Failure Detection on VLABench (Seen Tasks)
85.6Balanced Accuracy (bACC)Hide-and-Seek
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Hide-and-SeekPolicy=π0.52026.05 | 85.6 | 82.7 | 66.2 | |
| SAFE-MLPPolicy=π0.52026.05 | 78.8 | 76.6 | 58.1 | |
| PCA KMeansPolicy=π0.52026.05 | 64.2 | 66.9 | 52.9 | |
| MahalanobisPolicy=π0.52026.05 | 63.7 | 51.9 | 58.1 | |
| Cluster EntropyPolicy=π0.52026.05 | 62.9 | 65.6 | 60.2 | |
| SAFE-LSTMPolicy=π0.52026.05 | 62.9 | 52.4 | 58.7 | |
| RNDPolicy=π0.52026.05 | 62.8 | 61.8 | 54.5 | |
| LogpZOPolicy=π0.52026.05 | 61.8 | 50.3 | 55.2 | |
| STACPolicy=π0.52026.05 | 61.5 | 58.6 | 55.6 | |
| ACEPolicy=π0.52026.05 | 59.6 | 54.5 | 55.3 | |
| EigenScorePolicy=π0.52026.05 | 55.3 | 48.7 | 53.2 | |
| Cosine k-NNPolicy=π0.52026.05 | 53.5 | 38.4 | 48.5 |