Alignment Evaluation on Conflict resolution questions
62.6Win RateSingle-Agent
Evaluation Results
| Method | Links | |
|---|---|---|
| Single-AgentComparison=vs. Base, Decoding strategy=greedy2026.03 | 62.6 | |
| Multi-AgentComparison=vs. Base, Decoding strategy=sampling2026.03 | 62.2 | |
| Multi-AgentComparison=vs. Base, Decoding strategy=greedy2026.03 | 59.4 | |
| Single-AgentComparison=vs. Base, Decoding strategy=sampling2026.03 | 58 | |
| Multi-AgentComparison=vs. Single-Agent, Decoding strategy=sampling2026.03 | 51.4 | |
| Multi-AgentComparison=vs. Single-Agent, Decoding strategy=greedy2026.03 | 49.1 |