Human Evaluation on ScienceVQA (test)
8.35Faithfulness ScoreMMBoundary
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| MMBoundary2025.05 | 8.35 | 7.46 | 8.02 | 7.94 | |
| RCE2025.05 | 7.55 | 6.92 | 7.12 | 7.19 | |
| SaySelf2025.05 | 7.49 | 7.18 | 6.28 | 6.98 | |
| DRL2025.05 | 6.81 | 5.97 | 6.34 | 6.37 | |
| Conf-CSR2025.05 | 6.74 | 5.61 | 6.4 | 6.23 | |
| Multisample2025.05 | 4.77 | 5.24 | 4.53 | 4.85 |