Embodied Agent Task Execution on EB-ALFRED
36SRBenign SFT (GPT-4o)
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| Benign SFT (GPT-4o)Model=GPT-4o, D_benign=true, D_attack=false, D_contrast=false2025.10 | 36 | — | — | — | |
| BEAT (Qwen2-VL 7B)Model=Qwen2-VL 7B, D_benign=true, D_attack=true, D_contrast=true2025.10 | 34 | 59.2 | 0 | 0.721 | |
| Benign SFT (Qwen2-VL 7B)Model=Qwen2-VL 7B, D_benign=true, D_attack=false, D_contrast=false2025.10 | 32 | — | — | — | |
| BEAT (InternVL3 8B)Model=InternVL3 8B, D_benign=true, D_attack=true, D_contrast=true2025.10 | 26 | 80.8 | 0 | 0.872 | |
| Benign SFT (InternVL3 8B)Model=InternVL3 8B, D_benign=true, D_attack=false, D_contrast=false2025.10 | 24 | — | — | — | |
| BEAT w/o CTL (Qwen2-VL 7B)Model=Qwen2-VL 7B, D_benign=true, D_attack=true, D_contrast=false2025.10 | 17 | 40.2 | 22.5 | 0.667 | |
| BEAT w/o CTL (InternVL3 8B)Model=InternVL3 8B, D_benign=true, D_attack=true, D_contrast=false2025.10 | 16 | 69 | 81.3 | 0.655 | |
| BEAT w/o CTL (GPT-4o)Model=GPT-4o, D_benign=true, D_attack=true, D_contrast=false2025.10 | 14 | 55.8 | 19.5 | 0.663 | |
| Original (GPT-4o)Model=GPT-4o, D_benign=false, D_attack=false, D_contrast=false2025.10 | 11 | — | — | — | |
| Original (Qwen2-VL 7B)Model=Qwen2-VL 7B, D_benign=false, D_attack=false, D_contrast=false2025.10 | 0 | — | — | — | |
| Original (InternVL3 8B)Model=InternVL3 8B, D_benign=false, D_attack=false, D_contrast=false2025.10 | 0 | — | — | — |