Support-faithfulness Evaluation on PRB (Patient Response Bank) Psychiatric Consultation Dialogues
8.6FC ScoreMIND
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| MINDinstruction-tuning=true, fine-tuned=true2026.03 | 8.6 | 8.8 | 8.3 | 8.6 | |
| DoctorAgent-RLfine-tuned=true2026.03 | 8.1 | 8.2 | 7.7 | 8 | |
| Qwen3-32B†instruction-tuning=true2026.03 | 8 | 8.1 | 7.8 | 8 | |
| DDOfine-tuned=true2026.03 | 8 | 8.3 | 7.9 | 8.1 | |
| Qwen3-8B†instruction-tuning=true2026.03 | 7.8 | 7.9 | 7.6 | 7.8 | |
| DeepSeek-V32026.03 | 7.4 | 6.9 | 7.1 | 7.1 | |
| MRD-RAG2026.03 | 7.3 | 7.4 | 6.7 | 7.1 | |
| Qwen3-32B2026.03 | 7.2 | 6.7 | 6.9 | 6.9 | |
| Baichuan-M22026.03 | 7.1 | 6.6 | 6.8 | 6.8 | |
| Qwen3-8B2026.03 | 7 | 6.5 | 6.8 | 6.8 |