Conflict Resolution on MemoryAgentBench conflict-bearing subset official
100Final AccuracyOracle
Evaluation Results
| Method | Links | |||||
|---|---|---|---|---|---|---|
| OracleTasks=2822026.07 | 100 | 100 | 0 | — | — | |
| StateFuseTasks=2822026.07 | 97.5 | 100 | 2.5 | — | — | |
| MV registerTasks=2822026.07 | 97.5 | 100 | 2.5 | — | — | |
| Raw logTasks=2822026.07 | 97.5 | 0 | 2.5 | — | — | |
| CollapsedTasks=2822026.07 | 97.5 | 0 | 2.5 | — | — | |
| StateFuseMode=cons., Tasks=2822026.07 | 64.9 | 100 | 2.1 | — | — | |
| StateFuse coreTasks=2822026.07 | 64.9 | 100 | 2.1 | — | — | |
| MV registerMode=cons., Tasks=2822026.07 | 64.9 | 100 | 2.1 | — | — | |
| Provenance surfaceMode=cons., Tasks=2822026.07 | 64.9 | 100 | 2.1 | — | — | |
| Raw logMode=cons., Tasks=2822026.07 | 64.9 | 0 | 2.1 | — | — |