Share your thoughts, 1 month free Claude Pro on usSee more
WorkDL logo mark

Early Failure Detection on RoboCasa S (Unseen)

73.8ROC-AUC (q=0.25)

ActProbe

43.22451.16259.167.038Jun 7, 2026
Updated 1mo ago

Evaluation Results

MethodLinks
2026.06
73.8
2026.06
73.1
2026.06
71
2026.06
70.4
2026.06
67.2
2026.06
66
2026.06
62.4
2026.06
44.4