The paper presents a method for fine‑tuning a large language model (LLM) recommender to generate personalized, non‑harmful explanations for its recommendations. By training two LLM‑judge reward models and using constrained GRPO, the authors achieve a significant increase in the PASS rate for all three criteria, from 0.649 to 0.956 on their own judges and from 0.677 to 0.931 on an independent judge. The fine‑tuned model maintains its original recommendation performance, demonstrating that LLM‑based recommenders can be adapted to complex tasks without loss of effectiveness.
By Jiashu He, Emma Yanyang Kong, JJ Tan, David Fagnan
arXiv:2603. 21613v2 Announce Type: replace-cross Abstract: Recommender agents built on Large Language Models offer a promising paradigm for personalized recommendation.
By Tianyi Li, Zixuan Wang, Guidong Lei, Xiaodong Li, Hui Li
AgentRecommender is a method that uses large language model (LLM) agents to create customizable recommender systems on the user side, eliminating the need for additional data. By leveraging the agents' investigative capabilities and internal knowledge, users can build recommender systems tailored to their own preferences without relying on platform-driven algorithms. This approach addresses issues such as clickbait, filter bubbles, and the spread of fake news that arise from platform-centric recommender systems.
By Ryoma Sato
CRAMER is a framework that enables sequential recommendation models to adapt instantly to user requests by treating natural‑language requests as control signals and applying request‑aware masking to frozen backbone parameters. This approach avoids costly retraining or large language model inference, achieving minimal overhead. Experiments on large‑scale benchmarks demonstrate that CRAMER outperforms four state‑of‑the‑art request‑aware baselines while offering enhanced controllability and cross‑domain adaptability.
By Zhiyuan Julian Su, Naihe Feng, Zhen Luther Qin, Ga Wu
CORAL is an LLM‑native harness that automates continual optimization of production recommender systems. It operates in a closed loop: an agent observes system signals, reasons over past decisions, and uses tools—including a numerical optimizer—to reconfigure the recommender while staying within a fixed operating budget. In A/B experiments on two large social platforms, CORAL improved engagement without extra serving cost on one platform and reduced serving cost without harming engagement on the other, demonstrating that a single agentic loop can replace manual engineering for ongoing system tuning.
By Muhammad Rafay Azhar, Yuhang Zhou, Gilbert Jiang, Yuchen Wang, Rahul Sharma, Matthew DeSousa, Jiayi Liu, Xin Guo, Lizhu Zhang, Xiangjun Fan
arXiv:2607. 19739v1 Announce Type: cross Abstract: Although large language models (LLMs) have recently gained traction in recommender systems due to their strong reasoning capabilities and extensive world knowledge, previous LLM-based agents suffer from hallucination and context-length limitations, and thus are not suitable for full-ranking recommendation tasks.
By Mingdai Yang, Zhiwei Liu, Weizhi Zhang, Yibo Wang, Hao Peng, Philip Yu