Assistant Response Generation on WildBench v2
68.4Win RateSFTUSER
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| SFTUSEROpponent=Initial2026.05 | 68.4 | — | |
| SFTUSEROpponent=RPUSER12026.05 | 63.9 | — | |
| RPUSER3Opponent=Initial2026.05 | 61.3 | — | |
| SFTUSEROpponent=RPUSER22026.05 | 60 | — | |
| RPUSER2Opponent=Initial2026.05 | 59.9 | — | |
| SFTUSEROpponent=RPUSER32026.05 | 58.5 | — | |
| RPUSER3Opponent=RPUSER12026.05 | 56.6 | — | |
| RPUSER1Opponent=Initial2026.05 | 55.4 | — | |
| RPUSER2Opponent=RPUSER12026.05 | 55.2 | — | |
| RPUSER3Opponent=RPUSER22026.05 | 51 | — | |
| RPUSER2Opponent=RPUSER32026.05 | 49 | — | |
| RPUSER1Opponent=RPUSER22026.05 | 44.8 | — | |
| Initial assistantOpponent=RPUSER12026.05 | 44.6 | — | |
| RPUSER1Opponent=RPUSER32026.05 | 43.4 | — | |
| RPUSER3Opponent=SFTUSER2026.05 | 41.5 | — | |
| Initial assistantOpponent=RPUSER22026.05 | 40.1 | — | |
| RPUSER2Opponent=SFTUSER2026.05 | 40 | — | |
| Initial assistantOpponent=RPUSER32026.05 | 38.7 | — | |
| RPUSER1Opponent=SFTUSER2026.05 | 36.1 | — | |
| Initial assistantOpponent=SFTUSER2026.05 | 31.6 | — | |
| Initial assistant2026.05 | — | 51.1 | |
| RPUSER12026.05 | — | 53.6 | |
| RPUSER22026.05 | — | 56.6 | |
| RPUSER32026.05 | — | 57.7 | |
| SFTUSER2026.05 | — | 61.5 |