Fact Verification on M2RAG
70AccuracyOurs-top3
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Ours-top3Model=Qwen2.5-VL2026.04 | 70 | 73.75 | |
| Ours-top3Model=MM-RAIT (Qwen2.5-VL)2026.04 | 68 | 69.8 | |
| Ours-top1Model=MM-RAIT (Qwen2.5-VL)2026.04 | 66 | 67.43 | |
| jinarerank-top3Model=MM-RAIT (Qwen2.5-VL)2026.04 | 63 | 64.31 | |
| jinarerank-top3Model=Qwen2.5-VL2026.04 | 63 | 65.96 | |
| jinarerank-top1Model=MM-RAIT (Qwen2.5-VL)2026.04 | 62 | 63.65 | |
| top3Model=MM-RAIT (Qwen2.5-VL)2026.04 | 62 | 62.95 | |
| Ours-top1Model=Qwen2.5-VL2026.04 | 62 | 63.74 | |
| top1Model=MM-RAIT (Qwen2.5-VL)2026.04 | 60 | 62.44 | |
| top3Model=Qwen2.5-VL2026.04 | 60 | 62.26 | |
| top0Model=MM-RAIT (Qwen2.5-VL)2026.04 | 59 | 61.15 | |
| Ours-top1Model=LLaVA-NeXT2026.04 | 58 | 41.13 | |
| jinarerank-top1Model=Qwen2.5-VL2026.04 | 58 | 59.23 | |
| top1Model=Qwen2.5-VL2026.04 | 56 | 56.63 | |
| Ours-top3Model=LLaVA-NeXT2026.04 | 55 | 38.25 | |
| Ours-top3Model=InternVL-3.52026.04 | 55 | 55.04 | |
| jinarerank-top1Model=LLaVA-NeXT2026.04 | 52 | 34.93 | |
| Ours-top1Model=InternVL-3.52026.04 | 52 | 51.51 | |
| jinarerank-top3Model=InternVL-3.52026.04 | 52 | 52.02 | |
| top0Model=Qwen2.5-VL2026.04 | 51 | 50.67 | |
| top1Model=LLaVA-NeXT2026.04 | 50 | 33.18 | |
| jinarerank-top3Model=LLaVA-NeXT2026.04 | 50 | 34.26 | |
| top3Model=LLaVA-NeXT2026.04 | 47 | 31.14 | |
| top3Model=InternVL-3.52026.04 | 47 | 46.89 | |
| jinarerank-top1Model=InternVL-3.52026.04 | 45 | 44.81 | |
| top1Model=InternVL-3.52026.04 | 43 | 43.07 | |
| top0Model=InternVL-3.52026.04 | 41 | 38.96 | |
| top0Model=LLaVA-NeXT2026.04 | 39 | 35.76 |