Web Navigation on MiniWob++
53.26AccuracyExplorer-7B
Evaluation Results
| Method | Links | |
|---|---|---|
| Explorer-7BParameters=7B, zero-shot=true2025.02 | 53.26 | |
| GPT-4Model Type=API-based, zero-shot=true2025.02 | 53.04 | |
| Llama3-chat-70BModel Type=Open-source Instructed, Parameters=70B, zero-shot=true2025.02 | 48.7 | |
| Explorer-4BParameters=4B, zero-shot=true2025.02 | 46.74 | |
| AgentTrek-7BModel Type=Open-source Interactive Data Finetuned, Parameters=7B, zero-shot=true2025.02 | 45.28 | |
| GPT-3.5Model Type=API-based, zero-shot=true2025.02 | 39.57 | |
| Synatra-CodeLlama-7BModel Type=Open-source Interactive Data Finetuned, Parameters=7B, zero-shot=true2025.02 | 38.2 | |
| Qwen2-VL-7BModel Type=Open-source Instructed, Parameters=7B, zero-shot=true2025.02 | 36.96 | |
| AgentLM-70BModel Type=Open-source Interactive Data Finetuned, Parameters=70B, zero-shot=true2025.02 | 36.52 | |
| Phi-3.5VModel Type=Open-source Instructed, zero-shot=true2025.02 | 35.87 | |
| Llama3-chat-8BModel Type=Open-source Instructed, Parameters=8B, zero-shot=true2025.02 | 31.74 | |
| Lemur-chat-70BModel Type=Open-source Interactive Data Finetuned, Parameters=70B, zero-shot=true2025.02 | 21.3 | |
| AgentFlan-7BModel Type=Open-source Interactive Data Finetuned, Parameters=7B, zero-shot=true2025.02 | 20.87 | |
| AgentLM-7BModel Type=Open-source Interactive Data Finetuned, Parameters=7B, zero-shot=true2025.02 | 15.65 | |
| CodeActAgent-7BModel Type=Open-source Interactive Data Finetuned, Parameters=7B, zero-shot=true2025.02 | 9.78 |