Dialogue-level Guidance Quality Evaluation on Dialogue-level evaluation (N=54)
3.3State TrackingVIGiA
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| VIGiA2026.02 | 3.3 | 4.06 | 4.26 | |
| Qwen 3 VL2026.02 | 2.85 | 2.35 | 3.39 | |
| InternVL 3.52026.02 | 2.52 | 2.44 | 3.13 | |
| Qwen 2.5 VL2026.02 | 2.31 | 2.41 | 2.74 | |
| Llava OV2026.02 | 1.87 | 2.59 | 3 | |
| MM-PlanLLM2026.02 | 1.61 | 2.11 | 2.24 |