Embodied AI task execution on ALFWorld OOD v1
0.994SRAgentMark
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| AgentMarkTask=A1: Look At Obj In Light2026.01 | 0.994 | 11 | 1.38 | 30.3 | 0.07 | -0.0076 | |
| AgentMarkTask=A2: Pick And Place Simple2026.01 | 0.979 | 13.6 | 1.39 | 25.5 | 0.15 | -0.0022 | |
| AgentMarkTask=A4: Pick Cool Then Place2026.01 | 0.976 | 14.3 | 1.18 | 22.7 | -1.01 | -0.0011 | |
| AgentMarkTask=Avg.2026.01 | 0.975 | 14.1 | 1.34 | 28.4 | -0.18 | -0.003 | |
| AgentMarkTask=A6: Pick Two Obj And Place2026.01 | 0.971 | 24.1 | 1.41 | 38.2 | -0.18 | -0.0088 | |
| AgentMarkTask=A3: Pick Clean Then Place2026.01 | 0.968 | 15.2 | 1.29 | 24.5 | -0.18 | 0.0009 | |
| AgentMarkTask=A5: Pick Heat Then Place2026.01 | 0.913 | 17.1 | 1.44 | 33.4 | 0.02 | -0.0037 |