| Task Name | Dataset Name | SOTA Result | Trend | |
|---|---|---|---|---|
| Preference Alignment | Skywork | Win Rate (W)80 | 24 | |
| Language Modeling | Skywork | Finance Score3.25 | 8 | |
| Direct Preference Optimization | Skywork AlpacaEval 2.0 | LCWR20.56 | 7 | |
| Human Evaluation | Skywork (test) | Elo Rating1,610.1 | 5 |