Long-form generation on DiffTask
0.16PRRHUQ-SATRMD
Evaluation Results
| Method | Links | |
|---|---|---|
| HUQ-SATRMDModel=Llama 3.1-8B2026.04 | 0.16 | |
| SATMD-MSPModel=Llama 3.1-8B2026.04 | 0.15 | |
| SAPLMA (mid)Model=Llama 3.1-8B2026.04 | 0.1 | |
| SATRMD-MSPModel=Llama 3.1-8B2026.04 | 0.09 | |
| HUQ-SATMDModel=Llama 3.1-8B2026.04 | 0.04 | |
| HUQ-SATMDModel=Gemma-2-9B2026.04 | 0.02 | |
| SAPLMA (mid)Model=Gemma-2-9B2026.04 | 0.02 | |
| HBOModel=Llama 3.1-8B2026.04 | 0.01 | |
| MSPModel=Llama 3.1-8B2026.04 | -0.01 | |
| SATRMD-MSPModel=Gemma-2-9B2026.04 | -0.06 | |
| HBOModel=Gemma-2-9B2026.04 | -0.07 | |
| MSPModel=Gemma-2-9B2026.04 | -0.1 | |
| HUQ-SATRMDModel=Gemma-2-9B2026.04 | -0.11 | |
| SATMD-MSPModel=Gemma-2-9B2026.04 | -0.18 |