Dialogue Experience Evaluation on 20 popular topics 50 dialogues per topic
88.6InterestingStephanie2
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| Stephanie2Backbone Model=GPT5.2, Dialogue System Variant=Stephanie22026.01 | 88.6 | 84.4 | 91.9 | 92 | 89.7 | 86.5 | 89.8 | 89 | |
| Stephanie1Backbone Model=GPT5.2, Dialogue System Variant=Stephanie12026.01 | 86.9 | 83.8 | 91 | 90.8 | 88.4 | 84.9 | 82.3 | 86.9 | |
| Stephanie2Backbone Model=Deepseek-V3, Dialogue System Variant=Stephanie22026.01 | 86.1 | 78.4 | 89.1 | 89.6 | 86.4 | 83.7 | 88.1 | 85.9 | |
| Stephanie1Backbone Model=Deepseek-V3, Dialogue System Variant=Stephanie12026.01 | 83.3 | 76.5 | 87.5 | 87 | 84.5 | 80.6 | 78.7 | 82.6 | |
| PDBackbone Model=Deepseek-V3, Dialogue System Variant=PD2026.01 | 75.9 | 72.1 | 81.1 | 79 | 79.1 | 78 | 72.2 | 76.8 | |
| PDBackbone Model=GPT5.2, Dialogue System Variant=PD2026.01 | 75.2 | 81.3 | 81.5 | 79.2 | 81.8 | 84.8 | 72.7 | 79.5 | |
| Stephanie2Backbone Model=Llama3.1-8B, Dialogue System Variant=Stephanie22026.01 | 72.4 | 69.7 | 78.4 | 76.2 | 76.7 | 77.5 | 79.1 | 75.7 | |
| Stephanie1Backbone Model=Llama3.1-8B, Dialogue System Variant=Stephanie12026.01 | 68.8 | 68.9 | 75.3 | 72.6 | 73.4 | 74.1 | 68.3 | 71.6 | |
| PDBackbone Model=Llama3.1-8B, Dialogue System Variant=PD2026.01 | 59.1 | 65.4 | 67.7 | 62.5 | 66.6 | 71.5 | 61.7 | 64.9 |