| visual-reasoning-on-winoground | BLIP | – | Text Score: 39.0Image Score: 19.2Group Score: 15.0 |
| visual-reasoning-on-winoground | BLIP (+Graph Text) | – | Text Score: 40.3Image Score: 20.5Group Score: 16.5 |
| visual-reasoning-on-winoground | BLIP (+Graph Text, +Graph Neg) | – | Text Score: 40.5Image Score: 25.5Group Score: 19.0 |
| visual-reasoning-on-winoground | BLIP (SGVL) | – | Text Score: 42.8Image Score: 27.3Group Score: 21.5 |
| visual-reasoning-on-winoground | BLIP2 | – | Text Score: 42.0Image Score: 23.8Group Score: 19.0 |
| visual-reasoning-on-winoground | BLIP2 (SGVL) | – | Text Score: 42.8Image Score: 28.5Group Score: 23.3 |
| visual-reasoning-on-winoground | CLIP (SGVL) | – | Text Score: 32.0Image Score: 14.0Group Score: 9.8 |
| visual-reasoning-on-winoground | LLaVA | – | Text Score: 24.8Image Score: 25.0Group Score: 13.0 |
| visual-reasoning-on-winoground | MiniGPT-4 | – | Text Score: 23.3Image Score: 18.0Group Score: 9.5 |
| visual-reasoning-on-winoground | NegBLIP | – | Text Score: 42.5Image Score: 24.0Group Score: 18.5 |
| visual-reasoning-on-winoground | NegBLIP2 | – | Text Score: 41.5Image Score: 26.0Group Score: 20.5 |
| visual-reasoning-on-winoground | NegCLIP | – | Text Score: 29.5Image Score: 10.5Group Score: 8.0 |