Failure Detection on Franka unseen
0.215Brier ScoreSAFE-RNN-TDQC (Ours)
Evaluation Results
| Method | Links | |
|---|---|---|
| SAFE-RNN-TDQC (Ours)VLA Model=π0-FAST2026.04 | 0.215 | |
| RNN-TDQC (Ours)VLA Model=π0-FAST2026.04 | 0.228 | |
| SAFE-MLP-TDQC (Ours)VLA Model=π0-FAST2026.04 | 0.229 | |
| RNN-BCEVLA Model=π0-FAST2026.04 | 0.243 | |
| SAFE-MLP BCEVLA Model=π0-FAST2026.04 | 0.248 | |
| Avg entropyVLA Model=π0-FAST2026.04 | 0.281 | |
| SAFE-RNNVLA Model=π0-FAST2026.04 | 0.288 | |
| Avg prob.VLA Model=π0-FAST2026.04 | 0.294 | |
| Max prob.VLA Model=π0-FAST2026.04 | 0.323 | |
| Running Avg entropyVLA Model=π0-FAST2026.04 | 0.339 | |
| Running Avg prob.VLA Model=π0-FAST2026.04 | 0.361 |