Vision-Language Spatial Reasoning on WhatsUp A-OU 2x2 directional variants
99.03GroupScoreTTM
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| TTMProtocol=Algorithm 12025.10 | 99.03 | 20.4 | 25.9 | 95.5 | |
| SimpleMatchProtocol=GroupMatch2025.10 | 78.64 | — | — | — | |
| CLIP-B32Type=Raw performance, Protocol=GroupScore2025.10 | 3.88 | — | — | — |