LED-Merging: Mitigating Safety-Utility Conflicts in Model Merging with Location-Election-Disjoint

• URL: http://arxiv.org/abs/2502.16770v2
• Date: Thu, 14 Aug 2025 07:15:21 GMT
• Title: LED-Merging: Mitigating Safety-Utility Conflicts in Model Merging with Location-Election-Disjoint
• Authors: Qianli Ma, Dongrui Liu, Qian Chen, Linfeng Zhang, Jing Shao,
• Abstract summary: LED-Merging resolves safety-utility conflicts and provides a lightweight, training-free paradigm for constructing reliable multi-task LLMs.<n>$textbfL$ocates task-specific neurons via gradient-based attribution.<n>$textbfE$lects critical neurons through multi-model importance fusion.<n>$textbfD$isjoints conflicting updates through parameter isolation.
• Score: 42.98847958315427
• License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
• Abstract: Fine-tuning pre-trained Large Language Models (LLMs) for specialized tasks incurs substantial computational and data costs. While model merging offers a training-free solution to integrate multiple task-specific models, existing methods suffer from safety-utility conflicts where enhanced general capabilities degrade safety safeguards. We identify two root causes: $\textbf{neuron misidentification}$ due to simplistic parameter magnitude-based selection, and $\textbf{cross-task neuron interference}$ during merging. To address these challenges, we propose $\textbf{LED-Merging}$, a three-stage framework that $\textbf{L}$ocates task-specific neurons via gradient-based attribution, dynamically $\textbf{E}$lects critical neurons through multi-model importance fusion, and $\textbf{D}$isjoints conflicting updates through parameter isolation. Extensive experiments on Llama-3-8B, Mistral-7B, and Llama2-13B demonstrate that LED-Merging effectively reduces harmful response rates, showing a 31.4\% decrease on Llama-3-8B-Instruct on HarmBench, while simultaneously preserving 95\% of utility performance, such as achieving 52.39\% accuracy on GSM8K. LED-Merging resolves safety-utility conflicts and provides a lightweight, training-free paradigm for constructing reliable multi-task LLMs. Code is available at $\href{https://github.com/MqLeet/LED-Merging}{GitHub}$.

Related papers

• Replacing Multi-Step Assembly of Data Preparation Pipelines with One-Step LLM Pipeline Generation for Table QA [16.758340727602793]
  Table Question Answering (TQA) aims to answer natural language questions over structured tables.<n>Large Language Models (LLMs) enable promising solutions to this problem, with operator-centric solutions that generate table manipulation pipelines in a multi-step manner offering state-of-the-art performance.<n>We propose Operation-R1, the first framework that trains lightweight LLMs via a novel variant of reinforcement learning with verifiable rewards to produce high-quality data-preparation pipelines for TQA in a single inference step.
  arXiv  Detail & Related papers  (2026-02-26T07:49:50Z)
• Wireless Federated Multi-Task LLM Fine-Tuning via Sparse-and-Orthogonal LoRA [61.12136997430116]
  Decentralized federated learning (DFL) based on low-rank adaptation (LoRA) enables mobile devices with multi-task datasets to collaboratively fine-tune a large language model (LLM) by exchanging locally updated parameters with a subset of neighboring devices via wireless connections for knowledge integration.<n> directly aggregating parameters fine-tuned on heterogeneous datasets induces three primary issues across the DFL life-cycle: (i) catastrophic knowledge forgetting during fine-tuning process, arising from conflicting update directions caused by data heterogeneity; (ii) textitinefficient communication and convergence during model aggregation process,
  arXiv  Detail & Related papers  (2026-02-24T02:45:32Z)
• Model Merging in the Essential Subspace [78.5390284258307]
  Model merging aims to integrate multiple task-specific fine-tuned models into a single multi-task model without additional training.<n>Despite extensive research, task interference remains a major obstacle that often undermines the performance of merged models.<n>We propose ESM (Essential Subspace Merging), a robust framework for effective model merging.
  arXiv  Detail & Related papers  (2026-02-23T00:33:38Z)
• ReliabilityBench: Evaluating LLM Agent Reliability Under Production-Like Stress Conditions [0.32928123659012326]
  Existing benchmarks for tool-using LLM agents primarily report single-run success rates and miss reliability properties required in production.<n>We introduce textbfReliabilityBench, a benchmark for evaluating agent reliability across three dimensions.<n>We evaluate two models (Gemini 2.0 Flash, GPT-4o) and two agent architectures (ReAct, Reflexion) across four domains (scheduling, travel, customer support, e-commerce) over 1,280 episodes.
  arXiv  Detail & Related papers  (2026-01-03T13:41:33Z)
• Drag-and-Drop LLMs: Zero-Shot Prompt-to-Weights [75.83625828306839]
  textbfDrag-and-Drop LLMs (textitDnD) eliminates per-task training by mapping a handful of unlabeled task prompts directly to LoRA weight updates.<n>A lightweight text encoder distills each prompt batch into condition embeddings, which are then transformed by a cascaded hyper-convolutional decoder into the full set of LoRA matrices.
  arXiv  Detail & Related papers  (2025-06-19T15:38:21Z)
• Unraveling LoRA Interference: Orthogonal Subspaces for Robust Model Merging [38.12136955174922]
  Fine-tuning large language models (LMs) for individual tasks yields strong performance but is expensive for deployment and storage.<n>Recent works explore model merging to combine multiple task-specific models into a single multi-task model without additional training.<n>Existing merging methods often fail for models fine-tuned with low-rank adaptation (LoRA), due to significant performance degradation.
  arXiv  Detail & Related papers  (2025-05-28T23:28:12Z)
• Decouple and Orthogonalize: A Data-Free Framework for LoRA Merging [18.650279202312614]
  We propose a Decoupled and Orthogonal merging approach(DO-Merging)<n>By separating parameters into magnitude and direction components, we reduce the impact of magnitude differences on the directional alignment of the merged models.<n>We validate through extensive experiments across vision, language, and multi-modal domains that our proposed DO-Merging can achieve significantly higher performance than existing merging methods at a minimal cost.
  arXiv  Detail & Related papers  (2025-05-21T16:34:37Z)
• $\ extit{Agents Under Siege}$: Breaking Pragmatic Multi-Agent LLM Systems with Optimized Prompt Attacks [45.74758377276353]
  Multi-agent Large Language Model (LLM) systems create novel adversarial risks because their behavior depends on communication between agents and decentralized reasoning.<n>In this work, we innovatively focus on attacking pragmatic systems that have constrains such as limited token bandwidth, latency between message delivery, and defense mechanisms.<n>We design a $textitpermutation-invariant adversarial attack$ that optimize prompt distribution across latency and bandwidth-constraint network topologies to bypass distributed safety mechanisms.
  arXiv  Detail & Related papers  (2025-03-31T20:43:56Z)
• Optimal Brain Iterative Merging: Mitigating Interference in LLM Merging [11.708743111945727]
  Large Language Models (LLMs) have demonstrated impressive capabilities, but their high computational costs pose challenges for customization.<n>Model merging offers a cost-effective alternative, yet existing methods suffer from interference among parameters, leading to performance degradation.<n>We propose Optimal Brain Iterative Merging, a novel method designed to mitigate both intra-model and inter-model interference.
  arXiv  Detail & Related papers  (2025-02-17T09:07:49Z)
• Reasoning-Augmented Conversation for Multi-Turn Jailbreak Attacks on Large Language Models [53.580928907886324]
  Reasoning-Augmented Conversation is a novel multi-turn jailbreak framework. It reformulates harmful queries into benign reasoning tasks. We show that RACE achieves state-of-the-art attack effectiveness in complex conversational scenarios.
  arXiv  Detail & Related papers  (2025-02-16T09:27:44Z)
• Mix Data or Merge Models? Balancing the Helpfulness, Honesty, and Harmlessness of Large Language Model via Model Merging [36.00016254809852]
  This paper systematically compares the effectiveness of model merging and data mixture methods in constructing 3H-aligned LLMs.<n>We propose a novel textbfReweighting textbfEnhanced task textbfSingular textbfMerging method, textbfRESM, through outlier weighting and sparsity-aware rank selection strategies.
  arXiv  Detail & Related papers  (2025-02-08T11:56:58Z)
• R-MTLLMF: Resilient Multi-Task Large Language Model Fusion at the Wireless Edge [78.26352952957909]
  Multi-task large language models (MTLLMs) are important for many applications at the wireless edge, where users demand specialized models to handle multiple tasks efficiently.<n>The concept of model fusion via task vectors has emerged as an efficient approach for combining fine-tuning parameters to produce an MTLLM.<n>In this paper, the problem of enabling edge users to collaboratively craft such MTLMs via tasks vectors is studied, under the assumption of worst-case adversarial attacks.
  arXiv  Detail & Related papers  (2024-11-27T10:57:06Z)
• GenBFA: An Evolutionary Optimization Approach to Bit-Flip Attacks on LLMs [3.967858172081495]
  Large Language Models (LLMs) have revolutionized natural language processing (NLP)<n>Increasing adoption in mission-critical applications raises concerns about hardware-based threats, particularly bit-flip attacks (BFAs)
  arXiv  Detail & Related papers  (2024-11-21T00:01:51Z)
• Iterative Self-Tuning LLMs for Enhanced Jailbreaking Capabilities [50.980446687774645]
  We introduce ADV-LLM, an iterative self-tuning process that crafts adversarial LLMs with enhanced jailbreak ability.<n>Our framework significantly reduces the computational cost of generating adversarial suffixes while achieving nearly 100% ASR on various open-source LLMs.<n>It exhibits strong attack transferability to closed-source models, achieving 99% ASR on GPT-3.5 and 49% ASR on GPT-4, despite being optimized solely on Llama3.
  arXiv  Detail & Related papers  (2024-10-24T06:36:12Z)
• Is Parameter Collision Hindering Continual Learning in LLMs? [50.57658782050275]
  Large Language Models (LLMs) often suffer from catastrophic forgetting when learning multiple tasks sequentially.<n>We show that building non-collision parameters is a more critical interdependence factor in addressing CL challenges.<n>We propose Non-collision Low-Rank Adaptation (N-LoRA), a simple yet effective approach leveraging low collision rates to enhance CL in LLMs.
  arXiv  Detail & Related papers  (2024-10-14T05:54:11Z)
• SMILE: Zero-Shot Sparse Mixture of Low-Rank Experts Construction From Pre-Trained Foundation Models [85.67096251281191]
  We present an innovative approach to model fusion called zero-shot Sparse MIxture of Low-rank Experts (SMILE) construction. SMILE allows for the upscaling of source models into an MoE model without extra data or further training. We conduct extensive experiments across diverse scenarios, such as image classification and text generation tasks, using full fine-tuning and LoRA fine-tuning.
  arXiv  Detail & Related papers  (2024-08-19T17:32:15Z)
• Refuse Whenever You Feel Unsafe: Improving Safety in LLMs via Decoupled Refusal Training [67.30423823744506]
  We introduce a novel approach, Decoupled Refusal Training (DeRTa), designed to empower LLMs to refuse compliance to harmful prompts at any response position.<n>DeRTa incorporates two novel components: (1) Maximum Likelihood Estimation with Harmful Response Prefix, which trains models to recognize and avoid unsafe content by appending a segment of harmful response to the beginning of a safe response, and (2) Reinforced Transition Optimization (RTO), which equips models with the ability to transition from potential harm to safety refusal consistently throughout the harmful response sequence.
  arXiv  Detail & Related papers  (2024-07-12T09:36:33Z)
• SHERL: Synthesizing High Accuracy and Efficient Memory for Resource-Limited Transfer Learning [63.93193829913252]
  We propose an innovative METL strategy called SHERL for resource-limited scenarios. In the early route, intermediate outputs are consolidated via an anti-redundancy operation. In the late route, utilizing minimal late pre-trained layers could alleviate the peak demand on memory overhead.
  arXiv  Detail & Related papers  (2024-07-10T10:22:35Z)
• Red-Teaming Large Language Models using Chain of Utterances for Safety-Alignment [32.2246459413988]
  We propose a new safety evaluation benchmark RED-EVAL that carries out red-teaming. We show that even widely deployed models are susceptible to the Chain of Utterances-based (CoU) prompting. We also demonstrate the consistency of the RED-EVAL across 8 open-source LLMs in generating harmful responses in more than 86% of the red-teaming attempts.
  arXiv  Detail & Related papers  (2023-08-18T16:27:04Z)
• 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)

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.