Interactive Decision Making on ALFWorld (Efficiency Metrics)
91.73Avg@4 (Least-Time)TurnOPD
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| TurnOPDStudent=Qwen3-4B, Teacher=Qwen3-8B-GRPO2026.07 | 91.73 | 92.21 | 2.16 | 1.33 | |
| Vanilla OPDStudent=Qwen3-4B, Teacher=Qwen3-8B-GRPO2026.07 | 90.79 | 91.81 | 2.86 | 1 | |
| TeacherStudent=Qwen3-1.7B, Teacher=Qwen3-8B-GRPO2026.07 | 90.75 | 90.75 | — | — | |
| TeacherStudent=Qwen3-4B, Teacher=Qwen3-8B-GRPO2026.07 | 90.75 | 90.75 | — | — | |
| TCOD-F2BStudent=Qwen3-4B, Teacher=Qwen3-8B-GRPO2026.07 | 86.5 | 86.5 | 1.89 | 1.51 | |
| TurnOPDStudent=Qwen3-1.7B, Teacher=Qwen3-8B-GRPO2026.07 | 85.6 | 86.29 | 1.93 | 2.29 | |
| TCOD-F2BStudent=Qwen3-1.7B, Teacher=Qwen3-8B-GRPO2026.07 | 80.06 | 80.06 | 1.87 | 2.37 | |
| Vanilla OPDStudent=Qwen3-1.7B, Teacher=Qwen3-8B-GRPO2026.07 | 73.52 | 83 | 4.42 | 1 | |
| Zero-ShotStudent=Qwen3-4B, Teacher=Qwen3-8B-GRPO2026.07 | 8.17 | 8.17 | — | — | |
| Zero-ShotStudent=Qwen3-1.7B, Teacher=Qwen3-8B-GRPO2026.07 | 0 | 0 | — | — |