mlsanityData to 14 Oct 2026

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

arXiv 2508.18763 ↗First posted 26 Aug 2025cs.AIcited by 2

On the map: LLM reasoning → Test-time compute and efficient thinking

Abstract (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 multiple models to perform autoregressive reasoning. Contrary to the assumption that more models yield better results, we introduce a distribution distance-based dynamic selection strategy (DDS) to optimize the multi-model collaboration process. To address the critical challenge of vocabulary misalignment in multi-model collaboration, we propose the concept of minimal complete semantic units (MCSU), which is simple yet enables multiple language models to achieve natural alignment within the linguistic space. Experimental results across various benchmarks demonstrate the superiority of our method. The code will be available at https://github.com/Fanye12/DDS.

Key facts. Cited by 2 papers in the map (data through 14 Oct 2026). Placed in Test-time compute and efficient thinking, one of 17 directions in LLM reasoning. Builds on 39 of its 47 references that are in the map. First posted on arXiv 26 Aug 2025.

Builds on

39 of 47 references.

Closest ancestors · 1

Same direction, newest first

  1. 2023Self-Refine: Iterative Refinement with Self-Feedback30 Mar

Other references · 38

  1. 2023GPT-4 Technical Reportcited by 18,396
  2. 2024The Llama 3 Herd of Modelscited by 14,715
  3. 2023LLaMA: Open and Efficient Foundation Language Modelscited by 14,132
  4. 2015Distilling the Knowledge in a Neural Networkcited by 10,324
  5. 2021Training Verifiers to Solve Math Word Problemscited by 10,109
  6. 2021Evaluating Large Language Models Trained on Codecited by 8,320
  7. 2020Scaling Laws for Neural Language Modelscited by 6,586
  8. 2018Think you have Solved Question Answering? Try ARC, the AI2 Reasoning Challengecited by 4,886
  9. 2022Large Language Models are Zero-Shot Reasonerscited by 4,374
  10. 2021TruthfulQA: Measuring How Models Mimic Human Falsehoodscited by 2,975
  11. 2018Can a Suit of Armor Conduct Electricity? A New Dataset for Open Book Question Answeringcited by 2,164
  12. 2024Qwen2 Technical Reportcited by 2,120
26 more
  1. 2022Beyond the Imitation Game: Quantifying and extrapolating the capabilities of language modelscited by 1,998
  2. 2018CommonsenseQA: A Question Answering Challenge Targeting Commonsense Knowledgecited by 1,822
  3. 2023Improving Factuality and Reasoning in Language Models through Multiagent Debatecited by 1,784
  4. 2018Hierarchical Neural Story Generationcited by 1,193
  5. 2024ChatGLM: A Family of Large Language Models from GLM-130B to GLM-4 All Toolscited by 1,121
  6. 2023Encouraging Divergent Thinking in Large Language Models through Multi-Agent Debatecited by 1,011
  7. 2021Are NLP Models really able to Solve Simple Math Word Problems?cited by 878
  8. 2021Did Aristotle Use a Laptop? A Question Answering Benchmark with Implicit Reasoning Strategiescited by 743
  9. 2024Yi: Open Foundation Models by 01.AIcited by 719
  10. 2023C-Eval: A Multi-Level Multi-Discipline Chinese Evaluation Suite for Foundation Modelscited by 708
  11. 2017Program Induction by Rationale Generation : Learning to Solve and Explain Algebraic Word Problemscited by 650
  12. 2023LLM-Blender: Ensembling Large Language Models with Pairwise Ranking and Generative Fusioncited by 564
  13. 2022Complexity-Based Prompting for Multi-Step Reasoningcited by 388
  14. 2024Debating with More Persuasive LLMs Leads to More Truthful Answerscited by 266
  15. 2023Large Language Model Routing with Benchmark Datasetscited by 172
  16. 2023Exchange-of-Thought: Enhancing Large Language Model Capabilities through Cross-Model Communicationcited by 82
  17. 2023Corex: Pushing the Boundaries of Complex Reasoning through Multi-Model Collaborationcited by 70
  18. 2024Ensemble Learning for Heterogeneous Large Language Models with Deep Parallel Collaborationcited by 67
  19. 2024Learning to Decode Collaboratively with Multiple Language Modelscited by 62
  20. 2024Bridging the Gap between Different Vocabularies for LLM Ensemblecited by 45
  21. 2024Breaking the Ceiling of the LLM Community by Treating Token Generation as a Classification for Ensemblingcited by 33
  22. 2025PersonaX: A Recommendation Agent Oriented User Modeling Framework for Long Behavior Sequencecited by 33
  23. 2025ReSo: A Reward-driven Self-organizing LLM-based Multi-Agent System for Reasoning Taskscited by 31
  24. 2024Determine-Then-Ensemble: Necessity of Top-k Union for Large Language Model Ensemblingcited by 28
  25. 2024Enhancing Retrieval and Managing Retrieval: A Four-Module Synergy for Improved Quality and Efficiency in RAG Systemscited by 15
  26. 2025Uncertainty-Aware GUI Agent: Adaptive Perception through Component Recommendation and Human-in-the-Loop Refinementcited by 15

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2025: 1; 2026: 1.

2 papers.

  1. 2025Entropy-Aware Token Rejection for Improving Speculative Decodingcited by 3
  2. 2026GUI-C2: Coarse-to-Fine GUI Grounding via Difficulty-Aware Reinforcement Learningcited by 2

Similar, not cited

Close in wording, no citation either way

  1. 2026Soft Token Alignment for Cross-Lingual Reasoning
  2. 2026MentorCollab: Large-to-Small Inference-Time Mentorship for Concise Reasoning in Language Models
  3. 2026Local-First Intelligence via Proactive SLM Collaborating with Feedback-Providing LLM
  4. 2026PyroDash: Cost-Efficient Token-Level Small-Large Language Model Collaborative Inference
  5. 2025Synergistic Weak-Strong Collaboration by Aligning Preferences
  6. 2026The Single-Multi Evolution Loop for Self-Improving Model Collaboration Systems
  7. 2025Probabilistic Token Alignment for Large Language Model Fusion
  8. 2024Unconstrained Model Merging for Enhanced LLM Reasoning