arXiv cs.AI· Jaewon Cho, Minseo Yoon, Jinyoung Park, Jinwoo Seo, Jaewon Chu, Alice Oh, Hyunwoo J. Kim·· 4 小时前AI 评分38
TOPSD：用可教学前缀解决多语言推理蒸馏中的 Teacher Drift 问题
Knowing When to Teach: Learning from Teachable Prefixes in Multilingual Reasoning Distillation
AI 导读
一项 arXiv 研究提出 Teachable On-Policy Self-Distillation（TOPSD），解决多语言推理自蒸馏中教师模型跟随学生错误步骤的 "Teacher Drift" 现象，且该现象在低资源语言中出现更早。
正文 · 原文
Abstract:On-Policy Self-Distillation (OPSD) improves reasoning by training a student on its own reasoning, using the same model conditioned on reference solutions as a teacher. Applying OPSD to multilingual reasoning requires the teacher to provide useful feedback across languages, even when the student makes mistakes in intermediate steps. We observe a phenomenon we call Teacher Drift: as these mistakes accumulate, the teacher may follow the student's incorrect steps instead of correcting them. Our analysis shows that this shift tends to occur earlier in lower-resource languages. To account for these differences across languages, we propose Teachable On-Policy Self-Distillation (TOPSD), which determines where and whether teacher supervision should be applied to each reasoning trajectory. TOPSD uses teacher-student likelihood to estimate where drift begins and select a prefix for supervision. It then verifies whether the teacher can recover the correct answer from that prefix, applying distillation only to prefixes that pass verification. Experiments on multilingual mathematical reasoning across ten languages and three model scales show that TOPSD consistently outperforms conventional OPSD, particularly in low- and medium-resource languages.
| Subjects: | Computation and Language (cs.CL); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2611.08179 [cs.CL] |
| (or arXiv:2611.08179v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2611.08179 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Jaewon Cho [view email]
[v1]
Thu, 5 Nov 2026 15:34:23 UTC (1,658 KB)
来源：arXiv cs.AI · arxiv.org