Vision-Language Spatial Reasoning on WhatsUp A-LR 2x2 directional variants
95.87GroupScoreTTM
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| TTMProtocol=Algorithm 12025.10 | 95.87 | 55.1 | 135.1 | 93 | |
| SimpleMatchProtocol=GroupMatch2025.10 | 40.78 | — | — | — | |
| CLIP-B32Type=Raw performance, Protocol=GroupScore2025.10 | 0 | — | — | — |