x 10^-62026.03 | 61.56 | 82.32 | |
| SPLModel=Llama-3.2-3B2026.03 | 61.56 | — | |
| SPLModel=Llama-3.2-3B, beta=3 x 10^-72026.03 | 61.52 | 76.89 | |
| DPLModel=Llama-3.1-8B2026.03 | 57.66 | — | |
| BTLModel=Llama-3.1-8B2026.03 | 57.4 | — | |
| VPLModel=Llama-3.1-8B, beta=3 x 10^-72026.03 | 57.25 | 0 | |
| VPLModel=Llama-3.1-8B, beta=3 x 10^-52026.03 | 57.18 | 0 | |
| VPLModel=Llama-3.1-8B, beta=3 x 10^-62026.03 | 57.14 | 0 | |
| VPLModel=Llama-3.1-8B2026.03 | 57.14 | — | |
| BTLModel=Llama-3.2-3B2026.03 | 57.07 | — | |
| DPLModel=Llama-3.2-3B2026.03 | 57.04 | — | |
| VPLModel=Llama-3.2-3B, beta=3 x 10^-62026.03 | 57.03 | 0 | |
| VPLModel=Llama-3.2-3B2026.03 | 57.03 | — | |
| VPLModel=Llama-3.2-3B, beta=3 x 10^-52026.03 | 57 | 0 | |
| VPLModel=Llama-3.2-3B, beta=3 x 10^-72026.03 | 56.91 | 0 |