Continual Learning on Knowledge QA Domain-level stream
51.59AA ScoreMasFACT
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| MasFACTReplay-based=false2026.05 | 51.59 | 0.06 | |
| MasFACTReplay-based=false2026.05 | 51.59 | 0.06 | |
| MasRouterReplay-based=true2026.05 | 49.16 | 3.01 | |
| MasRouterReplay-based=true2026.05 | 49.16 | 3.01 | |
| ARGDesignerReplay-based=true2026.05 | 48.7 | 3.16 | |
| ARGDesignerReplay-based=true2026.05 | 48.7 | 3.16 | |
| GDesignerReplay-based=true2026.05 | 47.73 | 4.2 | |
| GDesignerReplay-based=true2026.05 | 47.73 | 4.2 | |
| ARGDesignerReplay-based=false2026.05 | 47.11 | 5.36 | |
| ARGDesignerReplay-based=false2026.05 | 47.11 | 5.36 | |
| MasRouterReplay-based=false2026.05 | 46.84 | 4.64 | |
| MasRouterReplay-based=false2026.05 | 46.84 | 4.64 | |
| GDesignerReplay-based=false2026.05 | 45.62 | 6.33 | |
| GDesignerReplay-based=false2026.05 | 45.62 | 6.33 | |
| AgentDropoutReplay-based=true2026.05 | 36.05 | 4.07 | |
| AgentDropoutReplay-based=true2026.05 | 36.05 | 4.07 | |
| CoT-ICLReplay-based=false2026.05 | 34.37 | 5.08 | |
| CoT-ICLReplay-based=false2026.05 | 34.37 | 5.08 | |
| AgentDropoutReplay-based=false2026.05 | 33.15 | 5.83 | |
| AgentDropoutReplay-based=false2026.05 | 33.15 | 5.83 | |
| SC(CoT)-ICLReplay-based=false2026.05 | 32.54 | 7.46 | |
| SC(CoT)-ICLReplay-based=false2026.05 | 32.54 | 7.46 | |
| CompleteReplay-based=false2026.05 | 32.24 | — | |
| Complete GraphReplay-based=false2026.05 | 32.24 | — | |
| TreeReplay-based=false2026.05 | 30.4 | — | |
| TreeReplay-based=false2026.05 | 30.4 | — | |
| AFLOWReplay-based=true2026.05 | 30.21 | 3.89 | |
| AFLOWReplay-based=true2026.05 | 30.21 | 3.89 | |
| ChainReplay-based=false2026.05 | 29.11 | — | |
| ChainReplay-based=false2026.05 | 29.11 | — | |
| AFLOWReplay-based=false2026.05 | 28.45 | 7.52 | |
| AFLOWReplay-based=false2026.05 | 28.45 | 7.52 | |
| VanillaReplay-based=false2026.05 | 25.57 | — | |
| VanillaReplay-based=false2026.05 | 25.57 | — |