Merge, Ensemble, and Cooperate! A Survey on Collaborative Strategies in the Era of Large Language Models

2024/07/08 by Jinliang Lu, Ziliang Pang, Lu, Jinliang +8 · 11 citations
Social Sciences · #Computation and Language (cs.CL) #FOS: Computer and information sciences #Wikis in Education and Collaboration

paper · pdf · doi:10.48550/arxiv.2407.06089

openalex publication_date 2024/07/08 · openalex created_date 2024/07/11 · openalex updated_date 2026/07/28

Abstract

The remarkable success of Large Language Models (LLMs) has ushered natural language processing (NLP) research into a new era. Despite their diverse capabilities, LLMs trained on different corpora exhibit varying strengths and weaknesses, leading to challenges in maximizing their overall efficiency and versatility. To address these challenges, recent studies have explored collaborative strategies for LLMs. This paper provides a comprehensive overview of this emerging research area, highlighting the motivation behind such collaborations. Specifically, we categorize collaborative strategies into three primary approaches: Merging, Ensemble, and Cooperation. Merging involves integrating multiple LLMs in the parameter space. Ensemble combines the outputs of various LLMs. Cooperation leverages different LLMs to allow full play to their diverse capabilities for specific tasks. We provide in-depth introductions to these methods from different perspectives and discuss their potential applications. Additionally, we outline future research directions, hoping this work will catalyze further studies on LLM collaborations and paving the way for advanced NLP applications.

Cited by

• VL-RouterBench: A Benchmark for Vision-Language Model Routing
• LLMBoost: Make Large Language Models Stronger with Boosting
• Knots: A Large-Scale Multi-Agent Enhanced Expert-Annotated Dataset and LLM Prompt Optimization for NOTAM Semantic Parsing
• Cache Merging as a Convergent Replicated State for Multi-Agent Latent Reasoning
• Beyond Consensus: Mitigating the Agreeableness Bias in LLM Judge Evaluations
• The Price Reversal Phenomenon: When Cheaper Reasoning Models Cost More
• LLM Chemistry Estimation for Multi-LLM Recommendation
• Beyond Majority Voting: LLM Aggregation by Leveraging Higher-Order Information
• PSO-Merging: Merging Models Based on Particle Swarm Optimization
• Cost-Aware Contrastive Routing for LLMs
• Quality-of-Service Aware LLM Routing for Edge Computing with Multiple Experts

Related

• How to Enable Effective Cooperation Between Humans and NLP Models: A Survey of Principles, Formalizations, and Beyond
• Efficient Large Language Models: A Survey
• Training-free LLM Merging for Multi-task Learning
• Model Merging in LLMs, MLLMs, and Beyond: Methods, Theories, Applications and Opportunities
• Exploring Collaboration Mechanisms for LLM Agents: A Social Psychology View