Customizing In-context Learning for Dynamic Interest Adaption in LLM-based Recommendation

Abstract

Frequently updating Large Language Model (LLM)-based recommender systems to adapt to dynamic user interests—as done for traditional ones—is impractical due to high training costs, even with acceleration methods. This work explores the possibility of adapting the model to dynamic user interests without any model-level updates via In-context Learning (ICL), which enables adaptation through few-shot examples within input prompts. While using recent user interactions as ICL demonstrations offers a potential solution for dynamic interest adaptation, existing LLM-based recommenders face critical limitations: recommendation-specific tuning often diminishes the model’s in-context learning ability, and the original LLM’s ICL lacks task-specific optimization for recommendations. To bridge this gap, we introduce RecICL, a framework that establishes recommendation-oriented in-context learning by structuring recent user interactions and current inputs into ICL formats. RecICL achieves dual objectives: (1) preserving fundamental ICL capabilities during recommendation adaptation and (2) dynamically capturing user preference evolution through the most recent interactions. Extensive experiments across multiple benchmarks demonstrate RecICL’s superior performance, achieving better results without model updates.

@inproceedings{bao2025recicl,
  title		=	{Customizing In-context Learning for Dynamic Interest Adaption in LLM-based Recommendation},
  author	=	{Keqin Bao, Ming Yan, Yang Zhang, Jizhi Zhang, Wenjie Wang, Fuli Feng, and Xiangnan He},
  booktitle	=	{ACL 2025},
  year 		= 	{2025}
}
Copy

Keqin Bao

鲍克勤

Jizhi Zhang

张及之

Wenjie Wang

王文杰 教授

Fuli Feng

冯福利 教授

Xiangnan He

何向南 教授