arXiv：cs.CL· Li Wang, Chen Zhang, Yirong Yang, Xingyu Chen, Jinkun Zhao, Tianhao Peng, Wenjun Wu·· 10 小时前
SkillGrain：从用户反馈中学习可迁移的个性化技能
Learning What to Keep: Transfer-Aware Personal Skill Learning from User Feedback
AI 导读
针对用户反馈中持久偏好与单次修正混杂导致负迁移的问题，研究者提出 SkillGrain 框架，将反馈拆解为原子规则并预测其可迁移性，只保留适合未来请求的规则。该方法结合 rubric 引导的技能奖励与细粒度信用分配，并通过强化学习联合优化技能归纳与技能条件生成。在涵盖五类个性化写作任务的基准上，SkillGrain 在三种基座模型的域内与域外设置中均取得最优综合平均表现。
正文
Abstract:User feedback enables LLMs to refine reusable personalized skills, but often mixes persistent preferences with request-specific corrections. Indiscriminately retaining both risks negative transfer, motivating selective retention at the level of individual requirements. We frame this challenge as selective transfer over atomic rules and propose SkillGrain, a framework for learning reusable personalized skills from user feedback. SkillGrain decomposes feedback into independently actionable rules, predicts their transferability, and retains only those suitable for future requests. To learn these decisions, we combine rubric-guided skill rewards with fine-grained credit assignment that separately evaluates rule content and retention decisions. We further jointly optimize skill induction and skill-conditioned generation through reinforcement learning, ensuring that learned skills support downstream personalization. We construct a benchmark spanning five personalized writing tasks to evaluate cross-request generalization. Experiments across three base models in both in-domain and out-of-domain settings show that SkillGrain achieves the best overall average performance among competing methods.
| Subjects: | Computation and Language (cs.CL) |
| Cite as: | arXiv:2610.19380 [cs.CL] |
| (or arXiv:2610.19380v1 [cs.CL] for this version) | |
| https://doi.org/10.48550/arXiv.2610.19380 arXiv-issued DOI via DataCite (pending registration) |
Submission history
From: Li Wang [view email]
[v1]
Wed, 14 Oct 2026 01:27:19 UTC (822 KB)
来源：arXiv：cs.CL · arxiv.org