dorsal/arxiv
View SchemaUnleashing the Native Recommendation Potential: LLM-Based Generative Recommendation via Structured Term Identifiers
| Authors | Zhiyang Zhang, Junda She, Kuo Cai, Bo Chen, Shiyao Wang, Xinchen Luo, Qiang Luo, Ruiming Tang, Han Li, Kun Gai, Guorui Zhou |
|---|---|
| Categories | |
| ArXiv ID | 2601.06798vv1 |
| URL | https://arxiv.org/abs/2601.06798 |
| License | http://arxiv.org/licenses/nonexclusive-distrib/1.0/ |
Abstract
Leveraging the vast open-world knowledge and understanding capabilities of Large Language Models (LLMs) to develop general-purpose, semantically-aware recommender systems has emerged as a pivotal research direction in generative recommendation. However, existing methods face bottlenecks in constructing item identifiers. Text-based methods introduce LLMs' vast output space, leading to hallucination, while methods based on Semantic IDs (SIDs) encounter a semantic gap between SIDs and LLMs' native vocabulary, requiring costly vocabulary expansion and alignment training. To address this, this paper introduces Term IDs (TIDs), defined as a set of semantically rich and standardized textual keywords, to serve as robust item identifiers. We propose GRLM, a novel framework centered on TIDs, employs Context-aware Term Generation to convert item's metadata into standardized TIDs and utilizes Integrative Instruction Fine-tuning to collaboratively optimize term internalization and sequential recommendation. Additionally, Elastic Identifier Grounding is designed for robust item mapping. Extensive experiments on real-world datasets demonstrate that GRLM significantly outperforms baselines across multiple scenarios, pointing a promising direction for generalizable and high-performance generative recommendation systems.
{
"annotation_id": "2c65da2b-be03-4324-b4df-7a1d78cc0acc",
"date_created": "2026-02-17T05:53:08.728000Z",
"date_modified": "2026-02-17T05:53:08.728000Z",
"file_hash": "a88e9ebe92a21aa78d039eb51c7a86f884ae8614cac09c993137e58efbd0c168",
"private": false,
"record": {
"abstract": "Leveraging the vast open-world knowledge and understanding capabilities of Large Language Models (LLMs) to develop general-purpose, semantically-aware recommender systems has emerged as a pivotal research direction in generative recommendation. However, existing methods face bottlenecks in constructing item identifiers. Text-based methods introduce LLMs\u0027 vast output space, leading to hallucination, while methods based on Semantic IDs (SIDs) encounter a semantic gap between SIDs and LLMs\u0027 native vocabulary, requiring costly vocabulary expansion and alignment training. To address this, this paper introduces Term IDs (TIDs), defined as a set of semantically rich and standardized textual keywords, to serve as robust item identifiers. We propose GRLM, a novel framework centered on TIDs, employs Context-aware Term Generation to convert item\u0027s metadata into standardized TIDs and utilizes Integrative Instruction Fine-tuning to collaboratively optimize term internalization and sequential recommendation. Additionally, Elastic Identifier Grounding is designed for robust item mapping. Extensive experiments on real-world datasets demonstrate that GRLM significantly outperforms baselines across multiple scenarios, pointing a promising direction for generalizable and high-performance generative recommendation systems.",
"arxiv_id": "2601.06798",
"authors": [
"Zhiyang Zhang",
"Junda She",
"Kuo Cai",
"Bo Chen",
"Shiyao Wang",
"Xinchen Luo",
"Qiang Luo",
"Ruiming Tang",
"Han Li",
"Kun Gai",
"Guorui Zhou"
],
"categories": [
"cs.IR"
],
"license": "http://arxiv.org/licenses/nonexclusive-distrib/1.0/",
"title": "Unleashing the Native Recommendation Potential: LLM-Based Generative Recommendation via Structured Term Identifiers",
"url": "https://arxiv.org/abs/2601.06798",
"version": "v1"
},
"schema_id": "dorsal/arxiv",
"source": {
"execution_id": "dbcec8e5-9ff1-4a6c-a9a3-5335238681a0",
"id": "arXiv Dataset",
"type": "Model",
"variant": "snapshot-2026-01-17",
"version": "0.1.0"
},
"user_id": 1000002
}