Mathematical Reasoning on AIME 2024 (pass@3)
48.75Pass@3HDPO
Evaluation results
| Method | Links | |
|---|---|---|
| HDPOBackbone=Qwen2.5-Math-7B, Number of samples=162026.06 | 48.75 | |
| Critique-GRPOBackbone=8B, Number of samples=162026.06 | 44.93 | |
| HDPOBackbone=Qwen3-4B, Number of samples=162026.06 | 43.56 | |
| GRPOBackbone=Qwen2.5-Math-7B, Number of samples=162026.06 | 40.76 | |
| OatBackbone=Qwen2.5-Math-7B, Number of samples=162026.06 | 38.91 | |
| HDPOBackbone=DS-R1-Llama-8B, Number of samples=162026.06 | 37.76 | |
| SimpleRLBackbone=Qwen-2.5-7B, Number of samples=162026.06 | 36.43 | |
| GRPOBackbone=Qwen3-4B, Number of samples=162026.06 | 32.08 | |
| Base ModelBackbone=Qwen2.5-Math-7B, Number of samples=162026.06 | 31.55 | |
| LUFFYBackbone=Qwen-Math-7B, Number of samples=162026.06 | 28.95 | |
| GRPOBackbone=DS-R1-Llama-8B, Number of samples=162026.06 | 24.55 | |
| Base ModelBackbone=Qwen3-4B, Number of samples=162026.06 | 20.36 | |
| Base ModelBackbone=DS-R1-Llama-8B, Number of samples=162026.06 | 16.43 | |
| EDGE-GRPOBackbone=Qwen-7B, Number of samples=162026.06 | 16.36 |