Offline Action Prediction on Mind2Web Cross-Domain v1.0 (test)
50.2Element AccuracyHTML-T5-XL
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| HTML-T5-XLTrain=SL2023.07 | 50.2 | 74.9 | 48.3 | |
| MindAct (Flan-T5-XL)Train=SL, LLM=Flan-T5-XL2023.07 | 42.1 | 66.5 | 39.6 | |
| MindAct (GPT-4)Train=ICL, LLM=GPT-42023.07 | 37.1 | 46.5 | 26.4 | |
| Synapse (GPT-3.5)Train=ICL, LLM=GPT-3.52023.07 | 29.4 | — | 25.9 |