LLM Alignment on HelpSteer (test)
8.34AlpacaEval 2 WRPD
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| PDModel=Qwen2.5-7B, Strategy Group=SELT2025.08 | 8.34 | 12.98 | — | — | |
| PD (rati.)Model=Qwen2.5-7B, Strategy Group=SELT2025.08 | 7.97 | 12.73 | — | — | |
| DMPOModel=Qwen2.5-7B, Strategy Group=FULL2025.08 | 7.6 | 12.26 | — | — | |
| AVG.Model=Qwen2.5-7B, Strategy Group=FULL2025.08 | 7.59 | 12.19 | — | — | |
| PD (ours)Model=Llama3.1-8B, Category=SELT2025.08 | 7.55 | 12.13 | 1.19 | 8.7 | |
| RAFModel=Llama3.1-8B, Category=SELT2025.08 | 6.58 | 10.46 | 1.1 | 6.6 | |
| PD (rati.)Model=Llama3.1-8B, Category=SELT2025.08 | 6.57 | 12.41 | 1.13 | 5 | |
| RAFModel=Qwen2.5-7B, Strategy Group=SELT2025.08 | 6.57 | 10.98 | — | — | |
| RANDModel=Qwen2.5-7B, Strategy Group=SELT2025.08 | 6.38 | 10.69 | — | — | |
| AVG.Model=Llama3.1-8B, Category=FULL2025.08 | 6.18 | 9.55 | 1.03 | 10 | |
| ALLModel=Qwen2.5-7B, Strategy Group=FULL2025.08 | 6.17 | 10.63 | — | — | |
| DMPOModel=Llama3.1-8B, Category=FULL2025.08 | 6.01 | 10.35 | 1.09 | 13.7 | |
| OVA.Model=Qwen2.5-7B, Strategy Group=FULL2025.08 | 5.87 | 11.71 | — | — | |
| OVA.Model=Llama3.1-8B, Category=FULL2025.08 | 5.75 | 9.95 | 1 | 10 | |
| ALLModel=Llama3.1-8B, Category=FULL2025.08 | 5.55 | 9.97 | 1.02 | 10 | |
| RANDModel=Llama3.1-8B, Category=SELT2025.08 | 5.53 | 8.5 | 0.98 | 5 | |
| SFTModel=Llama3.1-8B, Category=INIT2025.08 | 4.52 | 8.25 | 0.76 | 0 | |
| SFTModel=Qwen2.5-7B, Strategy Group=INIT2025.08 | 4.36 | 9.12 | — | — | |
| PDModel=Qwen2.5-7B, Strategy Group=SELT2025.08 | 1.22 | — | — | — | |
| PD (rati.)Model=Qwen2.5-7B, Strategy Group=SELT2025.08 | 1.21 | — | — | — | |
| DMPOModel=Qwen2.5-7B, Strategy Group=FULL2025.08 | 1.08 | — | — | — | |
| RAFModel=Qwen2.5-7B, Strategy Group=SELT2025.08 | 1.05 | — | — | — | |
| AVG.Model=Qwen2.5-7B, Strategy Group=FULL2025.08 | 1.03 | — | — | — | |
| ALLModel=Qwen2.5-7B, Strategy Group=FULL2025.08 | 1.03 | — | — | — | |
| OVA.Model=Qwen2.5-7B, Strategy Group=FULL2025.08 | 1 | — | — | — | |
| RANDModel=Qwen2.5-7B, Strategy Group=SELT2025.08 | 0.98 | — | — | — | |
| SFTModel=Qwen2.5-7B, Strategy Group=INIT2025.08 | 0.77 | — | — | — |