Expert Evaluation on AnnoMI Expert Evaluation Subset
4.06Cultivating Change TalkHQ (High-quality human sessions)
Evaluation Results
| Method | Links | ||||||||
|---|---|---|---|---|---|---|---|---|---|
| HQ (High-quality human sessions)Type=Human Counselor2025.02 | 4.06 | 3.9 | 4.26 | 4.26 | 4.18 | 2.68 | 4.68 | 4.68 | |
| CAMIFramework=STAR framework2025.02 | 3.68 | 3.32 | 3.9 | 4 | 3.94 | 2.4 | 3.6 | 4.32 | |
| CoSDescription=Chain-of-Strategy baseline2025.02 | 2.74 | 2.74 | 3.6 | 3.72 | 3 | 1.8 | 3.06 | 4 | |
| LQ (Low-quality human sessions)Type=Human Counselor2025.02 | 1.68 | 1.74 | 1.46 | 1.38 | 1.58 | 1.2 | 2.32 | 4 |