Reward Modeling on PPE (Specific Metrics)
76.9PPE Human PreferenceSkywork-Reward-V2
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| Skywork-Reward-V2Backbone=gemma-2-9b-it2025.07 | 76.9 | 82 | |
| Skywork-Reward-V2Backbone=gemma-3-1b-it2025.07 | 70.1 | 73.8 | |
| Skywork-Reward-V2Backbone=gemma-3-4b2025.07 | 68.9 | 73.7 | |
| Skywork-Reward-V2Backbone=gemma-2-2b-it2025.07 | 67.9 | 71.2 | |
| Skywork-Reward-V2Backbone=Qwen2.5-7B2025.07 | 66.4 | 73.9 | |
| INF-ORM-Llama3.1-70B2025.07 | 64.2 | 64.4 | |
| CIR-AMS/BTRM_Qwen2_7b_06132025.07 | 60 | 63.1 | |
| Ray2333/GRM-gemma2-2B-rewardmodel-ft2025.07 | 55.4 | 62 |