Mathematical Reasoning on AIME 2024 (Acc, Len, L-Acc)
63.3AccBingo-A
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| Bingo-ABackbone=DeepSeek-R1-Distill-Qwen-7B2025.06 | 63.3 | 4,670 | 41.5 | |
| Bingo-EBackbone=DeepSeek-R1-Distill-Qwen-7B2025.06 | 60 | 4,011 | 42.9 | |
| PPOBackbone=DeepSeek-R1-Distill-Qwen-7B2025.06 | 56.7 | 5,857 | 30.3 | |
| BaseBackbone=DeepSeek-R1-Distill-Qwen-7B2025.06 | 40 | 6,528 | 18 | |
| Bingo-ABackbone=DeepSeek-R1-Distill-Qwen-1.5B2025.06 | 33.3 | 2,943 | 26.7 | |
| Bingo-EBackbone=DeepSeek-R1-Distill-Qwen-1.5B2025.06 | 33.3 | 2,943 | 26.7 | |
| PPOBackbone=DeepSeek-R1-Distill-Qwen-1.5B2025.06 | 26.7 | 6,961 | 10.3 | |
| PPOBackbone=Qwen2.5-Math-7B-Instruct2025.06 | 20 | 1,260 | 15.2 | |
| Bingo-ABackbone=Qwen2.5-Math-7B-Instruct2025.06 | 20 | 892 | 16.8 | |
| BaseBackbone=DeepSeek-R1-Distill-Qwen-1.5B2025.06 | 16.7 | 7,027 | 6.3 | |
| BaseBackbone=Qwen2.5-Math-7B-Instruct2025.06 | 16.7 | 1,310 | 12.5 | |
| Bingo-EBackbone=Qwen2.5-Math-7B-Instruct2025.06 | 16.7 | 811 | 14.2 |