The Past Is Not Past: Memory-Enhanced Dynamic Reward Shaping
Open paper
Benchmark
Model
Rank
Results
mathematical-reasoning-on-math-500
Qwen3-8B + MEDS
#45
Accuracy: 92.51
mathematical-reasoning-on-aime24
Qwen3-8B + MEDS
#86
Acc: 45.78
Rank counts only results with a code link.