Eliminating Out-of-Domain Recommendations in LLM-based Recommender Systems: A Unified View

TL;DR

This paper introduces RecLM, a unified framework that effectively eliminates out-of-domain recommendations in LLM-based recommender systems by integrating three grounding paradigms, achieving state-of-the-art accuracy and practical applicability.

Contribution

RecLM unifies retrieval and generation paradigms in a single architecture to eliminate out-of-domain recommendations in LLM-based recommenders.

Findings

RecLM achieves zero out-of-domain recommendations (OOD@10=0) across all variants.

Constrained generation variants RecLM-cgen and RecLM-token outperform existing baselines.

The unified framework facilitates systematic comparison and practical application of LLMs in recommendation tasks.

Abstract

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.