Social Simulation on Sotopia standard (test)
8.92Goal ScoreThe Stackelberg Speaker + MetaMind
Evaluation results
| Method | Links | ||
|---|---|---|---|
| The Stackelberg Speaker + MetaMindBase Agent=MetaMind2025.10 | 8.92 | 4.2 | |
| MetaMind2025.10 | 8.7 | 4.03 | |
| The Stackelberg Speaker + ReActBase Agent=ReAct2025.10 | 8.58 | 3.95 | |
| ReCon2025.10 | 8.45 | 3.88 | |
| ReAct2025.10 | 8.43 | 3.85 | |
| SDPO2025.10 | 8.13 | 3.63 | |
| SOTOPIA-Ω2025.10 | 8.07 | 3.67 | |
| SOTOPIA-π2025.10 | 7.62 | 3.44 |