Mathematical Reasoning on Olympiad (avg@16)
54.5Avg@16 ScoreScaf-GRPO
Evaluation Results
| Method | Links | |
|---|---|---|
| Scaf-GRPOBackbone=DeepSeek-R1-Distill-Qwen-1.5B2025.10 | 54.5 | |
| Vanilla GRPOBackbone=DeepSeek-R1-Distill-Qwen-1.5B2025.10 | 54 | |
| Scaf-GRPOBackbone=Qwen2.5-Math-7B2025.10 | 42.1 | |
| Oat-ZeroBackbone=Qwen2.5-Math-7B2025.10 | 41.4 | |
| Scaf-GRPOBackbone=Qwen2.5-7B2025.10 | 41.2 | |
| Vanilla GRPOBackbone=Qwen2.5-Math-7B2025.10 | 41.1 | |
| LUFFYBackbone=Qwen2.5-Math-7B2025.10 | 40.6 | |
| Vanilla GRPOBackbone=Qwen2.5-7B2025.10 | 40.4 | |
| SimpleRL-ZeroBackbone=Qwen2.5-Math-7B2025.10 | 38.2 | |
| Scaf-GRPOBackbone=Qwen2.5-Math-1.5B2025.10 | 37.4 | |
| Vanilla GRPOBackbone=Qwen2.5-Math-1.5B2025.10 | 35 | |
| Scaf-GRPOBackbone=Llama-3.2-3B-Instruct2025.10 | 21.3 | |
| Vanilla GRPOBackbone=Llama-3.2-3B-Instruct2025.10 | 19.2 |