1bit-Merging: Dynamic Quantized Merging for Large Language Models

Recent advances in large language models have led to specialized models excelling in specific domains, creating a need for efficient model merging techniques. While traditional merging approaches combine parameters into a single static model, they often compromise task-specific performance. However, task-specific routing methods maintain accuracy but introduce substantial storage overhead. We present \texttt{1bit}-Merging, a novel framework that integrates task-specific routing with 1-bit quantized task vectors to balance performance and storage efficiency. Our approach leverages the observation that different task-specific models store knowledge in distinct layers-chat models primarily in attention layers and math/code models in MLP layers-enabling targeted compression strategies. Through extensive experiments with LLaMA2 and Mistral model families across chat, mathematical reasoning, and code generation tasks, we demonstrate that \texttt{1bit}-Merging achieves comparable or superior performance to existing methods while significantly reducing storage requirements. Our framework offers a practical solution for combining specialized models while maintaining their individual strengths and addressing the storage challenges of current approaches.

@article{liu2025_2502.10743,
  title={ 1bit-Merging: Dynamic Quantized Merging for Large Language Models },
  author={ Shuqi Liu and Han Wu and Bowei He and Zehua Liu and Xiongwei Han and Mingxuan Yuan and Linqi Song },
  journal={arXiv preprint arXiv:2502.10743},
  year={ 2025 }
}