Personalization on TruthfulQA
56Creative Score (ArmoRM)Amulet
Evaluation Results
| Method | Links | ||||
|---|---|---|---|---|---|
| AmuletBackbone=Llama-3.1-8B2025.09 | 56 | 44 | 52 | 49 | |
| T-POPBackbone=Qwen2-7B2025.09 | 53 | 47 | 53 | 58 | |
| T-POPBackbone=Llama-3.1-8B2025.09 | 52 | 54 | 51 | 53 | |
| T-POPBackbone=Mistral-7B2025.09 | 51 | 43 | 54 | 54 | |
| AmuletBackbone=Mistral-7B2025.09 | 41 | 32 | 49 | 47 | |
| LABackbone=Llama-3.1-8B2025.09 | 41 | 34 | 49 | 43 | |
| LABackbone=Mistral-7B2025.09 | 38 | 33 | 51 | 47 | |
| AmuletBackbone=Qwen2-7B2025.09 | 36 | 32 | 51 | 45 | |
| BS16Backbone=Mistral-7B2025.09 | 34 | 31 | 47 | 39 | |
| LABackbone=Qwen2-7B2025.09 | 34 | 33 | 54 | 44 | |
| PrefBackbone=Mistral-7B2025.09 | 33 | 31 | 44 | 38 | |
| PrefBackbone=Qwen2-7B2025.09 | 33 | 31 | 46 | 38 | |
| BS16Backbone=Qwen2-7B2025.09 | 33 | 32 | 50 | 39 | |
| PrefBackbone=Llama-3.1-8B2025.09 | 33 | 29 | 40 | 36 | |
| BS16Backbone=Llama-3.1-8B2025.09 | 33 | 31 | 42 | 37 | |
| BaseBackbone=Mistral-7B2025.09 | 32 | 30 | 41 | 36 | |
| BaseBackbone=Qwen2-7B2025.09 | 32 | 30 | 41 | 36 | |
| BaseBackbone=Llama-3.1-8B2025.09 | 31 | 29 | 37 | 34 |