Embodied Household Planning on ALFWorld
87.5Success RateTRIAGE
Evaluation Results
| Method | Links | |
|---|---|---|
| TRIAGEModel=Qwen2.5-7B-Instruct, Judge=Qwen3-8B-thinking, Evidence=true2026.06 | 87.5 | |
| TRIAGEModel=Qwen2.5-7B-Instruct, Judge=Qwen3-8B-thinking, Evidence=false2026.06 | 83.1 | |
| GRPOModel=Qwen2.5-7B-Instruct2026.06 | 79.6 | |
| TRIAGEModel=Qwen2.5-7B-Instruct, Judge=Qwen3-8B no-think2026.06 | 76.8 | |
| TRIAGEModel=Qwen3-1.7B-Instruct, Judge=Qwen3-8B-thinking, Evidence=true2026.06 | 56.4 | |
| TRIAGEModel=Qwen3-1.7B-Instruct, Judge=Qwen3-8B-thinking, Evidence=false2026.06 | 51.8 | |
| GRPOModel=Qwen3-1.7B-Instruct2026.06 | 45.2 | |
| TRIAGEModel=Qwen3-1.7B-Instruct, Judge=Qwen3-8B no-think2026.06 | 40.7 |