rDPO (Best-of-K)
| Method | Links | ||||
|---|---|---|---|---|---|
2026.03 | 11.62 | 7.1 | -2.58 | 316 | |
2026.03 | 11.58 | 6.78 | -1.98 | 325 | |
2026.03 | 5.86 | 6.97 | -8.08 | 309 | |
2026.03 | 5.83 | 6.57 | -7.31 | 290 | |
2026.03 | 5.77 | 6.9 | -8.03 | 282 | |
2026.03 | 5.68 | 6.68 | -7.68 | 232 | |
2026.03 | 5.66 | 8.74 | -11.81 | 311 | |
2026.03 | 5.3 | 6.63 | -7.96 | 278 | |
2026.03 | 5.22 | 6.4 | -7.58 | 274 | |
2026.03 | 4.52 | 7.46 | -10.4 | 331 | |
2026.03 | 4.49 | 6.21 | -7.93 | 321 | |
2026.03 | 4.38 | 5.6 | -6.82 | 319 | |
2026.03 | 4.37 | 6.31 | -8.25 | 326 | |
2026.03 | 4.31 | 5.82 | -7.33 | 321 | |
2026.03 | 4.27 | 5.91 | -7.55 | 320 | |
2026.03 | 4.25 | 5.67 | -7.33 | 319 | |
2026.03 | 4.12 | 6.24 | -8.36 | 330 | |
2026.03 | 4.02 | 3.74 | -3.46 | 267 | |
2026.03 | 3.29 | 3.11 | -2.93 | 256 | |
2026.03 | 3.18 | 2.36 | -1.54 | 261 | |
2026.03 | 3.17 | 2.71 | -2.25 | 257 | |
2026.03 | 3.14 | 2.68 | -2.22 | 261 | |
2026.03 | 3.14 | 2.66 | -2.18 | 256 | |
2026.03 | 3.13 | 2.79 | -2.45 | 253 | |
2026.03 | 3.12 | 2.48 | -1.85 | 257 | |
2026.03 | 3.09 | 2.62 | -2.15 | 259 |