TY  - RPRT
TI  - When Long Helps Short: How Context Length in Supervised Fine-tuning Affects Behavior of Large Language Models
AU  - Yingming Zheng
AU  - Hanqi Li
AU  - Kai Yu
AU  - Lu Chen
PY  - 2025
UR  - https://arxiv.org/abs/2509.18762
ID  - 2509.18762
ER  -