Not Every Token Is Worth Distilling: Selective Supervision for Direct-OPD
Open paper
Benchmark
Model
Rank
Results
mathematical-reasoning-on-aime-2026
Qwen3-8B + S²D-OPD (Nemotron-1.5B → QuestA-1.5B)
#40
Accuracy: 70.9
Rank counts only results with a code link.