Sens-Merging: Sensitivity-Guided Parameter Balancing for Merging Large Language Models

• URL: http://arxiv.org/abs/2502.12420v2
• Date: Wed, 19 Feb 2025 12:34:46 GMT
• Title: Sens-Merging: Sensitivity-Guided Parameter Balancing for Merging Large Language Models
• Authors: Shuqi Liu, Han Wu, Bowei He, Xiongwei Han, Mingxuan Yuan, Linqi Song,
• Abstract summary: Sens-Merging is a sensitivity-guided coefficient adjustment method for model merging. We show that Sens-Merging significantly improves performance across general knowledge, mathematical reasoning, and code generation tasks. Our findings reveal important trade-offs between task-specific and cross-task scalings, providing insights for future model merging strategies.
• Score: 20.741460682103863
• License:
• Abstract: Recent advances in large language models have led to numerous task-specialized fine-tuned variants, creating a need for efficient model merging techniques that preserve specialized capabilities while avoiding costly retraining. While existing task vector-based merging methods show promise, they typically apply uniform coefficients across all parameters, overlooking varying parameter importance both within and across tasks. We present Sens-Merging, a sensitivity-guided coefficient adjustment method that enhances existing model merging techniques by operating at both task-specific and cross-task levels. Our method analyzes parameter sensitivity within individual tasks and evaluates cross-task transferability to determine optimal merging coefficients. Extensive experiments on Mistral 7B and LLaMA2-7B/13B models demonstrate that Sens-Merging significantly improves performance across general knowledge, mathematical reasoning, and code generation tasks. Notably, when combined with existing merging techniques, our method enables merged models to outperform specialized fine-tuned models, particularly in code generation tasks. Our findings reveal important trade-offs between task-specific and cross-task scalings, providing insights for future model merging strategies.

Related papers

• 1bit-Merging: Dynamic Quantized Merging for Large Language Models [20.19975755949984]
  texttt1bit-Merging is a novel framework that integrates task-specific routing with 1-bit quantized task vectors to balance performance and storage efficiency. We demonstrate that texttt1bit-Merging achieves comparable or superior performance to existing methods while significantly reducing storage requirements.
  arXiv  Detail & Related papers  (2025-02-15T09:47:50Z)
• Merging Models on the Fly Without Retraining: A Sequential Approach to Scalable Continual Model Merging [75.93960998357812]
  Deep model merging represents an emerging research direction that combines multiple fine-tuned models to harness their capabilities across different tasks and domains. Current model merging techniques focus on merging all available models simultaneously, with weight matrices-based methods being the predominant approaches. We propose a training-free projection-based continual merging method that processes models sequentially.
  arXiv  Detail & Related papers  (2025-01-16T13:17:24Z)
• Modeling Multi-Task Model Merging as Adaptive Projective Gradient Descent [74.02034188307857]
  Merging multiple expert models offers a promising approach for performing multi-task learning without accessing their original data. We find existing methods inevitably discard task-specific information that, while causing conflicts, is crucial for performance. Our approach consistently outperforms previous methods, achieving state-of-the-art results across diverse architectures and tasks in both vision and NLP domains.
  arXiv  Detail & Related papers  (2025-01-02T12:45:21Z)
• Parameter Competition Balancing for Model Merging [13.66727853299506]
  PCB-Merging is a training-free technique that adjusts the coefficients of each parameter for effective model merging. PCB-Merging achieves substantial performance enhancements across multiple modalities, domains, model sizes, number of tasks, fine-tuning forms, and large language models.
  arXiv  Detail & Related papers  (2024-10-03T11:17:58Z)
• MergeNet: Knowledge Migration across Heterogeneous Models, Tasks, and Modalities [72.05167902805405]
  We present MergeNet, which learns to bridge the gap of parameter spaces of heterogeneous models. The core mechanism of MergeNet lies in the parameter adapter, which operates by querying the source model's low-rank parameters. MergeNet is learned alongside both models, allowing our framework to dynamically transfer and adapt knowledge relevant to the current stage.
  arXiv  Detail & Related papers  (2024-04-20T08:34:39Z)
• AdaMerging: Adaptive Model Merging for Multi-Task Learning [68.75885518081357]
  This paper introduces an innovative technique called Adaptive Model Merging (AdaMerging) It aims to autonomously learn the coefficients for model merging, either in a task-wise or layer-wise manner, without relying on the original training data. Compared to the current state-of-the-art task arithmetic merging scheme, AdaMerging showcases a remarkable 11% improvement in performance.
  arXiv  Detail & Related papers  (2023-10-04T04:26:33Z)
• TIES-Merging: Resolving Interference When Merging Models [95.59265307318752]
  Transfer learning can confer significant advantages, including improved downstream performance, faster convergence, and better sample efficiency. Model merging has emerged as a solution to combine multiple task-specific models into a single model without performing additional training. Existing merging methods often ignore the interference between parameters of different models, resulting in large performance drops when merging multiple models. We propose TIES-Merging, which introduces three novel steps when merging models: resetting parameters that only changed a small amount during fine-tuning, resolving sign conflicts, and merging only the parameters that are in alignment with the final agreed-upon sign.
  arXiv  Detail & Related papers  (2023-06-02T17:31:32Z)
• A Dirichlet Process Mixture of Robust Task Models for Scalable Lifelong Reinforcement Learning [11.076005074172516]
  reinforcement learning algorithms can easily encounter catastrophic forgetting or interference when faced with lifelong streaming information. We propose a scalable lifelong RL method that dynamically expands the network capacity to accommodate new knowledge. We show that our method successfully facilitates scalable lifelong RL and outperforms relevant existing methods.
  arXiv  Detail & Related papers  (2022-05-22T09:48:41Z)
• Towards a Unified View of Parameter-Efficient Transfer Learning [108.94786930869473]
  Fine-tuning large pre-trained language models on downstream tasks has become the de-facto learning paradigm in NLP. Recent work has proposed a variety of parameter-efficient transfer learning methods that only fine-tune a small number of (extra) parameters to attain strong performance. We break down the design of state-of-the-art parameter-efficient transfer learning methods and present a unified framework that establishes connections between them.
  arXiv  Detail & Related papers  (2021-10-08T20:22:26Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.