# A Unified Language Model for Large Scale Search, Recommendation, and Reasoning

arXiv: https://arxiv.org/abs/2603.17533 · HTML: https://mlsanity.com/paper/2603.17533/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Placed in LLM reasoning for recommendation, one of 17 directions in LLM reasoning.
- Builds on 30 of its 75 references that are in the map.
- First posted on arXiv 18 Mar 2026.

## Placed in

- LLM reasoning / [LLM reasoning for recommendation](https://mlsanity.com/lr/llm-reasoning-for-recommendation/)

## Builds on (most cited first, 30 of 30)

- [The Llama 3 Herd of Models](https://mlsanity.com/paper/2407.21783/) (2024)
- [Qwen3 Technical Report](https://mlsanity.com/paper/2505.09388/) (2025)
- [Efficient Estimation of Word Representations in Vector Space](https://mlsanity.com/paper/1301.3781/) (2013)
- [Qwen2-VL: Enhancing Vision-Language Model's Perception of the World at Any Resolution](https://mlsanity.com/paper/2409.12191/) (2024)
- [Qwen3 Embedding: Advancing Text Embedding and Reranking Through Foundation Models](https://mlsanity.com/paper/2506.05176/) (2025)
- [Chameleon: Mixed-Modal Early-Fusion Foundation Models](https://mlsanity.com/paper/2405.09818/) (2024)
- [Emu3: Next-Token Prediction is All You Need](https://mlsanity.com/paper/2409.18869/) (2024)
- [Finite Scalar Quantization: VQ-VAE Made Simple](https://mlsanity.com/paper/2309.15505/) (2023)
- [VILA-U: a Unified Foundation Model Integrating Visual Understanding and Generation](https://mlsanity.com/paper/2409.04429/) (2024)
- [Actions Speak Louder than Words: Trillion-Parameter Sequential Transducers for Generative Recommendations](https://mlsanity.com/paper/2402.17152/) (2024)
- [Emu: Generative Pretraining in Multimodality](https://mlsanity.com/paper/2307.05222/) (2023)
- [CoLLM: Integrating Collaborative Embeddings into Large Language Models for Recommendation](https://mlsanity.com/paper/2310.19488/) (2023)
- [Leveraging Large Language Models in Conversational Recommender Systems](https://mlsanity.com/paper/2305.07961/) (2023)
- [IDGenRec: LLM-RecSys Alignment with Textual ID Learning](https://mlsanity.com/paper/2403.19021/) (2024)
- [Better Generalization with Semantic IDs: A Case Study in Ranking for Recommendations](https://mlsanity.com/paper/2306.08121/) (2023)
- [OneCAT: Decoder-Only Auto-Regressive Model for Unified Understanding and Generation](https://mlsanity.com/paper/2509.03498/) (2025)
- [Vision as LoRA](https://mlsanity.com/paper/2503.20680/) (2025)
- [PLUM: Adapting Pre-trained Language Models for Industrial-scale Generative Recommendations](https://mlsanity.com/paper/2510.07784/) (2025)
- [Retrieval-Augmented Generation: A Comprehensive Survey of Architectures, Enhancements, and Robustness Frontiers](https://mlsanity.com/paper/2506.00054/) (2025)
- [CoST: Contrastive Quantization based Semantic Tokenization for Generative Recommendation](https://mlsanity.com/paper/2404.14774/) (2024)
- [Embedding-based Zero-shot Retrieval through Query Generation](https://mlsanity.com/paper/2009.10270/) (2020)
- [MIO: A Foundation Model on Multimodal Tokens](https://mlsanity.com/paper/2409.17692/) (2024)
- [OpenP5: An Open-Source Platform for Developing, Training, and Evaluating LLM-based Recommender Systems](https://mlsanity.com/paper/2306.11134/) (2023)
- [Embedding-to-Prefix: Parameter-Efficient Personalization for Pre-Trained Large Language Models](https://mlsanity.com/paper/2505.17051/) (2025)
- [STAR: A Simple Training-free Approach for Recommendations using Large Language Models](https://mlsanity.com/paper/2410.16458/) (2024)
- [TALKPLAY: Multimodal Music Recommendation with Large Language Models](https://mlsanity.com/paper/2502.13713/) (2025)
- [Learning Multi-Aspect Item Palette: A Semantic Tokenization Framework for Generative Recommendation](https://mlsanity.com/paper/2409.07276/) (2024)
- [TalkPlay-Tools: Conversational Music Recommendation with LLM Tool Calling](https://mlsanity.com/paper/2510.01698/) (2025)
- [Realizing Scaling Laws in Recommender Systems: A Foundation-Expert Paradigm for Hyperscale Model Deployment](https://mlsanity.com/paper/2508.02929/) (2025)
- [MVIGER: Multi-View Variational Integration of Complementary Knowledge for Generative Recommender](https://mlsanity.com/paper/2408.08686/) (2024)

## Built on by (most cited first, 2 of 2)

- [TwiSTAR:Think Fast, Think Slow, Then Act,Generative Recommendation with Adaptive Reasoning](https://mlsanity.com/paper/2605.11553/) (2026)
- [The Disconnect Between Better Descriptive Reasoning Trace Quality and Recommendation Effectiveness](https://mlsanity.com/paper/2608.23154/) (2026)

## Abstract (arXiv)

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, users, 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 as SIDs and trains a single model to interleave natural language and typed item identifiers within a shared sequence. Text 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. 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 cross-task transfer, demonstrating a practical path toward consolidating large-scale discovery capabilities into a single language-steerable generative model.