Safe evaluation on Simulator safe (held-out)
21AccuracyDistilled-Prompt VLM
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Distilled-Prompt VLMScene=sim2026.06 | 21 | 0.429 | |
| Naked-Modality VLMScene=sim, Modality=proprio2026.06 | 12 | — |
Loading the SOTA2 catalog…
| Method | Links | ||
|---|---|---|---|
| Distilled-Prompt VLMScene=sim2026.06 | 21 | 0.429 | |
| Naked-Modality VLMScene=sim, Modality=proprio2026.06 | 12 | — |