List your company
ResearchFormerly WizWand

Failure Detection on LIBERO seen

0.061Brier ScoreSAFE-RNN-TDQC (Ours)
0.027080.256040.4850.71396Apr 22, 2026
Updated 6mo ago

Evaluation results

MethodLinks
SAFE-RNN-TDQC (Ours)VLA Model=π02026.040.061
SAFE-RNN-TDQC (Ours)VLA Model=UniVLA2026.040.064
SAFE-MLP-TDQC (Ours)VLA Model=UniVLA2026.040.066
SAFE-MLP-TDQC (Ours)VLA Model=π02026.040.068
SAFE-MLP BCEVLA Model=π02026.040.075
SAFE-MLP BCEVLA Model=UniVLA2026.040.091
RNN-TDQC (Ours)VLA Model=UniVLA2026.040.1
SAFE-RNN-TDQC (Ours)VLA Model=π0-FAST2026.040.103
SAFE-MLP BCEVLA Model=π0-FAST2026.040.103
RNN-TDQC (Ours)VLA Model=π0-FAST2026.040.105
SAFE-RNNVLA Model=π0-FAST2026.040.106
SAFE-MLP-TDQC (Ours)VLA Model=π0-FAST2026.040.109
RNN-BCEVLA Model=π0-FAST2026.040.122
SAFE-RNNVLA Model=π02026.040.123
SAFE-RNNVLA Model=UniVLA2026.040.124
RNN-BCEVLA Model=UniVLA2026.040.138
RNN-TDQC (Ours)VLA Model=OpenVLA2026.040.191
SAFE-MLP BCEVLA Model=OpenVLA2026.040.192
SAFE-MLP-TDQC (Ours)VLA Model=OpenVLA2026.040.195
SAFE-RNN-TDQC (Ours)VLA Model=OpenVLA2026.040.197
RNN-BCEVLA Model=OpenVLA2026.040.199
SAFE-RNNVLA Model=OpenVLA2026.040.204
Avg entropyVLA Model=π0-FAST2026.040.209
Avg prob.VLA Model=π0-FAST2026.040.212
Running Avg prob.VLA Model=π0-FAST2026.040.244
Running Avg entropyVLA Model=OpenVLA2026.040.265
Running Avg entropyVLA Model=π0-FAST2026.040.279
Avg entropyVLA Model=OpenVLA2026.040.306
Max prob.VLA Model=π0-FAST2026.040.32
Running Avg prob.VLA Model=OpenVLA2026.040.338
Running Avg entropyVLA Model=UniVLA2026.040.343
Avg prob.VLA Model=OpenVLA2026.040.348
Max prob.VLA Model=OpenVLA2026.040.395
Avg entropyVLA Model=UniVLA2026.040.406
Avg prob.VLA Model=UniVLA2026.040.529
Running Avg prob.VLA Model=UniVLA2026.040.543
Max prob.VLA Model=UniVLA2026.040.909