Motivational Interviewing Dialogue Generation on StoryMI 1.0 (All 1,000 questionnaires)
90.7Dialogue EntropyGPT-5-Nano
Evaluation results
| Method | Links | ||||||||||
|---|---|---|---|---|---|---|---|---|---|---|---|
| GPT-5-NanoFramework=StoryMI2026.04 | 90.7 | 84.3 | 12.2 | 16.7 | 80 | 80.9 | 72.5 | 98.3 | 6.9 | 3.6 | |
| OpenChat-7BFramework=StoryMI2026.04 | 89.4 | 69.8 | 7.7 | 39.3 | 86.6 | 81.3 | 67 | 82.6 | 11.7 | 1.9 | |
| Gemma-7BFramework=StoryMI2026.04 | 89.3 | 64 | 7.3 | 49.3 | 79.7 | 80.1 | 72.1 | 97.4 | 29 | 4.4 | |
| Qwen 2.5-7BFramework=StoryMI2026.04 | 89.1 | 73 | 10 | 34.2 | 72.3 | 81 | 75.9 | 99 | 19.8 | 6.3 | |
| Overall (Mean)Framework=StoryMI2026.04 | 89.1 | 73.9 | 10.9 | 32 | 81.6 | 71.6 | 62.6 | 83 | 42.4 | 2.8 | |
| Phi-4-14BFramework=StoryMI2026.04 | 88.6 | 76.8 | 17.1 | 25.5 | 87.9 | 53 | 42.5 | 58 | 95.2 | 0.5 | |
| LLaMA 3.1-8BFramework=StoryMI2026.04 | 87.5 | 75.7 | 10.9 | 27 | 83.2 | 53 | 45.7 | 62.2 | 91.8 | 0.1 |