mlsanityData to 7 Oct 2026

TIES-Merging: Resolving Interference When Merging Models

arXiv 2306.01708 ↗First posted 2 Jun 2023cs.LG cs.AI cs.CLcited by 795

Not on the map. Why?

Cited by 43 in LLM safety and security, but placed by its own abstract, which didn’t fit.

Abstract (arXiv)

Transfer learning - i.e., further fine-tuning a pre-trained model on a downstream task - can confer significant advantages, including improved downstream performance, faster convergence, and better sample efficiency. These advantages have led to a proliferation of task-specific fine-tuned models, which typically can only perform a single task and do not benefit from one another. Recently, model merging techniques have emerged as a solution to combine multiple task-specific models into a single multitask model without performing additional training. However, existing merging methods often ignore the interference between parameters of different models, resulting in large performance drops when merging multiple models. In this paper, we demonstrate that prior merging techniques inadvertently lose valuable information due to two major sources of interference: (a) interference due to redundant parameter values and (b) disagreement on the sign of a given parameter's values across models. To address this, we propose our method, TRIM, ELECT SIGN & MERGE (TIES-Merging), which introduces three novel steps when merging models: (1) resetting parameters that only changed a small amount during fine-tuning, (2) resolving sign conflicts, and (3) merging only the parameters that are in alignment with the final agreed-upon sign. We find that TIES-Merging outperforms several existing methods in diverse settings covering a range of modalities, domains, number of tasks, model sizes, architectures, and fine-tuning settings. We further analyze the impact of different types of interference on model parameters, and highlight the importance of resolving sign interference. Our code is available at https://github.com/prateeky2806/ties-merging

Key facts. Cited by 795 papers in the map (data through 7 Oct 2026). Builds on 68 of its 89 references that are in the map. First posted on arXiv 2 Jun 2023.

Builds on

68 of 89 references.

References · 68

  1. 2017Attention Is All You Needcited by 59,932
  2. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,204
  3. 2021Learning Transferable Visual Models From Natural Language Supervisioncited by 33,456
  4. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 29,915
  5. 2019Exploring the Limits of Transfer Learning with a Unified Text-to-Text Transformercited by 14,491
  6. 2021LoRA: Low-Rank Adaptation of Large Language Modelscited by 14,084
  7. 2016Communication-Efficient Learning of Deep Networks from Decentralized Datacited by 7,592
  8. 2018GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understandingcited by 4,854
  9. 2016Overcoming catastrophic forgetting in neural networkscited by 4,718
  10. 2019HellaSwag: Can a Machine Really Finish Your Sentence?cited by 4,018
  11. 2021On the Opportunities and Risks of Foundation Modelscited by 3,961
  12. 2019HuggingFace's Transformers: State-of-the-art Natural Language Processingcited by 2,201
38 more
  1. 2013Describing Textures in the Wildcited by 2,110
  2. 2017EuroSAT: A Novel Dataset and Deep Learning Benchmark for Land Use and Land Cover Classificationcited by 1,488
  3. 2021Multitask Prompted Training Enables Zero-Shot Task Generalizationcited by 1,462
  4. 2022Model soups: averaging weights of multiple fine-tuned models improves accuracy without increasing inference timecited by 1,366
  5. 2019On the Convergence of FedAvg on Non-IID Datacited by 1,220
  6. 2018Averaging Weights Leads to Wider Optima and Better Generalizationcited by 1,194
  7. 2019A Comprehensive Survey on Transfer Learningcited by 1,154
  8. 2022Editing Models with Task Arithmeticcited by 1,076
  9. 2016An overview of gradient descent optimization algorithmscited by 1,021
  10. 2022Few-Shot Parameter-Efficient Fine-Tuning is Better and Cheaper than In-Context Learningcited by 886
  11. 2017A Survey on Multi-Task Learningcited by 838
  12. 2019Adversarial NLI: A New Benchmark for Natural Language Understandingcited by 798
  13. 2017Remote Sensing Image Scene Classification: Benchmark and State of the Artcited by 791
  14. 2021Robust fine-tuning of zero-shot modelscited by 785
  15. 2019Linear Mode Connectivity and the Lottery Ticket Hypothesiscited by 627
  16. 2021Merging Models with Fisher-Weighted Averagingcited by 613
  17. 2018Loss Surfaces, Mode Connectivity, and Fast Ensembling of DNNscited by 585
  18. 2020Federated Learning with Matched Averagingcited by 564
  19. 2021Sparsity in Deep Learning: Pruning and growth for efficient inference and training in neural networkscited by 521
  20. 2021SWAD: Domain Generalization by Seeking Flat Minimacited by 431
  21. 2019PAWS: Paraphrase Adversaries from Word Scramblingcited by 395
  22. 2022Dataless Knowledge Fusion by Merging Weights of Language Modelscited by 384
  23. 2018WiC: the Word-in-Context Dataset for Evaluating Context-Sensitive Meaning Representationscited by 370
  24. 2020What is being transferred in transfer learning?cited by 365
  25. 2018Essentially No Barriers in Neural Network Energy Landscapecited by 328
  26. 2022UL2: Unifying Language Learning Paradigmscited by 292
  27. 2022PromptSource: An Integrated Development Environment and Repository for Natural Language Promptscited by 288
  28. 2021The Role of Permutation Invariance in Linear Mode Connectivity of Neural Networkscited by 285
  29. 2019Cosmos QA: Machine Reading Comprehension with Contextual Commonsense Reasoningcited by 281
  30. 2019QASC: A Dataset for Question Answering via Sentence Compositioncited by 248
  31. 2021Efficiently Identifying Task Groupings for Multi-Task Learningcited by 218
  32. 2023Task Arithmetic in the Tangent Space: Improved Editing of Pre-Trained Modelscited by 218
  33. 2019Model Fusion via Optimal Transportcited by 217
  34. 2015Convergent Learning: Do different neural networks learn the same representations?cited by 215
  35. 2018Sentence Encoders on STILTs: Supplementary Training on Intermediate Labeled-data Taskscited by 214
  36. 2022Branch-Train-Merge: Embarrassingly Parallel Training of Expert Language Modelscited by 187
  37. 2022Patching open-vocabulary models by interpolating weightscited by 187
  38. 2022Diverse Weight Averaging for Out-of-Distribution Generalizationcited by 148

50 most cited of 68 All 68 (JSON)

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2023: 6; 2024: 168; 2025: 277; 2026: 344.

795 papers.

Sort:

By area:

  1. 2023LoraHub: Efficient Cross-Task Generalization via Dynamic LoRA Compositioncited by 359
  2. 2024Model Merging in LLMs, MLLMs, and Beyond: Methods, Theories, Applications and Opportunitiescited by 247
  3. 2025World Simulation with Video Foundation Models for Physical AIcited by 183
  4. 2025Light-R1: Curriculum SFT, DPO and RL for Long COT from Scratch and Beyondcited by 175
  5. 2024Task Singular Vectors: Reducing Task Interference in Model Mergingcited by 143
  6. 2023Model Breadcrumbs: Scaling Multi-Task Model Merging with Sparse Maskscited by 130
  7. 2024Localizing Task Information for Improved Model Merging and Compressioncited by 130
  8. 2024Simple and Scalable Strategies to Continually Pre-train Large Language Modelscited by 122
  9. 2024MiniCache: KV Cache Compression in Depth Dimension for Large Language Modelscited by 122
  10. 2024Aya Expanse: Combining Research Breakthroughs for a New Multilingual Frontiercited by 121
  11. 2024Model merging with SVD to tie the Knotscited by 118
  12. 2024Twin-Merging: Dynamic Integration of Modular Expertise in Model Mergingcited by 114
38 more
  1. 2024EMR-Merging: Tuning-Free High-Performance Model Mergingcited by 108
  2. 2025No Task Left Behind: Isotropic Model Merging with Common and Task-Specific Subspacescited by 90
  3. 2023Deep Model Fusion: A Surveycited by 89
  4. 2024Model Stock: All we need is just a few fine-tuned modelscited by 87
  5. 2024A Survey on Self-Evolution of Large Language Modelscited by 86
  6. 2024Towards Modular LLMs by Building and Reusing a Library of LoRAscited by 72
  7. 2024Parameter Competition Balancing for Model Mergingcited by 71
  8. 2024MagMax: Leveraging Model Merging for Seamless Continual Learningcited by 67
  9. 2024Learning to Route Among Specialized Experts for Zero-Shot Generalizationcited by 65
  10. 2026MOPD: Multi-Teacher On-Policy Distillation for Capability Integration in LLM Post-Trainingcited by 60
  11. 2025Whoever Started the Interference Should End It: Guiding Data-Free Model Merging via Task Vectorscited by 59
  12. 2024Merging LoRAs like Playing LEGO: Pushing the Modularity of LoRA to Extremes Through Rank-Wise Clusteringcited by 58
  13. 2025Unlocking Efficient Long-to-Short LLM Reasoning with Model Mergingcited by 58
  14. 2024MetaGPT: Merging Large Language Models Using Model Exclusive Task Arithmeticcited by 53
  15. 2024FuseChat: Knowledge Fusion of Chat Modelscited by 53
  16. 2024A Survey on Model MoErging: Recycling and Routing Among Specialized Experts for Collaborative Learningcited by 52
  17. 2024What Matters for Model Merging at Scale?cited by 52
  18. 2024Knowledge Mechanisms in Large Language Models: A Survey and Perspectivecited by 50
  19. 2024Localize-and-Stitch: Efficient Model Merging via Sparse Task Arithmeticcited by 50
  20. 2024LoRA Soups: Merging LoRAs for Practical Skill Composition Taskscited by 49
  21. 2025Reinforcement Learning Finetunes Small Subnetworks in Large Language Modelscited by 46
  22. 2024MedExpQA: Multilingual Benchmarking of Large Language Models for Medical Question Answeringcited by 43
  23. 2024Merge, Ensemble, and Cooperate! A Survey on Collaborative Strategies in the Era of Large Language Modelscited by 43
  24. 2024WARP: On the Benefits of Weight Averaged Rewarded Policiescited by 42
  25. 2024Model Merging and Safety Alignment: One Bad Model Spoils the Bunchcited by 41
  26. 2024A safety realignment framework via subspace-oriented model fusion for large language modelscited by 40
  27. 2025Bring Reason to Vision: Understanding Perception and Reasoning through Model Mergingcited by 40
  28. 2024LiNeS: Post-training Layer Scaling Prevents Forgetting and Enhances Model Mergingcited by 39
  29. 2025Modeling Multi-Task Model Merging as Adaptive Projective Gradient Descentcited by 39
  30. 2025mmBERT: A Modern Multilingual Encoder with Annealed Language Learningcited by 39
  31. 2026UI-Venus-1.5 Technical Reportcited by 39
  32. 2025When is Task Vector Provably Effective for Model Editing? A Generalization Analysis of Nonlinear Transformerscited by 38
  33. 2024Knowledge Composition using Task Vectors with Learned Anisotropic Scalingcited by 37
  34. 2025The Avengers: A Simple Recipe for Uniting Smaller Language Models to Challenge Proprietary Giantscited by 37
  35. 2023Concrete Subspace Learning based Interference Elimination for Multi-task Model Fusioncited by 36
  36. 2024Mitigating Catastrophic Forgetting in Language Transfer via Model Mergingcited by 36
  37. 2025Model Merging in Pre-training of Large Language Modelscited by 36
  38. 2025Don't Overthink It: A Survey of Efficient R1-style Large Reasoning Modelscited by 36

50 most cited of 795 All 795 (JSON)