TY  - RPRT
TI  - Easy Samples Are All You Need: Self-Evolving LLMs via Data-Efficient Reinforcement Learning
AU  - Zhiyin Yu
AU  - Bo Zhang
AU  - Qibin Hou
AU  - Zhonghai Wu
AU  - Xiao Luo
AU  - Lei Bai
PY  - 2026
UR  - https://arxiv.org/abs/2604.18639
ID  - 2604.18639
ER  -