Closed-ended reasoning on Closed regime n = 300
87.9Exact Match AccuracyTransactional leadership
Evaluation Results
| Method | Links | |
|---|---|---|
| Transactional leadershipModel=gpt-oss-120b2026.06 | 87.9 | |
| FlatModel=gpt-oss-120b2026.06 | 86.5 | |
| accept-only ablationModel=gpt-oss-120b2026.06 | 86.5 | |
| Situational leadershipModel=gpt-oss-120b2026.06 | 86.5 | |
| directives-only ablationModel=gpt-oss-120b2026.06 | 85.5 | |
| Theory-free controlModel=gpt-oss-120b2026.06 | 85.2 | |
| Transformational leadershipModel=gpt-oss-120b2026.06 | 83.8 | |
| Theory-free controlModel=llama-4-scout2026.06 | 80.7 | |
| Transformational leadershipModel=llama-4-scout2026.06 | 79.3 | |
| FlatModel=gemma-4-31B-it2026.06 | 79.3 | |
| FlatModel=llama-4-scout2026.06 | 78.7 | |
| directives-only ablationModel=llama-4-scout2026.06 | 78.7 | |
| Situational leadershipModel=llama-4-scout2026.06 | 78.3 | |
| Theory-free controlModel=gemma-4-31B-it2026.06 | 78.3 | |
| directives-only ablationModel=gemma-4-31B-it2026.06 | 78.3 | |
| Situational leadershipModel=gemma-4-31B-it2026.06 | 78.3 | |
| Transactional leadershipModel=llama-4-scout2026.06 | 78 | |
| accept-only ablationModel=llama-4-scout2026.06 | 78 | |
| Transformational leadershipModel=gemma-4-31B-it2026.06 | 77.3 | |
| Transactional leadershipModel=gemma-4-31B-it2026.06 | 76 | |
| accept-only ablationModel=gemma-4-31B-it2026.06 | 76 |