Lifelong Agent Interaction on StuLife
12.35Success Rate (SR)PROACTAGENT
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| PROACTAGENTRegime=Ours2026.04 | 12.35 | 19.26 | |
| GRPO+ReflexionRegime=Online baselines2026.04 | 9.37 | 13.51 | |
| Mem0Regime=Online baselines2026.04 | 7.34 | 10.83 | |
| ReflexionRegime=Online baselines2026.04 | 6.82 | 11.31 | |
| GRPO (online)Regime=Online baselines2026.04 | 6.71 | 10.28 | |
| MemoryBankRegime=Online baselines2026.04 | 6.05 | 7.69 | |
| SFTRegime=Offline baselines2026.04 | 5.11 | 7.27 | |
| AWMRegime=Online baselines2026.04 | 3.51 | 7.6 | |
| BaselineRegime=Offline baselines2026.04 | 2.56 | 6.33 |