No article text was recorded for this version.
weight matrix decomposition
Version 1 of 136 papers37 conceptsfrozen 2026-10-04
Papers in this version
- Blog: Task Singular Vectors: Reducing Task Interference in Model Merging
- Blog: KnowledgeSmith: Uncovering Knowledge Updating in LLMs with Model Editing and Unlearning
- Blog: AlphaTuning: Quantization-Aware Parameter-Efficient Adaptation of Large-Scale Pre-Trained Language Models
- Blog: Sculpting Subspaces: Constrained Full Fine-Tuning in LLMs for Continual Learning
- Blog: Shortcuts in the Tail: Debiasing via Post-Hoc Spectral Compression of Fine-Tuning Updates
- Blog: Rotation-Preserving Supervised Fine-Tuning
- Blog: A new low-rank adaptation method for brain structure and metastasis segmentation via decoupled principal weight direction and magnitude
- Blog: Auditing LLMs for Hidden Behaviors using Model Diffing
- Blog: When Shared Knowledge Hurts: Spectral Over-Accumulation in Model Merging
- Blog: Spectral-Aware Low-Rank Adaptation for Speaker Verification
- Blog: RL Fine-Tuning Heals OOD Forgetting in SFT
- Blog: Critical Points of Linear Neural Networks: Analytical Forms and Landscape Properties
- Blog: Continual Learning with Scaled Gradient Projection
- Blog: PiSSA: Principal Singular Values and Singular Vectors Adaptation of Large Language Models
- Blog: Learning dynamics in linear recurrent neural networks
- Blog: Exploring Data-Free LoRA Transferability for Video Diffusion Models
- Blog: Efficient Multi-Source Knowledge Transfer by Model Merging
- Blog: Federated Learning of Large Models at the Edge via Principal Sub-Model Training
- Blog: OptMerge: Unifying Multimodal LLM Capabilities and Modalities via Model Merging
- Blog: Singular Value Fine-tuning: Few-shot Segmentation requires Few-parameters Fine-tuning
- Blog: MASS: MoErging through Adaptive Subspace Selection
- Blog: OrthAlign: Orthogonal Subspace Decomposition for Non-Interfering Multi-Objective Alignment
- Blog: Watch the Weights: Unsupervised monitoring and control of fine-tuned LLMs
- Blog: Scalable Multi-Task Low-Rank Model Adaptation
- Blog: Parameter Efficient Continual Learning with Dynamic Low- Rank Adaptation
- Blog: Improve Safety Training of Large Language Models with Safety-Critical Singular Vectors Localization
- Blog: Real-Aware Residual Model Merging for Deepfake Detection
- Blog: CorDA: Context-Oriented Decomposition Adaptation of Large Language Models for Task-Aware Parameter-Efficient Fine-tuning
- Blog: Overcoming Resource Constraints in Federated Learning: Large Models Can Be Trained with only Weak Clients
- Blog: Training Networks in Null Space of Feature Covariance for Continual Learning
- Blog: STAR: Spectral Truncation and Rescale for Model Merging
- Blog: Energy-Regularized Sequential Model Editing on Hyperspheres
- Blog: Fine-tuning with Reserved Majority for Noise Reduction
- Blog: Redefining Experts: Interpretable Decomposition of Language Models for Toxicity Mitigation
- Blog: Low Rank and Sparse Fourier Structure in Recurrent Networks Trained on Modular Addition
- Blog: MODEL SOUPS NEED ONLY ONE INGREDIENT
History & citation
Current · Permanent link
Cite this direction
Latest version (a living page):
@misc{lacuna_weight_matrix_decomposition_e8212c01,
title = {weight matrix decomposition},
author = {{Lacuna}},
howpublished = {Lacuna research direction},
year = {2026},
url = {https://lacuna.tiptreesystems.com/direction/weight-matrix-decomposition/rdr_e8212c01ec284af681e5781b06784a35},
note = {Latest version of a living research direction; cite a version for a fixed text. Accessed 2026-11-20},
}This exact version (version 1, frozen text):
@misc{lacuna_weight_matrix_decomposition_e8212c01_v1,
title = {weight matrix decomposition},
author = {{Lacuna}},
howpublished = {Lacuna research direction},
year = {2026},
url = {https://lacuna.tiptreesystems.com/direction/weight-matrix-decomposition/rdr_e8212c01ec284af681e5781b06784a35/v/1},
note = {Version 1, frozen 2026-10-04. Accessed 2026-11-20},
}