Discriminative Hallucination Detection on DASH
79AccuracyOmniLMM
Evaluation Results
| Method | Links | |
|---|---|---|
| OmniLMMSize=12B2026.03 | 79 | |
| Qwen2.5-VL + FINER-TuningSize=7B2026.03 | 76.6 | |
| OmniLMM + RLAIF-VSize=12B2026.03 | 76.3 | |
| Qwen2.5-VLSize=7B2026.03 | 74.6 | |
| InternVL-3.5 + FINER-TuningSize=8B2026.03 | 74.5 | |
| InternVL-3.5Size=8B2026.03 | 68.3 | |
| InternVL-3.5 + FINER-TuningSize=14B2026.03 | 61.3 | |
| LLaVA-1.6Size=7B2026.03 | 58 | |
| LLaVA-1.6 + FINER-TuningSize=7B2026.03 | 57.4 | |
| InternVL-3.5Size=14B2026.03 | 55.8 |