arXiv：cs.AI· Xuancheng Li, Haitao Li, Yujia Zhou, Qingyi Pan, Heng Wang, Yiqun Liu, Min Zhang, Qingyao Ai·· 13 小时前
SLIFT：从用户反馈中选择性自学习的大语言模型框架
Different Feedback, Different Updates: Selective Self-Learning from User Interactions for Large Language Models
AI 导读
研究者提出 SLIFT，一个基于任务相对视角的选择性自学习框架，将每条用户反馈拆解为原子组件并标注为 Fix、Spec 或 Null 三类。SLIFT 在共享冻结主干上训练两个互补 LoRA 适配器，Generalist 通过反馈条件自蒸馏固化 Fix 需求，Specialist 为适用但未满足的 Spec 提供残差指导，Null 不产生正向更新。
正文
Abstract:User feedback offers natural supervision for persistent LLM improvement, but a single message may support multiple behavioral changes with different scopes of generalization. We introduce SLIFT, a selective self-learning framework built on a task-relative view of user feedback. SLIFT decomposes each feedback message into atomic components and interprets each component relative to the original task as Fix, Spec, or Null: requirements for task validity, compatible condition-specific refinements, or content with no reliable positive update direction. To incorporate each change at the appropriate scope, SLIFT trains two complementary LoRA adapters on a shared frozen backbone: a Generalist that consolidates Fix requirements into default behavior through feedback-conditioned self-distillation, and a Specialist that observes only the task and Generalist response to supply residual guidance for applicable, unmet Spec refinements. Null components induce no positive update. Across backbones, SLIFT achieves strong performance on both MemoryBench and WildFB, with targeted analyses further examining its underlying mechanisms. We release our code at this https URL.
| Comments: | Corrected a typo in the author metadata |
| Subjects: | Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2608.09109 [cs.AI] |
| (or arXiv:2608.09109v2 [cs.AI] for this version) | |
| https://doi.org/10.48550/arXiv.2608.09109 arXiv-issued DOI via DataCite |
Submission history
From: Xuancheng Li [view email]
[v1]
Mon, 10 Aug 2026 04:27:58 UTC (2,208 KB)
[v2]
Fri, 9 Oct 2026 03:20:42 UTC (2,208 KB)
来源：arXiv：cs.AI · arxiv.org