Multi-agent debate on Social regime
51.3AccuracySituational leadership controller
Evaluation Results
| Method | Links | |
|---|---|---|
| Situational leadership controllerModel=llama-4-scout2026.06 | 51.3 | |
| Transformational leadership controllerModel=llama-4-scout2026.06 | 46.7 | |
| FlatModel=llama-4-scout2026.06 | 45.3 | |
| MADModel=llama-4-scout2026.06 | 45 | |
| R0-voteModel=llama-4-scout2026.06 | 43.7 | |
| Transformational leadership controllerModel=gpt-oss-120b2026.06 | 43 | |
| Situational leadership controllerModel=gpt-oss-120b2026.06 | 42.7 | |
| R0-voteModel=gpt-oss-120b2026.06 | 41.3 | |
| FlatModel=gpt-oss-120b2026.06 | 41 | |
| MADModel=gpt-oss-120b2026.06 | 41 | |
| Situational leadership controllerModel=gemma-4-31B-it2026.06 | 40 | |
| R0-voteModel=gemma-4-31B-it2026.06 | 39 | |
| FlatModel=gemma-4-31B-it2026.06 | 39 | |
| Transformational leadership controllerModel=gemma-4-31B-it2026.06 | 38.7 | |
| MADModel=gemma-4-31B-it2026.06 | 37.7 |