# Which Tokens Should SFT Actually Learn? A Token-Trimming Perspective on Mathematical Reasoning

> *机器翻译 (AI-generated content)*

> 标题为英语

> 全文为英语

**Published:** 2026年9月10日
**Author:** Yaning Jia, Chunhui Zhang, Wenxuan Xu, Xingjian Diao, Xiaoyuan Wang, Soroush Vosoughi
**Source:** arXiv — cs.AI
**Language:** en-US
**HTML:** https://aiqazaqstan.com/zh/articles/which-tokens-should-sft-actually-learn-a-token-trimming-perspective-on-m

arXiv:2609.09707v1 Announce Type: new Abstract: Supervised fine-tuning (SFT) applies a uniform cross-entropy loss to all target tokens, even though different tokens provide unequal learning signals for mathematical reasoning. This uniform treatment can over-sharpen already…

> 内容来自原始来源： **arXiv — cs.AI**
> [在 arXiv — cs.AI 阅读全文](https://arxiv.org/abs/2609.09707)

**标签：** #arXiv #Meta AI

---

*AI Qazaqstan — non-commercial informational portal about AI in Kazakhstan. Machine translations and summaries are AI-generated.*