Conversational Search on Qulac
33.14ROUGE-LRAC_SFT
Evaluation results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| RAC_SFTTraining=Supervised Fine-Tuning2026.01 | 33.14 | 12.59 | 31.3 | 89.34 | 79.14 | 42.53 | |
| RAC_DPOTraining=Direct Preference Optimization2026.01 | 32.42 | 11.52 | 31.48 | 88.92 | 81.73 | 44.83 | |
| Q-Cond2026.01 | 29.44 | 10.51 | 25.92 | 88.24 | — | — | |
| QP-ZeroshotMode=Zero-shot2026.01 | 27.39 | 5.68 | 33.33 | 87.2 | — | — | |
| AT-CoTReasoning=Chain-of-Thought2026.01 | 17.97 | 2.77 | 20.81 | 84.72 | — | — |