Procedural Mistake Detection on Epic-tent-O
0.75PrecisionBERT
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| BERTStep Recog.=Oracle2024.04 | 0.75 | 0.056 | 0.104 | |
| PREGOStep Recog.=Oracle, Step Antic.=LLAMA2024.04 | 0.107 | 0.867 | 0.191 | |
| PREGOStep Recog.=Oracle, Step Antic.=GPT-3.52024.04 | 0.099 | 0.733 | 0.174 | |
| PREGOStep Recog.=OadTR [35], Step Antic.=LLAMA2024.04 | 0.095 | 0.933 | 0.172 | |
| PREGOStep Recog.=MiniRoad [2], Step Antic.=LLAMA2024.04 | 0.086 | 0.2 | 0.12 | |
| OadTR for MDStep Recog.=OadTR [35], Step Antic.=OadTR [35]2024.04 | 0.067 | 0.217 | 0.102 | |
| One-step memoryStep Recog.=Oracle2024.04 | 0.066 | 0.266 | 0.106 | |
| PREGOStep Recog.=MiniRoad [2], Step Antic.=GPT-3.52024.04 | 0.043 | 0.666 | 0.08 |