OneRec-Think: In-Text Reasoning for Generative Recommendation

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Liu, Zhanyu, Wang, Shiyao, Wang, Xingmei, Zhang, Rongzhou, Deng, Jiaxin, Bao, Honghui, Zhang, Jinghao, Li, Wuchao, Zheng, Pengfei, Wu, Xiangyu, Hu, Yifei, Hu, Qigen, Luo, Xinchen, Ren, Lejian, Zhang, Zixing, Wang, Qianqian, Cai, Kuo, Wu, Yunfan, Cheng, Hongtao, Cheng, Zexuan, Ren, Lu, Wang, Huanjie, Su, Yi, Tang, Ruiming, Gai, Kun, Zhou, Guorui
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866912700763734016
author Liu, Zhanyu
Wang, Shiyao
Wang, Xingmei
Zhang, Rongzhou
Deng, Jiaxin
Bao, Honghui
Zhang, Jinghao
Li, Wuchao
Zheng, Pengfei
Wu, Xiangyu
Hu, Yifei
Hu, Qigen
Luo, Xinchen
Ren, Lejian
Zhang, Zixing
Wang, Qianqian
Cai, Kuo
Wu, Yunfan
Cheng, Hongtao
Cheng, Zexuan
Ren, Lu
Wang, Huanjie
Su, Yi
Tang, Ruiming
Gai, Kun
Zhou, Guorui
author_facet Liu, Zhanyu
Wang, Shiyao
Wang, Xingmei
Zhang, Rongzhou
Deng, Jiaxin
Bao, Honghui
Zhang, Jinghao
Li, Wuchao
Zheng, Pengfei
Wu, Xiangyu
Hu, Yifei
Hu, Qigen
Luo, Xinchen
Ren, Lejian
Zhang, Zixing
Wang, Qianqian
Cai, Kuo
Wu, Yunfan
Cheng, Hongtao
Cheng, Zexuan
Ren, Lu
Wang, Huanjie
Su, Yi
Tang, Ruiming
Gai, Kun
Zhou, Guorui
contents The powerful generative capacity of Large Language Models (LLMs) has instigated a paradigm shift in recommendation. However, existing generative models (e.g., OneRec) operate as implicit predictors, critically lacking the capacity for explicit and controllable reasoning-a key advantage of LLMs. To bridge this gap, we propose OneRec-Think, a unified framework that seamlessly integrates dialogue, reasoning, and personalized recommendation. OneRec-Think incorporates: (1) Itemic Alignment: cross-modal Item-Textual Alignment for semantic grounding; (2) Reasoning Activation: Reasoning Scaffolding to activate LLM reasoning within the recommendation context; and (3) Reasoning Enhancement, where we design a recommendation-specific reward function that accounts for the multi-validity nature of user preferences. Experiments across public benchmarks show state-of-the-art performance. Moreover, our proposed "Think-Ahead" architecture enables effective industrial deployment on Kuaishou, achieving a 0.159\% gain in APP Stay Time and validating the practical efficacy of the model's explicit reasoning capability.
format Preprint
id arxiv_https___arxiv_org_abs_2510_11639
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle OneRec-Think: In-Text Reasoning for Generative Recommendation
Liu, Zhanyu
Wang, Shiyao
Wang, Xingmei
Zhang, Rongzhou
Deng, Jiaxin
Bao, Honghui
Zhang, Jinghao
Li, Wuchao
Zheng, Pengfei
Wu, Xiangyu
Hu, Yifei
Hu, Qigen
Luo, Xinchen
Ren, Lejian
Zhang, Zixing
Wang, Qianqian
Cai, Kuo
Wu, Yunfan
Cheng, Hongtao
Cheng, Zexuan
Ren, Lu
Wang, Huanjie
Su, Yi
Tang, Ruiming
Gai, Kun
Zhou, Guorui
Information Retrieval
The powerful generative capacity of Large Language Models (LLMs) has instigated a paradigm shift in recommendation. However, existing generative models (e.g., OneRec) operate as implicit predictors, critically lacking the capacity for explicit and controllable reasoning-a key advantage of LLMs. To bridge this gap, we propose OneRec-Think, a unified framework that seamlessly integrates dialogue, reasoning, and personalized recommendation. OneRec-Think incorporates: (1) Itemic Alignment: cross-modal Item-Textual Alignment for semantic grounding; (2) Reasoning Activation: Reasoning Scaffolding to activate LLM reasoning within the recommendation context; and (3) Reasoning Enhancement, where we design a recommendation-specific reward function that accounts for the multi-validity nature of user preferences. Experiments across public benchmarks show state-of-the-art performance. Moreover, our proposed "Think-Ahead" architecture enables effective industrial deployment on Kuaishou, achieving a 0.159\% gain in APP Stay Time and validating the practical efficacy of the model's explicit reasoning capability.
title OneRec-Think: In-Text Reasoning for Generative Recommendation
topic Information Retrieval
url https://arxiv.org/abs/2510.11639