Pairwise Preference Comparison on Qwen2.5-3B responses (test)
82.7Avg Preference ScoreNASH-MD-PG
Evaluation Results
| Method | Links | |
|---|---|---|
| NASH-MD-PGComparison model=RLOO2025.12 | 82.7 | |
| STACKELBERGGDARole=Follower, Comparison model=QWEN2.5-3B2025.12 | 82.1 | |
| NASH-MD-PGComparison model=QWEN2.5-0.5B2025.12 | 81.3 | |
| STACKELBERGGDARole=Follower, Comparison model=QWEN2.5-0.5B2025.12 | 81.3 | |
| NASH-MD-PGComparison model=QWEN2.5-3B2025.12 | 80.1 | |
| STACKELBERGGDARole=Follower, Comparison model=RLOO2025.12 | 79.9 | |
| STACKELBERGGDARole=Leader, Comparison model=QWEN2.5-0.5B2025.12 | 69.6 | |
| STACKELBERGGDARole=Follower, Comparison model=STACKELBERGGDA LEADER2025.12 | 68.8 | |
| STACKELBERGGDARole=Leader, Comparison model=QWEN2.5-3B2025.12 | 68.1 | |
| STACKELBERGGDARole=Leader, Comparison model=RLOO2025.12 | 66.2 | |
| NASH-MD-PGComparison model=STACKELBERGGDA LEADER2025.12 | 63.8 | |
| RLOOComparison model=QWEN2.5-0.5B2025.12 | 60.1 | |
| RLOOComparison model=QWEN2.5-3B2025.12 | 58.8 | |
| NASH-MD-PGComparison model=STACKELBERGGDA FOLLOWER2025.12 | 50.7 | |
| QWEN2.5-0.5BComparison model=QWEN2.5-3B2025.12 | 50.4 | |
| QWEN2.5-3BComparison model=QWEN2.5-0.5B2025.12 | 49.6 | |
| STACKELBERGGDARole=Follower, Comparison model=NASH-MD-PG2025.12 | 49.3 | |
| QWEN2.5-3BComparison model=RLOO2025.12 | 41.2 | |
| QWEN2.5-0.5BComparison model=RLOO2025.12 | 39.9 | |
| STACKELBERGGDARole=Leader, Comparison model=NASH-MD-PG2025.12 | 36.2 | |
| RLOOComparison model=STACKELBERGGDA LEADER2025.12 | 33.8 | |
| QWEN2.5-3BComparison model=STACKELBERGGDA LEADER2025.12 | 31.9 | |
| STACKELBERGGDARole=Leader, Comparison model=STACKELBERGGDA FOLLOWER2025.12 | 31.2 | |
| QWEN2.5-0.5BComparison model=STACKELBERGGDA LEADER2025.12 | 30.4 | |
| RLOOComparison model=STACKELBERGGDA FOLLOWER2025.12 | 20.1 | |
| QWEN2.5-3BComparison model=NASH-MD-PG2025.12 | 19.9 | |
| QWEN2.5-0.5BComparison model=NASH-MD-PG2025.12 | 18.7 | |
| QWEN2.5-0.5BComparison model=STACKELBERGGDA FOLLOWER2025.12 | 18.7 | |
| QWEN2.5-3BComparison model=STACKELBERGGDA FOLLOWER2025.12 | 17.9 | |
| RLOOComparison model=NASH-MD-PG2025.12 | 17.3 |