Role-playing on RoleBench Chinese (instruction generalization)
36.4Win Rate (vs GPT-4)RoleGLM
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| RoleGLM2023.10 | 36.4 | 52.4 | |
| ChatPLUG2023.10 | 28.9 | 19.9 | |
| Character.AI2023.10 | 28.2 | 19 | |
| ChatGLM22023.10 | 24.2 | 19.6 |
Loading the SOTA2 catalog…