Stylized Dialogue on 148-query style-3 persona (test)
4.696ContextSFR
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| SFRTemperature=0.8, Backbone=Qwen3-32B, LLM-judge=Qwen3-235B2026.05 | 4.696 | 4.851 | 4.73 | 4.98 | |
| SFTTemperature=0.8, Backbone=Qwen3-32B, LLM-judge=Qwen3-235B2026.05 | 4.568 | 4.818 | 3.932 | 4.831 | |
| DeepSeek-R1-promptTemperature=0.8, Backbone=DeepSeek-R1 (frozen), LLM-judge=Qwen3-235B2026.05 | 4.511 | 4.759 | 3.891 | 4.796 |