Failure Taxonomy Recovery on RoboFail expert taxonomy
92CPUnsupervised Failure Taxonomy Discovery Framework
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Unsupervised Failure Taxonomy Discovery FrameworkMode=Aggregation2025.06 | 92 | 100 | 95.8 | |
| BERTopic-LLM2025.06 | 87.5 | 87.5 | 87.5 | |
| Unsupervised Failure Taxonomy Discovery FrameworkMode=Single Run2025.06 | 81.8 | 90 | 84.9 | |
| BERTopic2025.06 | 78.5 | 62.5 | 69.6 |