Information Following on InfoBench
89.2Easy ScoreQwen3-14B + RL w/ PaTaRM
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Qwen3-14B + RL w/ PaTaRMTraining=RL w/ PaTaRM2025.10 | 89.2 | 89.2 | 89.2 | |
| Qwen3-14B + DPOTraining=DPO2025.10 | 88.7 | 86.5 | 87.2 | |
| GPT-4o2025.10 | 87.9 | 87.6 | 87.1 | |
| Qwen3-14B + SFTTraining=SFT2025.10 | 87.4 | 86 | 86.4 | |
| Qwen3-14B + RL w/ SkyworkTraining=RL w/ Skywork2025.10 | 87.1 | 88.1 | 87.8 | |
| Qwen2.5-7B-Instruct + RL w/ PaTaRMTraining=RL w/ PaTaRM2025.10 | 86.6 | 82.8 | 83.9 | |
| Qwen3-8B + SFTTraining=SFT2025.10 | 86.3 | 84 | 84.7 | |
| Qwen3-14BTraining=Base2025.10 | 86.3 | 86.9 | 86.7 | |
| Qwen3-8BTraining=Base2025.10 | 86.2 | 85.4 | 85.6 | |
| Qwen3-8B + RL w/ PaTaRMTraining=RL w/ PaTaRM2025.10 | 86 | 87.7 | 87.2 | |
| Qwen3-8B + RL w/ SkyworkTraining=RL w/ Skywork2025.10 | 85.9 | 85.6 | 85.7 | |
| Qwen2.5-7B-Base + DPOTraining=DPO2025.10 | 85.6 | 77.2 | 79.8 | |
| Qwen2.5-7B-Instruct + SFTTraining=SFT2025.10 | 85.4 | 79.4 | 81.2 | |
| Qwen3-8B + DPOTraining=DPO2025.10 | 85.4 | 85.1 | 85.2 | |
| Qwen2.5-7B-Instruct + RL w/ SkyworkTraining=RL w/ Skywork2025.10 | 84.8 | 82.2 | 83 | |
| Qwen2.5-7B-Base + RL w/ PaTaRMTraining=RL w/ PaTaRM2025.10 | 83.7 | 84.6 | 84.3 | |
| Qwen2.5-7B-InstructTraining=Instruct2025.10 | 83.2 | 78.6 | 80 | |
| Qwen2.5-7B-Instruct + DPOTraining=DPO2025.10 | 82.4 | 82.7 | 82.6 | |
| Qwen2.5-7B-Base + SFTTraining=SFT2025.10 | 80.9 | 67.8 | 71.8 | |
| Qwen2.5-7B-Base + RL w/ SkyworkTraining=RL w/ Skywork2025.10 | 77.1 | 73.6 | 78.7 | |
| Qwen2.5-7B-BaseTraining=Base2025.10 | 67.6 | 65.2 | 66.7 |