Papers
arxiv:2604.20313

Formalising the Logit Shift Induced by LoRA: A Technical Note

Published on Apr 22
Authors:
,
,

Abstract

First-order formalization of logit shift and fact-margin change in LoRA through Fréchet approximation reveals decomposition into layerwise contributions and inter-layer coupling effects.

AI-generated summary

This technical note provides a first-order formalisation of the logit shift and fact-margin change induced by Low-Rank Adaptation (LoRA). Using a first-order Fréchet approximation around the base model trajectory, we show that the multi-layer LoRA effect can be decomposed into a linear summation of layerwise contributions and a higher-order remainder term representing inter-layer coupling.

Community

Sign up or log in to comment

Get this paper in your agent:

hf papers read 2604.20313
Don't have the latest CLI?
curl -LsSf https://hf.co/cli/install.sh | bash

Models citing this paper 0

No model linking this paper

Cite arxiv.org/abs/2604.20313 in a model README.md to link it from this page.

Datasets citing this paper 0

No dataset linking this paper

Cite arxiv.org/abs/2604.20313 in a dataset README.md to link it from this page.

Spaces citing this paper 0

No Space linking this paper

Cite arxiv.org/abs/2604.20313 in a Space README.md to link it from this page.

Collections including this paper 0

No Collection including this paper

Add this paper to a collection to link it from this page.