When Shared Knowledge Hurts: Spectral Over-Accumulation in Model Merging

• URL: http://arxiv.org/abs/2602.05536v1
• Date: Thu, 05 Feb 2026 10:52:36 GMT
• Title: When Shared Knowledge Hurts: Spectral Over-Accumulation in Model Merging
• Abstract summary: Singular Value (SVC) is a training-free and data-free post-processing method that quantifies subspace overlap and rescales singular values to restore a balanced spectrum.<n>SVC consistently improves strong merging baselines and achieves state-of-the-art performance across vision and language benchmarks.
• Score: 39.696384608482255
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: Model merging combines multiple fine-tuned models into a single model by adding their weight updates, providing a lightweight alternative to retraining. Existing methods primarily target resolving conflicts between task updates, leaving the failure mode of over-counting shared knowledge unaddressed. We show that when tasks share aligned spectral directions (i.e., overlapping singular vectors), a simple linear combination repeatedly accumulates these directions, inflating the singular values and biasing the merged model toward shared subspaces. To mitigate this issue, we propose Singular Value Calibration (SVC), a training-free and data-free post-processing method that quantifies subspace overlap and rescales inflated singular values to restore a balanced spectrum. Across vision and language benchmarks, SVC consistently improves strong merging baselines and achieves state-of-the-art performance. Furthermore, by modifying only the singular values, SVC improves the performance of Task Arithmetic by 13.0%. Code is available at: https://github.com/lyymuwu/SVC.

Related papers

• Saliency-Aware Model Merging [47.65036144170475]
  Model merging aims to consolidate multiple task-specific models fine-tuned on different datasets into a unified architecture that performs cross-domain proficiency.<n>Current data-free model merging methods often struggle to scale as they rely on simple parameter-levels that ignore inter-layer dependencies and non-uniform distribution of expertise.<n>This work proposes SA-Merging, which is built upon connectivity-based saliency formulations from structural pruning (e.g., SynFlow) and extends them to the data-free model merging setting.
  arXiv  Detail & Related papers  (2026-05-30T04:00:44Z)
• Label-Free Cross-Task LoRA Merging with Null-Space Compression [50.63908869296697]
  We introduce Null-Space Compression (NSC) Merging, a label-free, output-agnostic method that sets merge weights from adapter geometry.<n>NSC achieves state-of-the-art performance across twenty heterogeneous vision tasks with balanced gains where prior methods overfit subsets of tasks.
  arXiv  Detail & Related papers  (2026-03-27T11:34:41Z)
• Model Merging via Multi-Teacher Knowledge Distillation [11.543771846135021]
  We introduce a novel flatness-aware PAC-Bayes generalization bound specifically for the model merging setting.<n>We frame model merging as multi-teacher knowledge distillation on scarce, unlabeled data.<n>We formally demonstrate that minimizing the student-teacher Kullback-Leibler divergence directly tightens the upper bound on the merged model's excess risk.
  arXiv  Detail & Related papers  (2025-12-24T17:10:44Z)
• Towards Reversible Model Merging For Low-rank Weights [5.100622189286672]
  Model merging aims to combine multiple fine-tuned models into a single set of weights that performs well across all source tasks.<n>We show that applying conventional merging methods to low-rank weights leads to severe performance degradation in the merged model.<n>We propose a fundamentally different approach: instead of collapsing all adapters into one set of weights, we construct a compact basis.<n>This reframes merging as generating a reconstruction-capable model space rather than producing a single merged model.
  arXiv  Detail & Related papers  (2025-10-15T23:22:38Z)
• Intrinsic Training Signals for Federated Learning Aggregation [13.540945877050525]
  Federated Learning (FL) enables collaborative model training across distributed clients while preserving data privacy.<n>This work demonstrates that effective model merging can be achieved solely through existing training signals.
  arXiv  Detail & Related papers  (2025-07-09T13:03:23Z)
• NAN: A Training-Free Solution to Coefficient Estimation in Model Merging [61.36020737229637]
  We show that the optimal merging weights should scale with the amount of task-specific information encoded in each model.<n>We propose NAN, a simple yet effective method that estimates model merging coefficients via the inverse of parameter norm.<n>NAN is training-free, plug-and-play, and applicable to a wide range of merging strategies.
  arXiv  Detail & Related papers  (2025-05-22T02:46:08Z)
• Modeling Multi-Task Model Merging as Adaptive Projective Gradient Descent [72.10987117380584]
  Merging multiple expert models offers a promising approach for performing multi-task learning without accessing their original data.<n>We find existing methods discard task-specific information that, while causing conflicts, is crucial for performance.<n>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)
• If You Can't Use Them, Recycle Them: Optimizing Merging at Scale Mitigates Performance Tradeoffs [48.95875673503714]
  We study merging "generalist" models trained on many tasks.<n>Our algorithm tunes the weight of each checkpoint in a linear combination, resulting in an optimal model.<n>Good merges tend to include almost all checkpoints with non-zero weights, indicating that even seemingly bad initial checkpoints can contribute to good final merges.
  arXiv  Detail & Related papers  (2024-12-05T13:12:51Z)
• LiNeS: Post-training Layer Scaling Prevents Forgetting and Enhances Model Merging [80.17238673443127]
  LiNeS is a post-training editing technique designed to preserve pre-trained generalization while enhancing fine-tuned task performance.<n>LiNeS demonstrates significant improvements in both single-task and multi-task settings across various benchmarks in vision and natural language processing.
  arXiv  Detail & Related papers  (2024-10-22T16:26:05Z)
• EMR-Merging: Tuning-Free High-Performance Model Merging [55.03509900949149]
  We show that Elect, Mask & Rescale-Merging (EMR-Merging) shows outstanding performance compared to existing merging methods. EMR-Merging is tuning-free, thus requiring no data availability or any additional training while showing impressive performance.
  arXiv  Detail & Related papers  (2024-05-23T05:25:45Z)
• Training-Free Pretrained Model Merging [38.16269074353077]
  We propose an innovative model merging framework, coined as merging under dual-space constraints (MuDSC) In order to enhance usability, we have also incorporated adaptations for group structure, including Multi-Head Attention and Group Normalization.
  arXiv  Detail & Related papers  (2024-03-04T06:19:27Z)
• AvgOut: A Simple Output-Probability Measure to Eliminate Dull Responses [97.50616524350123]
  We build dialogue models that are dynamically aware of what utterances or tokens are dull without any feature-engineering. The first model, MinAvgOut, directly maximizes the diversity score through the output distributions of each batch. The second model, Label Fine-Tuning (LFT), prepends to the source sequence a label continuously scaled by the diversity score to control the diversity level. The third model, RL, adopts Reinforcement Learning and treats the diversity score as a reward signal.
  arXiv  Detail & Related papers  (2020-01-15T18:32:06Z)

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.