Web Navigation on WebShop Implicit Hidden Drift
100Success Rate (Source)Llama3.3-70B + GLOVE
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Llama3.3-70B + GLOVEBackbone=Llama3.3-70B, GLOVE Augmentation=true2026.01 | 100 | 83.4 | 6.8 | |
| GPT-4oBackbone=GPT-4o, GLOVE Augmentation=false2026.01 | 99.4 | 76.6 | — | |
| GPT-4o + GLOVEBackbone=GPT-4o, GLOVE Augmentation=true2026.01 | 98.4 | 91.6 | 15 | |
| Grok-3Backbone=Grok-3, GLOVE Augmentation=false2026.01 | 96.2 | 78.4 | — | |
| Llama3.3-70BBackbone=Llama3.3-70B, GLOVE Augmentation=false2026.01 | 95.3 | 76.6 | — | |
| DeepSeek-V3.2Backbone=DeepSeek-V3.2, GLOVE Augmentation=false2026.01 | 95.3 | 76.9 | — | |
| Grok-3 + GLOVEBackbone=Grok-3, GLOVE Augmentation=true2026.01 | 95 | 98.4 | 20 | |
| DeepSeek-V3.2 + GLOVEBackbone=DeepSeek-V3.2, GLOVE Augmentation=true2026.01 | 94.4 | 97.5 | 20.6 | |
| Qwen3-30BBackbone=Qwen3-30B, GLOVE Augmentation=false2026.01 | 75 | 100 | — | |
| Qwen3-30B + GLOVEBackbone=Qwen3-30B, GLOVE Augmentation=true2026.01 | 75 | 100 | — | |
| Qwen2.5-7B + GLOVEBackbone=Qwen2.5-7B, GLOVE Augmentation=true2026.01 | 46.9 | 75 | 5.3 | |
| Llama3.1-8B + GLOVEBackbone=Llama3.1-8B, GLOVE Augmentation=true2026.01 | 42.5 | 61.6 | 25.4 | |
| Qwen2.5-7BBackbone=Qwen2.5-7B, GLOVE Augmentation=false2026.01 | 35.6 | 69.7 | — | |
| Llama3.1-8BBackbone=Llama3.1-8B, GLOVE Augmentation=false2026.01 | 34.4 | 36.2 | — |