Generalist Alignment on AlpacaEval 2.0
0.66Average RatingDEPO
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| DEPOIteration=32026.05 | 0.66 | 92.8 | |
| XPOIteration=32026.05 | 0.64 | 91.6 | |
| POPOIteration=32026.05 | 0.6 | 92.4 | |
| DPOIteration=32026.05 | -0.02 | 91.2 | |
| DEPOIteration=22026.05 | -2.18 | 89.5 | |
| POPOIteration=22026.05 | -2.45 | 89.1 | |
| XPOIteration=22026.05 | -2.6 | 88 | |
| DPOIteration=22026.05 | -3.34 | 87.1 | |
| DEPOIteration=12026.05 | -5.61 | 80.2 | |
| POPOIteration=12026.05 | -5.68 | 80 | |
| XPOIteration=12026.05 | -5.79 | 78.3 | |
| DPOIteration=12026.05 | -6.02 | 78.1 | |
| Llama-3-8B-SFT2026.05 | -7.92 | — |