Multi-party cognitive stimulation dialogue on Real-world Cognitive Stimulation Dialogue dataset (test)
75Win RateGCSD-3b
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| GCSD-3bbaseline_comparison=ERNIE2026.03 | 75 | 10 | 15 | |
| GCSD-3bbaseline_comparison=GLM-4-Plus2026.03 | 73 | 12 | 15 | |
| GCSD-3bbaseline_comparison=Doubao-Pro2026.03 | 60 | 23 | 17 | |
| GCSD-3bbaseline_comparison=ChatGPT2026.03 | 58 | 27 | 15 | |
| GCSD-3bbaseline_comparison=LLaMA-3.1-405B2026.03 | 58 | 34 | 8 | |
| GCSD-3bbaseline_comparison=Qwen-Max2026.03 | 55 | 29 | 16 | |
| GCSD-3bbaseline_comparison=GPT-4o2026.03 | 50 | 39 | 11 | |
| GCSD-3bbaseline_comparison=DeepSeek-671B2026.03 | 43 | 40 | 17 |