MixLM: High-Throughput and Effective LLM Ranking via Text-Embedding Mix-Interaction

Fuente: arXiv
Gespeichert in:
Bibliographische Detailangaben
Hauptverfasser: Li, Guoyao, He, Ran, Jing, Shusen, Behdin, Kayhan, Wang, Yubo, Ramachandran, Sundara Raman, Nguyen, Chanh, Sheng, Jian, Ma, Xiaojing, Zhu, Chuanrui, Vasudevan, Sriram, Wu, Muchen, Ghosh, Sayan, Su, Lin, Song, Qingquan, Wang, Xiaoqing, Wang, Zhipeng, Lan, Qing, Chen, Yanning, Wu, Jingwei, Simon, Luke, Zhang, Wenjing, Guo, Qi, Borisyuk, Fedor
Format: Preprint
Veröffentlicht: 2025
Schlagworte:
Online-Zugang:
Tags: Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
_version_ 1866917238015000576
author Li, Guoyao
He, Ran
Jing, Shusen
Behdin, Kayhan
Wang, Yubo
Ramachandran, Sundara Raman
Nguyen, Chanh
Sheng, Jian
Ma, Xiaojing
Zhu, Chuanrui
Vasudevan, Sriram
Wu, Muchen
Ghosh, Sayan
Su, Lin
Song, Qingquan
Wang, Xiaoqing
Wang, Zhipeng
Lan, Qing
Chen, Yanning
Wu, Jingwei
Simon, Luke
Zhang, Wenjing
Guo, Qi
Borisyuk, Fedor
author_facet Li, Guoyao
He, Ran
Jing, Shusen
Behdin, Kayhan
Wang, Yubo
Ramachandran, Sundara Raman
Nguyen, Chanh
Sheng, Jian
Ma, Xiaojing
Zhu, Chuanrui
Vasudevan, Sriram
Wu, Muchen
Ghosh, Sayan
Su, Lin
Song, Qingquan
Wang, Xiaoqing
Wang, Zhipeng
Lan, Qing
Chen, Yanning
Wu, Jingwei
Simon, Luke
Zhang, Wenjing
Guo, Qi
Borisyuk, Fedor
contents Large language models (LLMs) excel at capturing semantic nuances and therefore show impressive relevance ranking performance in modern recommendation and search systems. However, they suffer from high computational overhead under industrial latency and throughput requirements. In particular, cross-encoder ranking systems often create long context prefill-heavy workloads, as the model has to be presented with the user, query and item information. To this end, we propose MixLM, a novel LLM-based ranking framework, which significantly improves the system throughput via reducing the input context length, while preserving the semantic strength of cross-encoder rankers. In contrast to a standard ranking system where the context is presented to the model as pure text, we propose to use mix-interaction, a mixture of text and embedding tokens to represent the input. Specifically, MixLM encodes all items in the catalog into a few embedding tokens and stores in a nearline cache. The encoded item descriptions are used during online inference, effectively reducing the item length from a few thousand text tokens to a few embedding tokens. We share insights from deploying our MixLM framework to a real-world search application at LinkedIn, including a detailed discussion of our training pipelines, as well as a thorough analysis of our online serving infrastructure optimization. With the same latency budget and on-par relevance metrics, MixLM increased throughput by 10.0x comparing with strong baselines, 75.9x over full-text LLM rerankers. The efficiency gains delivered by MixLM enabled full-traffic deployment of LLM-powered search, which resulted in a significant 0.47\% increase in Daily Active Users (DAU) in online A/B tests.
format Preprint
id arxiv_https___arxiv_org_abs_2512_07846
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle MixLM: High-Throughput and Effective LLM Ranking via Text-Embedding Mix-Interaction
Li, Guoyao
He, Ran
Jing, Shusen
Behdin, Kayhan
Wang, Yubo
Ramachandran, Sundara Raman
Nguyen, Chanh
Sheng, Jian
Ma, Xiaojing
Zhu, Chuanrui
Vasudevan, Sriram
Wu, Muchen
Ghosh, Sayan
Su, Lin
Song, Qingquan
Wang, Xiaoqing
Wang, Zhipeng
Lan, Qing
Chen, Yanning
Wu, Jingwei
Simon, Luke
Zhang, Wenjing
Guo, Qi
Borisyuk, Fedor
Information Retrieval
Artificial Intelligence
Computation and Language
Machine Learning
Large language models (LLMs) excel at capturing semantic nuances and therefore show impressive relevance ranking performance in modern recommendation and search systems. However, they suffer from high computational overhead under industrial latency and throughput requirements. In particular, cross-encoder ranking systems often create long context prefill-heavy workloads, as the model has to be presented with the user, query and item information. To this end, we propose MixLM, a novel LLM-based ranking framework, which significantly improves the system throughput via reducing the input context length, while preserving the semantic strength of cross-encoder rankers. In contrast to a standard ranking system where the context is presented to the model as pure text, we propose to use mix-interaction, a mixture of text and embedding tokens to represent the input. Specifically, MixLM encodes all items in the catalog into a few embedding tokens and stores in a nearline cache. The encoded item descriptions are used during online inference, effectively reducing the item length from a few thousand text tokens to a few embedding tokens. We share insights from deploying our MixLM framework to a real-world search application at LinkedIn, including a detailed discussion of our training pipelines, as well as a thorough analysis of our online serving infrastructure optimization. With the same latency budget and on-par relevance metrics, MixLM increased throughput by 10.0x comparing with strong baselines, 75.9x over full-text LLM rerankers. The efficiency gains delivered by MixLM enabled full-traffic deployment of LLM-powered search, which resulted in a significant 0.47\% increase in Daily Active Users (DAU) in online A/B tests.
title MixLM: High-Throughput and Effective LLM Ranking via Text-Embedding Mix-Interaction
topic Information Retrieval
Artificial Intelligence
Computation and Language
Machine Learning
url https://arxiv.org/abs/2512.07846