# Bumblebee: Interleaved Mixed-Layer Building Blocks for Large-Scale Recommendation Systems

arXiv: https://arxiv.org/abs/2607.24804 · HTML: https://mlsanity.com/paper/2607.24804/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 25 of its 32 references that are in the map.
- First posted on arXiv 6 Jul 2026.

## Builds on (most cited first, 25 of 25)

- [Attention Is All You Need](https://mlsanity.com/paper/1706.03762/) (2017)
- [Scaling Laws for Neural Language Models](https://mlsanity.com/paper/2001.08361/) (2020)
- [FlashAttention: Fast and Memory-Efficient Exact Attention with IO-Awareness](https://mlsanity.com/paper/2205.14135/) (2022)
- [Wide & Deep Learning for Recommender Systems](https://mlsanity.com/paper/1606.07792/) (2016)
- [DeepFM: A Factorization-Machine based Neural Network for CTR Prediction](https://mlsanity.com/paper/1703.04247/) (2017)
- [Deep & Cross Network for Ad Click Predictions](https://mlsanity.com/paper/1708.05123/) (2017)
- [TALLRec: An Effective and Efficient Tuning Framework to Align Large Language Model with Recommendation](https://mlsanity.com/paper/2305.00447/) (2023)
- [AutoInt: Automatic Feature Interaction Learning via Self-Attentive Neural Networks](https://mlsanity.com/paper/1810.11921/) (2018)
- [Large Language Models are Zero-Shot Rankers for Recommender Systems](https://mlsanity.com/paper/2305.08845/) (2023)
- [Actions Speak Louder than Words: Trillion-Parameter Sequential Transducers for Generative Recommendations](https://mlsanity.com/paper/2402.17152/) (2024)
- [Search-based User Interest Modeling with Lifelong Sequential Behavior Data for Click-Through Rate Prediction](https://mlsanity.com/paper/2006.05639/) (2020)
- [OneRec: Unifying Retrieve and Rank with Generative Recommender and Iterative Preference Alignment](https://mlsanity.com/paper/2502.18965/) (2025)
- [How Can Recommender Systems Benefit from Large Language Models: A Survey](https://mlsanity.com/paper/2306.05817/) (2023)
- [One Model to Serve All: Star Topology Adaptive Recommender for Multi-Domain CTR Prediction](https://mlsanity.com/paper/2101.11427/) (2021)
- [Software-Hardware Co-design for Fast and Scalable Training of Deep Learning Recommendation Models](https://mlsanity.com/paper/2104.05158/) (2021)
- [Wukong: Towards a Scaling Law for Large-Scale Recommendation](https://mlsanity.com/paper/2403.02545/) (2024)
- [PEPNet: Parameter and Embedding Personalized Network for Infusing with Personalized Prior Information](https://mlsanity.com/paper/2302.01115/) (2023)
- [LONGER: Scaling Up Long Sequence Modeling in Industrial Recommenders](https://mlsanity.com/paper/2505.04421/) (2025)
- [TWIN: TWo-stage Interest Network for Lifelong User Behavior Modeling in CTR Prediction at Kuaishou](https://mlsanity.com/paper/2302.02352/) (2023)
- [TWIN V2: Scaling Ultra-Long User Behavior Sequence Modeling for Enhanced CTR Prediction at Kuaishou](https://mlsanity.com/paper/2407.16357/) (2024)
- [Scaling Law for Recommendation Models: Towards General-purpose User Representations](https://mlsanity.com/paper/2111.11294/) (2021)
- [Sampling Is All You Need on Modeling Long-Term User Behaviors for CTR Prediction](https://mlsanity.com/paper/2205.10249/) (2022)
- [Massive Memorization with Hundreds of Trillions of Parameters for Sequential Transducer Generative Recommenders](https://mlsanity.com/paper/2510.22049/) (2025)
- [Kunlun: Establishing Scaling Laws for Massive-Scale Recommendation Systems through Unified Architecture Design](https://mlsanity.com/paper/2602.10016/) (2026)
- [Bending the Scaling Law Curve in Large-Scale Recommendation Systems](https://mlsanity.com/paper/2602.16986/) (2026)

## Abstract (arXiv)

Recommendation systems have undergone significant transformations in the past years. The transition from traditional feature interaction modules to generative next-action prediction has pushed the boundaries of personalized content. Developments have largely evolved along two separate tracks. Sequence modeling approaches on the one hand and feature interaction methods on the other. In this paper, we introduce Bumblebee, a recommendation architecture that addresses the lack of interaction between the two directions through an interleaved, stackable block design. Each block implements a micro-pipeline of layers combining sequence personalization, attention-based encoding, and feature crossing into a self-contained unit. Every block produces a joint representation of both feature modalities which is consumed by the next block in the sequence. This mechanism encourages early and repeated mixture of modalities and enriches downstream features with additional contextual information. Residual connections between blocks create cross-modal information pathways and yield additional predictive performance without adding additional parameters. Blocks can be specialized by selectively dropping components, enabling flexible trade-offs between quality and throughput. We evaluate our approach on large-scale industrial data and show consistent improvements over comparable baseline models across several classification and regression tasks. Furthermore, we conduct ablation studies to confirm that the interleaved composition itself is the primary driver of these improvements. Our results suggest that interleaving heterogeneous functional units, rather than composing deep stacks, is a promising paradigm for future-generation recommendation architectures.