Token Level Routing Inference System for Edge Devices
She, Jianshu Zheng, Wenhao Liu, Zhengzhong Wang, Hongyi Xing, Eric P. Yao, Huaxiu Ho, Qirong
She, Jianshu
Zheng, Wenhao
Liu, Zhengzhong
Wang, Hongyi
Xing, Eric P.
Yao, Huaxiu
Ho, Qirong
Supervisor
Department
Machine Learning
Embargo End Date
Type
Conference proceeding
Date
2025
License
Language
English
Collections
Research Projects
Organizational Units
Journal Issue
Abstract
The computational complexity of large language model (LLM) inference significantly constrains their deployment efficiency on edge devices. In contrast, small language models offer faster decoding and lower resource consumption but often suffer from degraded response quality and heightened susceptibility to hallucinations. To address this trade-off, collaborative decoding, in which a large model assists in generating critical tokens, has emerged as a promising solution. This paradigm leverages the strengths of both model types by enabling high-quality inference through selective intervention of the large model, while maintaining the speed and efficiency of the smaller model. In this work, we present a novel collaborative decoding inference system that allows small models to perform on-device inference while selectively consulting a cloud-based large model for critical token generation. Remarkably, the system achieves a 60% performance gain on CommonsenseQA using only a 0.5B model on an M1 MacBook, with under 7% of tokens generation uploaded to the large model in the cloud.
Citation
J. She et al., “Token Level Routing Inference System for Edge Devices,” vol. 3, pp. 159–166, Aug. 2025, doi: 10.18653/V1/2025.ACL-DEMO.16
Source
Proceedings of the Annual Meeting of the Association for Computational Linguistics
Conference
63rd Annual Meeting of the Association for Computational Linguistics, ACL 2025
Keywords
Subjects
Source
63rd Annual Meeting of the Association for Computational Linguistics, ACL 2025
Publisher
Association for Computational Linguistics