Recommender systems based on Large Language Models (LLMs) are often plagued by hallucinations of out-of-domain (OOD) items. To address this, we propose RecLM, a unified framework that bridges the gap between retrieval and generation by instantiating three grounding paradigms under a single architecture: embedding-based retrieval, constrained generation over rewritten item titles, and discrete item-tokenizer generation. Using the same backbone LLM and prompts, we systematically compare these three views on public benchmarks. RecLM strictly eradicates OOD recommendations (OOD@10 = 0) across all variants, and the constrained generation variants RecLM-cgen and RecLM-token achieve overall state-of-the-art accuracy compared to both strong ID-based and LLM-based baselines. Our unified view provides a systematic basis for comparing three distinct paradigms to reduce item hallucinations, offering a practical framework to facilitate the application of LLMs to recommendation tasks. Source code is at https://github.com/microsoft/RecAI.
@article{arxiv.2505.03336,
title = {Eliminating Out-of-Domain Recommendations in LLM-based Recommender Systems: A Unified View},
author = {Hao Liao and Jiwei Zhang and Jianxun Lian and Wensheng Lu and Mingqi Wu and Shuo Wang and Yong Zhang and Yitian Huang and Mingyang Zhou and Rui Mao},
journal= {arXiv preprint arXiv:2505.03336},
year = {2026}
}