Denser neq Better: Limits of On-Policy Self-Distillation for Continual Post-Training
Paper • 2607.01763 • Published • 3
GRPO
GRPO checkpoints (every 10th step) and Dataset
-
Moenupa/Dolci-Think-RL-7B
Viewer • Updated • 72.2k • 32Note Dataset
Moenupa/verl
Viewer • Updated • 216k • 29Note Dataset
-
Moenupa/Qwen3-4B-Thinking-2507-GRPO-Tool
Updated -
Moenupa/Qwen3-4B-Thinking-2507-GRPO-Math
Updated -
Moenupa/Qwen3-4B-Thinking-2507-GRPO-Chem
Updated -
Moenupa/Qwen3-4B-Thinking-2507-GRPO-MathChem
Updated -
Moenupa/Qwen3-4B-Thinking-2507-GRPO-MathChemTool
Updated -
Moenupa/Qwen3-4B-Thinking-2507-GRPO-MathChemToolCode
Updated