Failure Detection on RoboFail (Out-Of-Domain)
86Execution Binary AccGuardian-8B-Thinking
Evaluation results
| Method | Links | ||
|---|---|---|---|
| Guardian-8B-ThinkingModel Category=Proposed Method, Strategy=Thinking, Base Model=InternVL3-8B2025.12 | 86 | 70 | |
| Qwen3-VL-235B-A22BModel Category=Large-scale generalist models2025.12 | 82 | 70 | |
| GPT4.1Model Category=Large-scale generalist models2025.12 | 82 | 67 | |
| SentinelModel Category=Specialized / small models2025.12 | 79 | — | |
| InternVL3-8BModel Category=Specialized / small models2025.12 | 77 | 53 | |
| Cosmos-Reason1-7BModel Category=Specialized / small models2025.12 | 67 | 70 | |
| AHA-13BModel Category=Specialized / small models2025.12 | 64 | — | |
| CLIP+MLPModel Category=Specialized / small models2025.12 | 42 | 43 |