Hallucination Evaluation on MME Total Scores
590MME Score (Zh)CLAIM
Evaluation Results
| Method | Links | |||||||||
|---|---|---|---|---|---|---|---|---|---|---|
| CLAIMBackbone=Qwen-VL-Chat2025.06 | 590 | 571.7 | 566.6 | 385 | — | — | 603 | — | 543.3 | |
| BaselineBackbone=Qwen-VL-Chat2025.06 | 553.3 | 646.7 | 548.3 | 500 | — | — | 315 | 601.7 | 503.7 | |
| VCDBackbone=Qwen-VL-Chat2025.06 | 553.3 | — | 556.7 | 503.3 | — | — | 295 | 576.7 | 497 | |
| CLAIMBackbone=LLaVA-1.52025.06 | 538.3 | 506.7 | 505 | 541.7 | 481.7 | — | — | — | 514.7 | |
| VCDBackbone=LLaVA-1.52025.06 | 396.7 | — | 441.3 | 343.3 | 393.3 | 378.3 | — | — | 390.6 | |
| BaselineBackbone=LLaVA-1.52025.06 | 378.3 | 628.3 | 383.3 | 333.3 | 395 | 336.7 | — | — | 365.3 |