Research

Paper

AI LLM February 26, 2026

Fine-grained Semantics Integration for Large Language Model-based Recommendation

Authors

Jiawen Feng, Xiaoyu Kong, Leheng Sheng, Bin Wu, Chao Yi, Feifang Yang, Xiang-Rong Sheng, Han Zhu, Xiang Wang, Jiancan Wu, Xiangnan He

Abstract

Recent advances in Large Language Models (LLMs) have shifted in recommendation systems from the discriminative paradigm to the LLM-based generative paradigm, where the recommender autoregressively generates sequences of semantic identifiers (SIDs) for target items conditioned on historical interaction. While prevalent LLM-based recommenders have demonstrated performance gains by aligning pretrained LLMs between the language space and the SID space, modeling the SID space still faces two fundamental challenges: (1) Semantically Meaningless Initialization: SID tokens are randomly initialized, severing the semantic linkage between the SID space and the pretrained language space at start point, and (2) Coarse-grained Alignment: existing SFT-based alignment tasks primarily focus on item-level optimization, while overlooking the semantics of individual tokens within SID sequences.To address these challenges, we propose TS-Rec, which can integrate Token-level Semantics into LLM-based Recommenders. Specifically, TS-Rec comprises two key components: (1) Semantic-Aware embedding Initialization (SA-Init), which initializes SID token embeddings by applying mean pooling to the pretrained embeddings of keywords extracted by a teacher model; and (2) Token-level Semantic Alignment (TS-Align), which aligns individual tokens within the SID sequence with the shared semantics of the corresponding item clusters. Extensive experiments on two real-world benchmarks demonstrate that TS-Rec consistently outperforms traditional and generative baselines across all standard metrics. The results demonstrate that integrating fine-grained semantic information significantly enhances the performance of LLM-based generative recommenders.

Metadata

arXiv ID: 2602.22632
Provider: ARXIV
Primary Category: cs.IR
Published: 2026-02-26
Fetched: 2026-02-28 05:45

Related papers

Raw Data (Debug)
{
  "raw_xml": "<entry>\n    <id>http://arxiv.org/abs/2602.22632v1</id>\n    <title>Fine-grained Semantics Integration for Large Language Model-based Recommendation</title>\n    <updated>2026-02-26T05:17:24Z</updated>\n    <link href='https://arxiv.org/abs/2602.22632v1' rel='alternate' type='text/html'/>\n    <link href='https://arxiv.org/pdf/2602.22632v1' rel='related' title='pdf' type='application/pdf'/>\n    <summary>Recent advances in Large Language Models (LLMs) have shifted in recommendation systems from the discriminative paradigm to the LLM-based generative paradigm, where the recommender autoregressively generates sequences of semantic identifiers (SIDs) for target items conditioned on historical interaction. While prevalent LLM-based recommenders have demonstrated performance gains by aligning pretrained LLMs between the language space and the SID space, modeling the SID space still faces two fundamental challenges: (1) Semantically Meaningless Initialization: SID tokens are randomly initialized, severing the semantic linkage between the SID space and the pretrained language space at start point, and (2) Coarse-grained Alignment: existing SFT-based alignment tasks primarily focus on item-level optimization, while overlooking the semantics of individual tokens within SID sequences.To address these challenges, we propose TS-Rec, which can integrate Token-level Semantics into LLM-based Recommenders. Specifically, TS-Rec comprises two key components: (1) Semantic-Aware embedding Initialization (SA-Init), which initializes SID token embeddings by applying mean pooling to the pretrained embeddings of keywords extracted by a teacher model; and (2) Token-level Semantic Alignment (TS-Align), which aligns individual tokens within the SID sequence with the shared semantics of the corresponding item clusters. Extensive experiments on two real-world benchmarks demonstrate that TS-Rec consistently outperforms traditional and generative baselines across all standard metrics. The results demonstrate that integrating fine-grained semantic information significantly enhances the performance of LLM-based generative recommenders.</summary>\n    <category scheme='http://arxiv.org/schemas/atom' term='cs.IR'/>\n    <published>2026-02-26T05:17:24Z</published>\n    <arxiv:primary_category term='cs.IR'/>\n    <author>\n      <name>Jiawen Feng</name>\n    </author>\n    <author>\n      <name>Xiaoyu Kong</name>\n    </author>\n    <author>\n      <name>Leheng Sheng</name>\n    </author>\n    <author>\n      <name>Bin Wu</name>\n    </author>\n    <author>\n      <name>Chao Yi</name>\n    </author>\n    <author>\n      <name>Feifang Yang</name>\n    </author>\n    <author>\n      <name>Xiang-Rong Sheng</name>\n    </author>\n    <author>\n      <name>Han Zhu</name>\n    </author>\n    <author>\n      <name>Xiang Wang</name>\n    </author>\n    <author>\n      <name>Jiancan Wu</name>\n    </author>\n    <author>\n      <name>Xiangnan He</name>\n    </author>\n  </entry>"
}