Preference Alignment on Board Game Playtesting Dataset
0.6576MAEMeepleLM
Evaluation results
| Method | Links | |||
|---|---|---|---|---|
| MeepleLM2026.01 | 0.6576 | 0.2205 | 0.2817 | |
| MeepleLMAblation=w/o Rulebook2026.01 | 0.7043 | 0.5496 | 0.0026 | |
| MeepleLMAblation=w/o MDA2026.01 | 0.7395 | 0.4148 | 0.2271 | |
| MeepleLMAblation=w/o Persona2026.01 | 0.7887 | 0.363 | 0.1348 | |
| Qwen3-8B2026.01 | 0.8906 | 1.0119 | 0.0492 | |
| GPT-5.12026.01 | 0.9874 | 0.9496 | 0.2555 | |
| Qwen3-235B2026.01 | 1.2288 | 0.635 | 0.1449 | |
| Gemini3-Pro2026.01 | 1.4277 | 0.5092 | 0.2465 |