Spatial-relation hallucination detection on R-Bench Image
81.13AccuracyRVE
Evaluation Results
| Method | Links | |
|---|---|---|
| RVEBackbone=LLaVA-NeXT-7B2026.05 | 81.13 | |
| RVEBackbone=ShareGPT4V-7B2026.05 | 80.31 | |
| ShareGPT4V-7BModel Version=Vanilla2026.05 | 79.67 | |
| LLaVA-NeXT-7BModel Version=Vanilla2026.05 | 79.35 | |
| RVEBackbone=InstructBLIP-7B2026.05 | 79.21 | |
| RVEBackbone=LLaVA-1.5-7B2026.05 | 79.08 | |
| LLaVA-1.5-7BModel Version=Vanilla2026.05 | 78.21 | |
| InstructBLIP-7BModel Version=Vanilla2026.05 | 77.62 |