Language Agent Task on WebShop (OOD)
49.6ScoreTask-Perturbed NLL Optimization
Evaluation Results
| Method | Links | |
|---|---|---|
| Task-Perturbed NLL OptimizationBackbone=Qwen3-8B2026.06 | 49.6 | |
| Task AugBackbone=Qwen3-8B2026.06 | 47.5 | |
| SFTBackbone=Qwen3-8B2026.06 | 47.1 | |
| COINBackbone=Qwen3-8B2026.06 | 45.2 |