List your company

Mathematical Reasoning on GSM8K (Exact Match, Avg, ΔScore)

92.89Exact Match AccuracyR1-p
76.478880.73948589.2606Mar 13, 2026

Evaluation results

MethodLinks
R1-pBackbone=Qwen2.5-3B-Instruct, Supervision Paradigm=20% GT2026.0392.8975.67-3.45
MemRewardBackbone=Qwen2.5-3B-Instruct, Supervision Paradigm=20% GT + GNN2026.0392.8977.02-2.1
R1-OracleBackbone=Qwen2.5-3B-Instruct, Supervision Paradigm=100% GT2026.0392.8979.120
MemRewardBackbone=Qwen2.5-1.5B-Instruct, Supervision Paradigm=20% GT + GNN2026.0388.6768.1-2.37
R1-OracleBackbone=Qwen2.5-1.5B-Instruct, Supervision Paradigm=100% GT2026.0386.4470.470
R1-pBackbone=Qwen2.5-1.5B-Instruct, Supervision Paradigm=20% GT2026.0377.1162.72-7.75