Individualized Feedback Simulation on Individualized Feedback 210 instances, 50 players (test)
1.19MAEBG-Persona
Evaluation Results
| Method | Links | ||||||
|---|---|---|---|---|---|---|---|
| BG-Personadecoding=greedy2026.06 | 1.19 | 84.3 | 76 | 5.7 | 5 | 4.3 | |
| GPT-5.4decoding=greedy2026.06 | 1.24 | 73.4 | 68 | 5.4 | 4.7 | 3.9 | |
| Qwen3.5-397Bdecoding=greedy2026.06 | 1.45 | 64.6 | 60 | 5 | 4.2 | 3.4 | |
| Gemini-3.1-Flashdecoding=greedy2026.06 | 1.49 | 70.7 | 70 | 5.3 | 4.5 | 3.6 | |
| Qwen3.5-27Bdecoding=greedy2026.06 | 1.9 | 65.7 | 52 | 5 | 3.8 | 3.2 |