Vision-Language Spatial Reasoning on WhatsUp B-FB 2x2 directional variants
66.67GroupScoreTTM
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| TTMProtocol=Algorithm 12025.10 | 66.67 | 19.6 | 41.7 | 37 | |
| SimpleMatchProtocol=GroupMatch2025.10 | 47.06 | — | — | — | |
| CLIP-B32Type=Raw performance, Protocol=GroupScore2025.10 | 0 | — | — | — |