Comprehensive Document Safety Analysis on T-SROIE dense text for robustness (test)
62.5M-F1DocShield
Evaluation Results
| Method | Links | |
|---|---|---|
| DocShield2026.04 | 62.5 | |
| Gemini-2.5-Pro2026.04 | 57 | |
| qwen3-vl-8B2026.04 | 55.6 | |
| InternVL-3.5-8B2026.04 | 53.7 | |
| gpt-4o-o3-04162026.04 | 52.8 | |
| FakeShield2026.04 | 26.8 | |
| Qwen2.5-VL-7B2026.04 | 26.2 | |
| DeepSeekVL-7B2026.04 | 24.8 |