Vision-Language Spatial Reasoning on WhatsUp B-LR 2x2 directional variants
82.84GroupScoreTTM
Evaluation results
| Method | Links | ||||
|---|---|---|---|---|---|
| TTMProtocol=Algorithm 12025.10 | 82.84 | 27 | 48.2 | 61.1 | |
| SimpleMatchProtocol=GroupMatch2025.10 | 55.88 | — | — | — | |
| CLIP-B32Type=Raw performance, Protocol=GroupScore2025.10 | 0 | — | — | — |