Share your thoughts, 1 month free Claude Pro on us
See more
Home
/
Benchmarks
Early Failure Detection on RoboCasa S (Unseen)
Loading...
73.8
ROC-AUC (q=0.25)
ActProbe
43.224
51.162
59.1
67.038
Jun 7, 2026
ROC-AUC (q=0.25)
Updated 1mo ago
Evaluation Results
Method
Method
Links
ROC-AUC (q=0.25)
ActProbe
Policy=π0.5
2026.06
73.8
SAFE-LSTM
Policy=π0.5
2026.06
73.1
SAFE-MLP
Policy=π0.5
2026.06
71
SAFE-MLP-TDQC
Policy=π0.5
2026.06
70.4
LogpZO
Policy=π0.5
2026.06
67.2
STAC-Single
Policy=π0.5
2026.06
66
SAFE-LSTM-TDQC
Policy=π0.5
2026.06
62.4
Cosine k-NN
Policy=π0.5
2026.06
44.4
Feedback
Search any
task
Search any
task