SOURCE-LINKED INTELLIGENCE

Learning from Hard Prompts: Difficulty-aware Advantage Amplification in Dynamic Sampling

arXiv · AI, language, vision and robotics · article · Aug 28, 2026 · UTC

Decoupled Clip and Dynamic Sampling Policy Optimization (DAPO) is a prominent variant of Group Relative Policy Optimization (GRPO). DAPO introduces several improvements over GRPO. Among these, Dynamic Sampling contributes the most to DAPO's accuracy gains relative to GRPO. To improve accuracy, Dynamic Sampling enhances training stability by eliminating zero policy gradients from zero advantages. Specifically, it avoids such zero gradients by filtering out prompts where sampled responses are either entirely correct or incorrect. However, our theoretical analysis shows that Dynamic Sampling decr

Read original source ↗ Open in workspace

recordType

paper

region

Global