A Unified Model for Personalization: Language-Steerable Generative Recommendation, Search, and User Understanding
Authors:
- Marco De NadaiMNMarco De Nadai
- Edoardo D'AMicoEDEdoardo D'AMico
- Max LefarovMLMax Lefarov
- Alexandre TamborrinoATAlexandre Tamborrino
- Divita VohraDVDivita Vohra
- Mark VanMiddlesworthMVMark VanMiddlesworth
- Shawn LinSLShawn Lin
- Jacqueline WoodJWJacqueline Wood
- Jan StypkaJSJan Stypka
- Eliza KlyceEKEliza Klyce
- Keshi DaiKDKeshi Dai
- Matthew N.K. SmithMSMatthew N.K. Smith
Open Access
- 8 August 2026
- conference paper
- Published by Association for Computing Machinery (ACM)
- p. 779-790
Abstract
Large language models (LLMs) are increasingly applied to recommendation, retrieval, and reasoning, yet deploying a single end-to-end model that can jointly support these behaviors over large, heterogeneous catalogs remains challenging. Such systems must generate unambiguous references to real items, handle multiple entity types, and operate under strict latency and reliability constraints requirements that are difficult to satisfy with text-only generation. While tool-augmented recommender systems address parts of this problem, they introduce orchestration complexity and limit end-to-end optimization. We view this setting as an instance of a broader research problem: how to adapt LLMs to reason jointly over multiple-domain entities, user behavior, and language in a fully self-contained manner. To this end, we introduce NEO, a framework that adapts a pre-trained decoder-only LLM into a tool-free, catalog-grounded generator. NEO represents items using semantic identifiers (SIDs) and trains a single model to interleave natural language and typed item identifiers within a shared sequence. Natural-language prompts control the task, target entity type, and output format (IDs, text, or mixed), while constrained decoding guarantees catalog-valid item generation without restricting free-form text. We refer to this instruction-conditioned controllability as language-steerability. Inspired by multimodal alignment, we treat SIDs as a distinct modality and study design choices for integrating discrete entity representations into LLMs via staged alignment and instruction tuning. We evaluate NEO at scale on a real-world catalog of over 10M items across multiple media types and discovery tasks, including recommendation, search, and user understanding. In offline experiments, NEO consistently outperforms strong task-specific baselines and exhibits positive cross-task transfer, demonstrating a practical path toward consolidating large-scale discovery capabilities into a single language-steerable generative model.Keywords
This publication has 13 references indexed in Scilit:
- Semantic IDs for Joint Generative Search and Recommendation2025
- Evaluating Podcast Recommendations with Profile-Aware LLM-as-a-Judge2025
- User-LLM: Efficient LLM Contextualization with User Embeddings2025
- EAGER-LLM: Enhancing Large Language Models as Recommenders through Exogenous Behavior-Semantic Integration2025
- STAR: A Simple Training-free Approach for Recommendations using Large Language ModelsarXiv, 2025
- CoST: Contrastive Quantization based Semantic Tokenization for Generative RecommendationarXiv, 2024
- Personalized Audiobook Recommendations at Spotify Through Graph Neural Networks2024
- Exploring Data Splitting Strategies for the Evaluation of Recommendation Models2020
- Robust Training of Vector Quantized Bottleneck Models2020
- Similarity estimation techniques from rounding algorithms2002