On-Policy Distillation with Negative-Policy Rollouts
Open paper
Benchmark
Model
Rank
Results
mathematical-reasoning-on-aime24
Qwen3-8B + NP-OPD (OPD 2, alpha=0.5, thinking mode)
#24
Acc: 81.7
mathematical-reasoning-on-math-500
Qwen3-8B + NP-OPD (OPD 2, alpha=0.5, thinking mode)
#50
Accuracy: 91.8
Rank counts only results with a code link.