Action Prediction on AndroidControl High-level
87.4Click TMRClean
Evaluation results
| Method | Links | ||||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| CleanBackbone=OS-Atlas-Base-7B2025.05 | 87.4 | 55.5 | 95.7 | 51.3 | 94.1 | 74 | 85.6 | 78.7 | 60.1 | 70.5 | 86 | 61 | |
| ICLAttackBackbone=OS-Atlas-Base-7B2025.05 | 86.6 | 55.7 | 96.8 | 50.5 | 93.1 | 74.1 | 78.2 | 71 | 60.4 | 68.5 | 84.4 | 59.7 | |
| AgentGhostBackbone=OS-Atlas-Base-7B2025.05 | 86.5 | 57.2 | 93.3 | 47.6 | 89.8 | 73.5 | 70.8 | 62.9 | 55.3 | 69 | 82.9 | 59.3 | |
| CleanModel=Qwen2-VL-7B2025.05 | 86.1 | 43.1 | 95.4 | 51.9 | 93.4 | 72.7 | 85.1 | 76.5 | 63.3 | 72.5 | 85.2 | 53.4 | |
| ICLAttackModel=Qwen2-VL-2B2025.05 | 85.7 | 53.4 | 94.8 | 52 | 82.5 | 67.5 | 75.2 | 66.5 | 58.1 | 62.8 | 81.9 | 56.8 | |
| ICLAttackModel=Qwen2-VL-7B2025.05 | 85.5 | 42.3 | 97.3 | 51.8 | 90 | 69.7 | 80.9 | 71 | 53.8 | 65.3 | 83.2 | 50.7 | |
| CleanModel=Qwen2-VL-2B2025.05 | 85 | 54.2 | 95.6 | 53 | 91.1 | 73 | 82.1 | 74.4 | 64.4 | 68.3 | 83.9 | 59.7 | |
| AgentGhostModel=Qwen2-VL-7B2025.05 | 84.3 | 41.8 | 93.3 | 47 | 93.4 | 72.8 | 77.5 | 66.3 | 60.2 | 71.1 | 83 | 50.6 | |
| AgentGhostModel=Qwen2-VL-2B2025.05 | 83.2 | 50.4 | 90.3 | 46.3 | 93.2 | 72.3 | 69.8 | 58.8 | 57.2 | 68.3 | 81 | 54.6 | |
| AddSentBackbone=OS-Atlas-Base-7B2025.05 | 66 | 46.7 | 69.6 | 38.4 | 93.1 | 76.3 | 63.3 | 57.4 | 49.5 | 53 | 65.8 | 49.8 | |
| AddSentModel=Qwen2-VL-2B2025.05 | 65.6 | 40.3 | 68.5 | 39.2 | 89.7 | 72 | 60.6 | 55.5 | 50.8 | 46.1 | 64.5 | 45 | |
| AddSentModel=Qwen2-VL-7B2025.05 | 65.2 | 32.2 | 70.9 | 39.2 | 91 | 68.5 | 64.7 | 58.4 | 49.2 | 51.1 | 65.4 | 40.6 | |
| SynAttackModel=Qwen2-VL-7B2025.05 | 63.9 | 30.4 | 67.8 | 34.1 | 89.8 | 68.1 | 63.8 | 57.3 | 50.3 | 52.2 | 64.3 | 39 | |
| SynAttackBackbone=OS-Atlas-Base-7B2025.05 | 63.6 | 41.3 | 67.9 | 35.3 | 91 | 70.2 | 62.9 | 58.1 | 52.5 | 50.6 | 64 | 45.9 | |
| SynAttackModel=Qwen2-VL-2B2025.05 | 63.6 | 39.8 | 67.4 | 36.2 | 88.6 | 68.8 | 60.8 | 55.7 | 50.8 | 47.8 | 63.3 | 44.4 |