paper-with-me

홈 › Papers

Token-Level LLM Collaboration via FusionRoute

2026-01-08 · Nuoya Xiong, Yuhang Zhou, Hanqing Zeng, Zhaorun Chen, Furong Huang, Shuchao Bi, Lizhu Zhang, Zhuokai Zhao arxiv

Large language models (LLMs) exhibit strengths across diverse domains. However, achieving strong performance across these domains with a single general-purpose model typically requires scaling to sizes that are prohibitively expensive to train and deploy. On the other hand, while smaller domain-specialized models are much more efficient, they struggle to generalize beyond their training distributions. To address this dilemma, we propose FusionRoute, a robust and effective token-level multi-LLM collaboration framework in which a lightweight router simultaneously (i) selects the most suitable expert at each decoding step and (ii) contributes a complementary logit that refines or corrects the selected expert's next-token distribution via logit addition. Unlike existing token-level collaboration methods that rely solely on fixed expert outputs, we provide a theoretical analysis showing that pure expert-only routing is fundamentally limited: unless strong global coverage assumptions hold, it cannot in general realize the optimal decoding policy. By augmenting expert selection with a trainable complementary generator, FusionRoute expands the effective policy class and enables recovery of optimal value functions under mild conditions. Empirically, across both Llama-3 and Gemma-2 families and diverse benchmarks spanning mathematical reasoning, code generation, and instruction following, FusionRoute outperforms both sequence- and token-level collaboration, model merging, and direct fine-tuning, while remaining competitive with domain experts on their respective tasks.

📄 PDF Abstract BibTeX arXiv:2601.05106

Code (0)

등록된 구현이 없습니다.

Tasks

Mathematical ReasoningInstruction FollowingCode Generation

Similar Papers 제목 키워드 기반

Dynamic Collaboration of Multi-Language Models based on Minimal Complete Semantic Units

2025-08-26 · Chao Hao, Zezheng Wang, Yanhua Huang, Ruiwen Xu 외 arxiv

This paper investigates the enhancement of reasoning capabilities in language models through token-level multi-model collaboration. Our approach selects the optimal tokens from the next token distributions provided by mu…

Beyond the Target: From Imitation to Collaboration in Speculative Decoding

2026-05-24 · Jinze Li, Yixing Xu, Guanchen Li, Jinfeng Xu 외 arxiv

Speculative decoding (SPD) accelerates large language model (LLM) inference by letting a smaller draft model propose multiple future tokens that are verified in parallel by a larger target model. The dominant SPD paradig…

Reinforcement Learning

Learning to Decode Collaboratively with Multiple Language Models

2024-03-06 · Shannon Zejiang Shen, Hunter Lang, Bailin Wang, Yoon Kim 외

We propose a method to teach multiple large language models (LLM) to collaborate by interleaving their generations at the token level. We model the decision of which LLM generates the next token as a latent variable. By …

Instruction Following

The Consensus Trap: Rescuing Multi-Agent LLMs from Adversarial Majorities via Token-Level Collaboration

2026-04-18 · Jiayuan Liu, Shiyi Du, Weihua Du, Mingyu Guo 외 arxiv

Multi-agent large language model (LLM) architectures increasingly rely on response-level aggregation, such as Majority Voting (MAJ), to raise reasoning ceilings. However, in open environments, agents are highly susceptib…

Optimizing Diversity and Quality through Base-Aligned Model Collaboration

2025-11-07 · Yichen Wang, Chenghao Yang, Tenghao Huang, Muhao Chen 외 arxiv

Alignment has greatly improved large language models (LLMs)' output quality at the cost of diversity, yielding highly similar outputs across generations, especially in open-ended generation tasks. We propose Base-Aligned…