The Efficiency Gap in Byte Modeling

Modern language models have historically relied on two dominant design choices: subword tokenization and autoregressive (AR) ordering. These design decisions bake in priors that dictate a model's learning. Recently, two alternative paradigms have challenged this: byte-level modeling, which bypasses static statistically-derived token vocabularies, and masked diffusion modeling (MDM), which conducts parallel, non-sequential generation. Their intersection represents a fully end-to-end modality-agnostic generative prototype; however, removing these structural priors incurs a significant computational cost. In this work, we investigate this cost through a compute-matched scaling study. Our results reveal that the performance penalty of byte modeling is not uniform; across scale, the scaling overhead of byte modeling is worse for MDM than for AR. We hypothesize that this disparity stems from context fragility: while AR's stable causal history allows models to naturally rediscover subword patterns, the MDM objective destroys the local contiguity required to efficiently resolve semantics from raw bytes. Our findings from controlled permutation experiments suggest that future modality-agnostic designs must incorporate alternative structural biases to maintain viable scaling trajectories in the byte regime.

ORBIT: Preserving Foundational Language Capabilities in GenRetrieval via Origin-Regulated Merging

Despite the rapid advancements in large language model (LLM) development, fine-tuning them for specific tasks often results in the catastrophic forgetting of their general, language-based reasoning abilities. This work investigates and addresses this challenge in the context of the Generative Retrieval (GenRetrieval) task. During GenRetrieval fine-tuning, we find this forgetting occurs rapidly and correlates with the distance between the fine-tuned and original model parameters. Given these observations, we propose ORBIT, a novel approach that actively tracks the distance between fine-tuned and initial model weights, and uses a weight averaging strategy to constrain model drift during GenRetrieval fine-tuning when this inter-model distance exceeds a maximum threshold. Our results show that ORBIT retains substantial text and retrieval performance by outperforming both common continual learning baselines and related regularization methods that also employ weight averaging.

Vectorizing the Trie: Efficient Constrained Decoding for LLM-based Generative Retrieval on Accelerators

Generative retrieval has emerged as a powerful paradigm for LLM-based recommendation. However, industrial recommender systems often benefit from restricting the output space to a constrained subset of items based on business logic (e.g. enforcing content freshness or product category), which standard autoregressive decoding cannot natively support. Moreover, existing constrained decoding methods that make use of prefix trees (Tries) incur severe latency penalties on hardware accelerators (TPUs/GPUs). In this work, we introduce STATIC (Sparse Transition Matrix-Accelerated Trie Index for Constrained Decoding), an efficient and scalable constrained decoding technique designed specifically for high-throughput LLM-based generative retrieval on TPUs/GPUs. By flattening the prefix tree into a static Compressed Sparse Row (CSR) matrix, we transform irregular tree traversals into fully vectorized sparse matrix operations, unlocking massive efficiency gains on hardware accelerators. We deploy STATIC on a large-scale industrial video recommendation platform serving billions of users. STATIC produces significant product metric impact with minimal latency overhead (0.033 ms per step and 0.25% of inference time), achieving a 948x speedup over a CPU trie implementation and a 47-1033x speedup over a hardware-accelerated binary-search baseline. Furthermore, the runtime overhead of STATIC remains extremely low across a wide range of practical configurations. To the best of our knowledge, STATIC enables the first production-scale deployment of strictly constrained generative retrieval. In addition, evaluation on academic benchmarks demonstrates that STATIC can considerably improve cold-start performance for generative retrieval. Our code is available at https://github.com/youtube/static-constraint-decoding.

PLUM: Adapting Pre-trained Language Models for Industrial-scale Generative Recommendations

Large Language Models (LLMs) pose a new paradigm of modeling and computation for information tasks. Recommendation systems are a critical application domain poised to benefit significantly from the sequence modeling capabilities and world knowledge inherent in these large models. In this paper, we introduce PLUM, a framework designed to adapt pre-trained LLMs for industry-scale recommendation tasks. PLUM consists of item tokenization using Semantic IDs, continued pre-training (CPT) on domain-specific data, and task-specific fine-tuning for recommendation objectives. For fine-tuning, we focus particularly on generative retrieval, where the model is directly trained to generate Semantic IDs of recommended items based on user context. We conduct comprehensive experiments on large-scale internal video recommendation datasets. Our results demonstrate that PLUM achieves substantial improvements for retrieval compared to a heavily-optimized production model built with large embedding tables. We also present a scaling study for the model's retrieval performance, our learnings about CPT, a few enhancements to Semantic IDs, along with an overview of the training and inference methods that enable launching this framework to billions of users in YouTube.

Bridging the Gap: Unpacking the Hidden Challenges in Knowledge Distillation for Online Ranking Systems

Knowledge Distillation (KD) is a powerful approach for compressing a large model into a smaller, more efficient model, particularly beneficial for latency-sensitive applications like recommender systems. However, current KD research predominantly focuses on Computer Vision (CV) and NLP tasks, overlooking unique data characteristics and challenges inherent to recommender systems. This paper addresses these overlooked challenges, specifically: (1) mitigating data distribution shifts between teacher and student models, (2) efficiently identifying optimal teacher configurations within time and budgetary constraints, and (3) enabling computationally efficient and rapid sharing of teacher labels to support multiple students. We present a robust KD system developed and rigorously evaluated on multiple large-scale personalized video recommendation systems within Google. Our live experiment results demonstrate significant improvements in student model performance while ensuring consistent and reliable generation of high quality teacher labels from a continuous data stream of data.

LLMs for User Interest Exploration: A Hybrid Approach

Traditional recommendation systems are subject to a strong feedback loop by learning from and reinforcing past user-item interactions, which in turn limits the discovery of novel user interests. To address this, we introduce a hybrid hierarchical framework combining Large Language Models (LLMs) and classic recommendation models for user interest exploration. The framework controls the interfacing between the LLMs and the classic recommendation models through "interest clusters", the granularity of which can be explicitly determined by algorithm designers. It recommends the next novel interests by first representing "interest clusters" using language, and employs a fine-tuned LLM to generate novel interest descriptions that are strictly within these predefined clusters. At the low level, it grounds these generated interests to an item-level policy by restricting classic recommendation models, in this case a transformer-based sequence recommender to return items that fall within the novel clusters generated at the high level. We showcase the efficacy of this approach on an industrial-scale commercial platform serving billions of users. Live experiments show a significant increase in both exploration of novel interests and overall user enjoyment of the platform.

Large Language Models as Data Augmenters for Cold-Start Item Recommendation

The reasoning and generalization capabilities of LLMs can help us better understand user preferences and item characteristics, offering exciting prospects to enhance recommendation systems. Though effective while user-item interactions are abundant, conventional recommendation systems struggle to recommend cold-start items without historical interactions. To address this, we propose utilizing LLMs as data augmenters to bridge the knowledge gap on cold-start items during training. We employ LLMs to infer user preferences for cold-start items based on textual description of user historical behaviors and new item descriptions. The augmented training signals are then incorporated into learning the downstream recommendation models through an auxiliary pairwise loss. Through experiments on public Amazon datasets, we demonstrate that LLMs can effectively augment the training signals for cold-start items, leading to significant improvements in cold-start item recommendation for various recommendation models.

Gemini: A Family of Highly Capable Multimodal Models

This report introduces a new family of multimodal models, Gemini, that exhibit remarkable capabilities across image, audio, video, and text understanding. The Gemini family consists of Ultra, Pro, and Nano sizes, suitable for applications ranging from complex reasoning tasks to on-device memory-constrained use-cases. Evaluation on a broad range of benchmarks shows that our most-capable Gemini Ultra model advances the state of the art in 30 of 32 of these benchmarks - notably being the first model to achieve human-expert performance on the well-studied exam benchmark MMLU, and improving the state of the art in every one of the 20 multimodal benchmarks we examined. We believe that the new capabilities of Gemini models in cross-modal reasoning and language understanding will enable a wide variety of use cases and we discuss our approach toward deploying them responsibly to users.

Beyond ChatBots: ExploreLLM for Structured Thoughts and Personalized Model Responses

Large language model (LLM) powered chatbots are primarily text-based today, and impose a large interactional cognitive load, especially for exploratory or sensemaking tasks such as planning a trip or learning about a new city. Because the interaction is textual, users have little scaffolding in the way of structure, informational "scent", or ability to specify high-level preferences or goals. We introduce ExploreLLM that allows users to structure thoughts, help explore different options, navigate through the choices and recommendations, and to more easily steer models to generate more personalized responses. We conduct a user study and show that users find it helpful to use ExploreLLM for exploratory or planning tasks, because it provides a useful schema-like structure to the task, and guides users in planning. The study also suggests that users can more easily personalize responses with high-level preferences with ExploreLLM. Together, ExploreLLM points to a future where users interact with LLMs beyond the form of chatbots, and instead designed to support complex user tasks with a tighter integration between natural language and graphical user interfaces.

Better Generalization with Semantic IDs: A case study in Ranking for Recommendations

Training good representations for items is critical in recommender models. Typically, an item is assigned a unique randomly generated ID, and is commonly represented by learning an embedding corresponding to the value of the random ID. Although widely used, this approach have limitations when the number of items are large and items are power-law distributed -- typical characteristics of real-world recommendation systems. This leads to the item cold-start problem, where the model is unable to make reliable inferences for tail and previously unseen items. Removing these ID features and their learned embeddings altogether to combat cold-start issue severely degrades the recommendation quality. Content-based item embeddings are more reliable, but they are expensive to store and use, particularly for users' past item interaction sequence. In this paper, we use Semantic IDs, a compact discrete item representations learned from content embeddings using RQ-VAE that captures hierarchy of concepts in items. We showcase how we use them as a replacement of item IDs in a resource-constrained ranking model used in an industrial-scale video sharing platform. Moreover, we show how Semantic IDs improves the generalization ability of our system, without sacrificing top-level metrics.