Instruction Following Evaluation on FollowBench OOD
78.06HSRPrompt
Evaluation Results
| Method | Links | ||
|---|---|---|---|
| PromptModel=Deepseek-V4-Flash, Category=Vanilla2026.06 | 78.06 | 84.62 | |
| PromptModel=Qwen3-32B, Category=Vanilla2026.06 | 64.66 | 77.12 | |
| EvoRubricsModel=Qwen3-8B, Category=Ours2026.06 | 63.01 | 74.7 | |
| OnlineRubricsModel=Qwen3-8B, Category=RL w/ Rubric2026.06 | 60.73 | 74.9 | |
| SFTModel=Qwen3-8B, Category=Supervised2026.06 | 59.33 | 72.4 | |
| GoldenRubricsModel=Qwen3-8B, Category=RL w/ Rubric2026.06 | 58.31 | 70.61 | |
| EvoRubricsModel=Qwen3-4B, Category=Ours2026.06 | 53.92 | 66.2 | |
| PromptModel=Qwen3-4B, Category=Vanilla2026.06 | 47.76 | 64.09 | |
| RuscaRLModel=Qwen3-8B, Category=RL w/ Rubric2026.06 | 39.34 | 61.59 | |
| SFTModel=Qwen3-4B, Category=Supervised2026.06 | 38.76 | 56.31 | |
| GoldenRubricsModel=Qwen3-4B, Category=RL w/ Rubric2026.06 | 32.5 | 50.4 | |
| RuscaRLModel=Qwen3-4B, Category=RL w/ Rubric2026.06 | 31.89 | 52.1 | |
| PromptModel=Qwen3-8B, Category=Vanilla2026.06 | 23.89 | 36.49 | |
| OnlineRubricsModel=Qwen3-4B, Category=RL w/ Rubric2026.06 | 23.66 | 41.47 |