Mathematical Reasoning on GSM8K (Exact Match, Avg, ΔScore)
92.89Exact Match AccuracyR1-p
Evaluation Results
| Method | Links | |||
|---|---|---|---|---|
| R1-pBackbone=Qwen2.5-3B-Instruct, Supervision Paradigm=20% GT2026.03 | 92.89 | 75.67 | -3.45 | |
| MemRewardBackbone=Qwen2.5-3B-Instruct, Supervision Paradigm=20% GT + GNN2026.03 | 92.89 | 77.02 | -2.1 | |
| R1-OracleBackbone=Qwen2.5-3B-Instruct, Supervision Paradigm=100% GT2026.03 | 92.89 | 79.12 | 0 | |
| MemRewardBackbone=Qwen2.5-1.5B-Instruct, Supervision Paradigm=20% GT + GNN2026.03 | 88.67 | 68.1 | -2.37 | |
| R1-OracleBackbone=Qwen2.5-1.5B-Instruct, Supervision Paradigm=100% GT2026.03 | 86.44 | 70.47 | 0 | |
| R1-pBackbone=Qwen2.5-1.5B-Instruct, Supervision Paradigm=20% GT2026.03 | 77.11 | 62.72 | -7.75 |