Empathetic Response Generation on ReddiSupp User Study (test)
3.98EmpathyMindTailor
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| MindTailorBackbone=Qwen2.5-14B, Evaluation Protocol=User study2026.06 | 3.98 | 3.56 | 3.7 | 3.76 | 3.8 | 3.52 | 3.72 | 2.04 | |
| ES-VRBackbone=Qwen2.5-14B, Evaluation Protocol=User study2026.06 | 3.8 | 3.3 | 3.46 | 3.56 | 3.68 | 3.22 | 3.5 | 2.42 | |
| VanillaBackbone=Qwen2.5-14B, Evaluation Protocol=User study2026.06 | 3.76 | 2.82 | 3.12 | 3.42 | 3.28 | 2.92 | 3.22 | 2.96 | |
| MentalAgoraBackbone=Qwen2.5-14B, Evaluation Protocol=User study2026.06 | 3.24 | 3.56 | 3.4 | 3.68 | 3.48 | 3.16 | 3.42 | 2.58 |