Hallucination Evaluation on HallBench avg
58.1Hallucination ScoreQwen2-VL
Evaluation Results
| Method | Links | |
|---|---|---|
| Qwen2-VLNumber of Parameters=72B2025.02 | 58.1 | |
| InternVL2.5Number of Parameters=78B2025.02 | 57.4 | |
| Claude-3.5 Sonnet-0620Model Version=06202025.02 | 55.5 | |
| Qwen2.5-VLNumber of Parameters=72B2025.02 | 55.2 | |
| GPT-4o 0513Model Version=05132025.02 | 55 | |
| Qwen2.5-VLNumber of Parameters=7B2025.02 | 52.9 | |
| Qwen2.5-VLNumber of Parameters=3B2025.02 | 46.3 |