GUI Automation on MiniWoB++ All - 53 types (test)
900Token UsageAutoRPA (code only)
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| AutoRPA (code only)LLM Backbone=GPT-4.12026.05 | 900 | 92.5 | |
| AutoRPALLM Backbone=GPT-4.12026.05 | 1,400 | 95.4 | |
| AdaPlanner (one demo)LLM Backbone=GPT-4.1, human examples=12026.05 | 4,500 | 74.3 | |
| AutoManualLLM Backbone=GPT-4.12026.05 | 4,600 | 95.2 | |
| AdaPlannerLLM Backbone=GPT-4.12026.05 | 6,100 | 90.3 | |
| ReAct†LLM Backbone=GPT-4.12026.05 | 9,200 | 92.8 | |
| RCILLM Backbone=GPT-4.12026.05 | 10,200 | 87.2 |