Class-level Continual Learning on MMLU-PRO
50.72Average Accuracy (AA)MasFACT
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| MasFACTSetting=Standard, Replay-based=false2026.05 | 50.72 | 0.13 | |
| MasFACTEvaluation Style=Standard2026.05 | 50.72 | 0.13 | |
| MasFACTSetting=Few-shot, Replay-based=false2026.05 | 49.18 | -0.19 | |
| MasFACTEvaluation Style=Few-shot2026.05 | 49.18 | -0.19 | |
| ARGDesignerSetting=Standard, Replay-based=true2026.05 | 48.69 | 1.33 | |
| ARGDesigner†Evaluation Style=Standard, Replay=true2026.05 | 48.69 | 1.33 | |
| GDesignerSetting=Standard, Replay-based=true2026.05 | 47.87 | 1.22 | |
| GDesigner†Evaluation Style=Standard, Replay=true2026.05 | 47.87 | 1.22 | |
| MasRouterSetting=Standard, Replay-based=true2026.05 | 47.53 | 0.83 | |
| MasRouter†Evaluation Style=Standard, Replay=true2026.05 | 47.53 | 0.83 | |
| MasRouterSetting=Few-shot, Replay-based=true2026.05 | 47.4 | 1.43 | |
| MasRouter†Evaluation Style=Few-shot, Replay=true2026.05 | 47.4 | 1.43 | |
| ARGDesignerSetting=Few-shot, Replay-based=true2026.05 | 46.87 | 1.27 | |
| ARGDesigner†Evaluation Style=Few-shot, Replay=true2026.05 | 46.87 | 1.27 | |
| GDesignerSetting=Standard, Replay-based=false2026.05 | 46.72 | 3.61 | |
| GDesignerEvaluation Style=Standard2026.05 | 46.72 | 3.61 | |
| MasRouterSetting=Standard, Replay-based=false2026.05 | 46.17 | 2.81 | |
| MasRouterEvaluation Style=Standard2026.05 | 46.17 | 2.81 | |
| GDesignerSetting=Few-shot, Replay-based=true2026.05 | 46.16 | 1.49 | |
| GDesigner†Evaluation Style=Few-shot, Replay=true2026.05 | 46.16 | 1.49 | |
| MasRouterSetting=Few-shot, Replay-based=false2026.05 | 45.94 | 2.27 | |
| MasRouterEvaluation Style=Few-shot2026.05 | 45.94 | 2.27 | |
| AgentDropoutSetting=Standard, Replay-based=true2026.05 | 45.84 | 1.49 | |
| AgentDropout†Evaluation Style=Standard, Replay=true2026.05 | 45.84 | 1.49 | |
| ARGDesignerSetting=Standard, Replay-based=false2026.05 | 45.52 | 3.39 | |
| ARGDesignerEvaluation Style=Standard2026.05 | 45.52 | 3.39 | |
| GDesignerSetting=Few-shot, Replay-based=false2026.05 | 45.04 | 3.28 | |
| GDesignerEvaluation Style=Few-shot2026.05 | 45.04 | 3.28 | |
| ARGDesignerSetting=Few-shot, Replay-based=false2026.05 | 44.72 | 3.25 | |
| ARGDesignerEvaluation Style=Few-shot2026.05 | 44.72 | 3.25 | |
| AgentDropoutSetting=Few-shot, Replay-based=true2026.05 | 43.03 | 1.72 | |
| AgentDropout†Evaluation Style=Few-shot, Replay=true2026.05 | 43.03 | 1.72 | |
| AFLOWSetting=Standard, Replay-based=true2026.05 | 42.89 | 2.19 | |
| AFLOW†Evaluation Style=Standard, Replay=true2026.05 | 42.89 | 2.19 | |
| AFLOWSetting=Few-shot, Replay-based=true2026.05 | 42.52 | 1.21 | |
| AFLOW†Evaluation Style=Few-shot, Replay=true2026.05 | 42.52 | 1.21 | |
| AgentDropoutSetting=Standard, Replay-based=false2026.05 | 42.48 | 3.83 | |
| AgentDropoutEvaluation Style=Standard2026.05 | 42.48 | 3.83 | |
| AFLOWSetting=Standard, Replay-based=false2026.05 | 42.37 | 4.32 | |
| AFLOWEvaluation Style=Standard2026.05 | 42.37 | 4.32 | |
| AgentDropoutSetting=Few-shot, Replay-based=false2026.05 | 41.07 | 3.13 | |
| AgentDropoutEvaluation Style=Few-shot2026.05 | 41.07 | 3.13 | |
| SC(CoT)-ICLSetting=Few-shot, Replay-based=false2026.05 | 40.52 | 4.95 | |
| SC(CoT)-ICLEvaluation Style=Few-shot2026.05 | 40.52 | 4.95 | |
| AFLOWSetting=Few-shot, Replay-based=false2026.05 | 40.12 | 3.97 | |
| AFLOWEvaluation Style=Few-shot2026.05 | 40.12 | 3.97 | |
| CoT-ICLSetting=Few-shot, Replay-based=false2026.05 | 39.6 | 4.01 | |
| CoT-ICLEvaluation Style=Few-shot2026.05 | 39.6 | 4.01 | |
| CompleteSetting=Few-shot, Replay-based=false2026.05 | 38.64 | — | |
| CompleteEvaluation Style=Few-shot2026.05 | 38.64 | — | |
| TreeSetting=Few-shot, Replay-based=false2026.05 | 36.82 | — | |
| TreeEvaluation Style=Few-shot2026.05 | 36.82 | — | |
| ChainSetting=Few-shot, Replay-based=false2026.05 | 36.48 | — | |
| ChainEvaluation Style=Few-shot2026.05 | 36.48 | — | |
| VanillaSetting=Few-shot, Replay-based=false2026.05 | 26.17 | — | |
| VanillaEvaluation Style=Few-shot2026.05 | 26.17 | — |