Target product selection on Webshop
98.3PRACPRAC
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| PRACModel=Kimi-VL†, Agent Setting=Action, Threat Model=ℓ∞2026.04 | 98.3 | — | — | — | — | — | |
| PRACModel=Kimi-VL†, Agent Setting=ReAct-F, Threat Model=ℓ∞2026.04 | 96.8 | — | — | — | — | — | |
| PRACModel=Kimi-VL†, Agent Setting=ReAct-S, Threat Model=ℓ∞2026.04 | 96.2 | — | — | — | — | — | |
| PRACModel=Qwen3-VL, Agent Setting=ReAct-S, Threat Model=ℓ∞2026.04 | 95.3 | — | — | — | — | — | |
| PRACModel=Qwen3-VL, Agent Setting=ReAct-F, Threat Model=ℓ∞2026.04 | 93.6 | — | — | — | — | — | |
| PRACModel=Qwen3-VL, Agent Setting=Action, Threat Model=ℓ∞2026.04 | 82.5 | — | — | — | — | — | |
| MeanModel=Aggregate2026.04 | 82.3 | 20.8 | 66.5 | 19.7 | 17.3 | 35.7 | |
| PRACModel=GLM4.6V‡, Agent Setting=ReAct-S, Threat Model=ℓ∞2026.04 | 80.1 | — | — | — | — | — | |
| PRACModel=GLM4.6V‡, Agent Setting=ReAct-F, Threat Model=ℓ∞2026.04 | 78.4 | — | — | — | — | — | |
| PRACModel=GLM4.6V‡, Agent Setting=Action, Threat Model=ℓ∞2026.04 | 77.7 | — | — | — | — | — | |
| PRACModel=EvoCUA‡, Agent Setting=ReAct-F, Threat Model=ℓ∞2026.04 | 68.1 | — | — | — | — | — | |
| PRACModel=EvoCUA‡, Agent Setting=ReAct-S, Threat Model=ℓ∞2026.04 | 64.6 | — | — | — | — | — | |
| PRACModel=EvoCUA‡, Agent Setting=Action, Threat Model=ℓ∞2026.04 | 60.4 | — | — | — | — | — | |
| AAModel=Qwen3-VL, Agent Setting=ReAct-F, Threat Model=ℓ∞, BB2026.04 | — | — | — | 17.4 | — | — | |
| AAModel=Qwen3-VL, Agent Setting=ReAct-S, Threat Model=ℓ∞, BB2026.04 | — | — | — | 16.7 | — | — | |
| AAModel=Qwen3-VL, Agent Setting=Action, Threat Model=ℓ∞, BB2026.04 | — | — | — | 19.7 | — | — | |
| CE-tModel=Qwen3-VL, Agent Setting=ReAct-F, Threat Model=ℓ∞2026.04 | — | — | 64.8 | — | — | — | |
| CE-tModel=Qwen3-VL, Agent Setting=ReAct-S, Threat Model=ℓ∞2026.04 | — | — | 70 | — | — | — | |
| CE-tModel=Qwen3-VL, Agent Setting=Action, Threat Model=ℓ∞2026.04 | — | — | 34.3 | — | — | — | |
| CleanModel=Qwen3-VL, Agent Setting=ReAct-F2026.04 | — | 19.5 | — | — | — | — | |
| CleanModel=Qwen3-VL, Agent Setting=ReAct-S2026.04 | — | 20.8 | — | — | — | — | |
| CleanModel=Qwen3-VL, Agent Setting=Action2026.04 | — | 20.2 | — | — | — | — | |
| TOModel=Qwen3-VL, Agent Setting=ReAct-F, Threat Model=BB2026.04 | — | — | — | — | — | 43.2 | |
| TOModel=Qwen3-VL, Agent Setting=ReAct-S, Threat Model=BB2026.04 | — | — | — | — | — | 55 | |
| TOModel=Qwen3-VL, Agent Setting=Action, Threat Model=BB2026.04 | — | — | — | — | — | 24.3 | |
| TRAPModel=Qwen3-VL, Agent Setting=ReAct-F, Threat Model=BB2026.04 | — | — | — | — | 12.6 | — | |
| TRAPModel=Qwen3-VL, Agent Setting=ReAct-S, Threat Model=BB2026.04 | — | — | — | — | 11.7 | — | |
| TRAPModel=Qwen3-VL, Agent Setting=Action, Threat Model=BB2026.04 | — | — | — | — | 19.6 | — |