Mathematical Reasoning on GSM8K (train)
89.8AccuracyJustGRPO
Evaluation Results
| Method | Links | |
|---|---|---|
| JustGRPOSeq Len=5122026.01 | 89.8 | |
| JustGRPOSeq Len=2562026.01 | 89.1 | |
| LLaDOUSeq Len=2562026.01 | 88.1 | |
| SPGSeq Len=2562026.01 | 86.1 | |
| GDPOSeq Len=5122026.01 | 84.5 | |
| SPGSeq Len=5122026.01 | 84.5 | |
| JustGRPOSeq Len=1282026.01 | 83.8 | |
| ESPOSeq Len=5122026.01 | 83.7 | |
| LLaDA-1.5Seq Len=2562026.01 | 83.3 | |
| GDPOSeq Len=2562026.01 | 82.8 | |
| d-TreeRPOSeq Len=5122026.01 | 82.6 | |
| wd1Seq Len=5122026.01 | 82.3 | |
| ESPOSeq Len=2562026.01 | 82.3 | |
| d1Seq Len=5122026.01 | 82.1 | |
| d-TreeRPOSeq Len=2562026.01 | 81.2 | |
| d1Seq Len=2562026.01 | 81.1 | |
| wd1Seq Len=2562026.01 | 80.8 | |
| ESPOSeq Len=1282026.01 | 80 | |
| SPGSeq Len=1282026.01 | 78.5 | |
| GDPOSeq Len=1282026.01 | 78.4 | |
| wd1Seq Len=1282026.01 | 77.2 | |
| d1Seq Len=1282026.01 | 73.2 |