LLM2Rec: Large Language Models Are Powerful Embedding Models for Sequential Recommendation
Fuente:
arXiv
Saved in:
| Main Authors: | , , , , |
|---|---|
| Format: | Preprint |
| Published: |
2025
|
| Subjects: | |
| Online Access: | |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| _version_ | 1866915361183498240 |
|---|---|
| author | He, Yingzhi Liu, Xiaohao Zhang, An Ma, Yunshan Chua, Tat-Seng |
| author_facet | He, Yingzhi Liu, Xiaohao Zhang, An Ma, Yunshan Chua, Tat-Seng |
| contents | Sequential recommendation aims to predict users' future interactions by modeling collaborative filtering (CF) signals from historical behaviors of similar users or items. Traditional sequential recommenders predominantly rely on ID-based embeddings, which capture CF signals through high-order co-occurrence patterns. However, these embeddings depend solely on past interactions, lacking transferable knowledge to generalize to unseen domains. Recent advances in large language models (LLMs) have motivated text-based recommendation approaches that derive item representations from textual descriptions. While these methods enhance generalization, they fail to encode CF signals-i.e., latent item correlations and preference patterns-crucial for effective recommendation. We argue that an ideal embedding model should seamlessly integrate CF signals with rich semantic representations to improve both in-domain and out-of-domain recommendation performance.
To this end, we propose LLM2Rec, a novel embedding model tailored for sequential recommendation, integrating the rich semantic understanding of LLMs with CF awareness. Our approach follows a two-stage training framework: (1) Collaborative Supervised Fine-tuning, which adapts LLMs to infer item relationships based on historical interactions, and (2) Item-level Embedding Modeling, which refines these specialized LLMs into structured item embedding models that encode both semantic and collaborative information. Extensive experiments on real-world datasets demonstrate that LLM2Rec effectively improves recommendation quality across both in-domain and out-of-domain settings. Our findings highlight the potential of leveraging LLMs to build more robust, generalizable embedding models for sequential recommendation. Our codes are available at https://github.com/HappyPointer/LLM2Rec. |
| format | Preprint |
| id |
arxiv_https___arxiv_org_abs_2506_21579 |
| institution | arXiv |
| publishDate | 2025 |
| record_format | arxiv |
| spellingShingle | LLM2Rec: Large Language Models Are Powerful Embedding Models for Sequential Recommendation He, Yingzhi Liu, Xiaohao Zhang, An Ma, Yunshan Chua, Tat-Seng Information Retrieval Artificial Intelligence Sequential recommendation aims to predict users' future interactions by modeling collaborative filtering (CF) signals from historical behaviors of similar users or items. Traditional sequential recommenders predominantly rely on ID-based embeddings, which capture CF signals through high-order co-occurrence patterns. However, these embeddings depend solely on past interactions, lacking transferable knowledge to generalize to unseen domains. Recent advances in large language models (LLMs) have motivated text-based recommendation approaches that derive item representations from textual descriptions. While these methods enhance generalization, they fail to encode CF signals-i.e., latent item correlations and preference patterns-crucial for effective recommendation. We argue that an ideal embedding model should seamlessly integrate CF signals with rich semantic representations to improve both in-domain and out-of-domain recommendation performance. To this end, we propose LLM2Rec, a novel embedding model tailored for sequential recommendation, integrating the rich semantic understanding of LLMs with CF awareness. Our approach follows a two-stage training framework: (1) Collaborative Supervised Fine-tuning, which adapts LLMs to infer item relationships based on historical interactions, and (2) Item-level Embedding Modeling, which refines these specialized LLMs into structured item embedding models that encode both semantic and collaborative information. Extensive experiments on real-world datasets demonstrate that LLM2Rec effectively improves recommendation quality across both in-domain and out-of-domain settings. Our findings highlight the potential of leveraging LLMs to build more robust, generalizable embedding models for sequential recommendation. Our codes are available at https://github.com/HappyPointer/LLM2Rec. |
| title | LLM2Rec: Large Language Models Are Powerful Embedding Models for Sequential Recommendation |
| topic | Information Retrieval Artificial Intelligence |
| url | https://arxiv.org/abs/2506.21579 |