Entity · paper

TailLoR: Protecting Principal Components in Parameter-Efficient Continual Learning

paperactivetaillor-protecting-principal-components-in-parameter-efficient-continual-learning-1dbc2fe8·1 events·first seen Jun 5, 2026

Aliases: TailLoR: Protecting Principal Components in Parameter-Efficient Continual Learning

Co-occurring entities

TailLoR

More like this (12)

Sparse Subspace-to-Expert Sharing for Task-Agnostic Continual Learning The Stable Recovery Manifold: Geometric Principles Governing Recoverability in Continual Learning Continual LLM Upcycling: A Predictor-Gated Bank-Wise Sparsity Training Recipe for Dense-to-Sparse LLMs PC Layer: Polynomial Weight Preconditioning for Improving LLM Pre-Training Towards Continual Motion-Language Agents: LoRA Variants for Incremental Motion Understanding and Generation Range Penalization: Theoretical Insights with Applications in Federated Learning QVal: Cheaply Evaluating Dense Supervision Signals for Long-Horizon LLM Agents QVal: Cheaply Evaluating Dense Supervision Signals for Long-Horizon LLM Agents FlashbackCL: Mitigating Temporal Forgetting in Federated Learning KL-regularized RL Linear Independent Component Analysis via Optimal Transport Learning from the Self-future: On-policy Self-distillation for dLLMs

Recent events (1)

4arXiv · cs.LG·Jun 5, 2026·source ↗

TailLoR: Spectral-domain continual learning via protected principal components

TailLoR is a new parameter-efficient finetuning method for continual learning that uses the singular value decomposition of pre-trained weights as a fixed reference frame, applying low-rank updates only to the singular value matrix. A soft spectral penalty discourages updates aligned with dominant singular directions, reducing catastrophic interference while routing adaptation into long-tail spectral coordinates. The approach targets the forgetting problem in continual learning through a principled spectral lens.

Open Weights Progress TailLoR: Protecting Principal Components in Parameter-Efficient Continual Learning TailLoR