| Dataset Name | SOTA Method | Metric | Trend | ||
|---|---|---|---|---|---|
| MRBench Chinese 1.0 | Doubao-Seed-1.6-250615 | Memory Adherence (SI)8.85 | 12 | 2mo ago | |
| MRBench English 1.0 | Doubao-Seed-1.6-250615 | MA-SI Score9.13 | 12 | 2mo ago | |
| DynSess-Eval | Average Performance (Auto)4.35 | 8 | 6d ago | ||
| RoleChat (val) | RoleJudge | Overall MSE0.21 | 5 | 1mo ago | |
| RoleBench | DPO-Qwen3-8B | Win Rate37.1 | 4 | 15d ago |