Efficient Layer-wise LLM Fine-tuning for Revision Intention Prediction

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Liu, Zhexiong, Litman, Diane
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866909818125549568
author Liu, Zhexiong
Litman, Diane
author_facet Liu, Zhexiong
Litman, Diane
contents Large Language Models (LLMs) have shown extraordinary success across various text generation tasks; however, their potential for simple yet essential text classification remains underexplored, as LLM pre-training tends to emphasize generation over classification. While LLMs with instruction tuning can transform classification into a generation task, they often struggle to categorize nuanced texts. One such example is text revision, which involves nuanced edits between pairs of texts. Although simply fine-tuning LLMs for revision classification seems plausible, it requires a large amount of revision annotations, which are exceptionally expensive and scarce in the community. To address this issue, we introduce a plug-and-play layer-wise parameter-efficient fine-tuning (PEFT) framework, i.e., IR-Tuning, which fine-tunes a subset of important LLM layers that are dynamically selected based on their gradient norm distribution, while freezing those of redundant layers. Extensive experiments suggest that IR-Tuning surpasses several layer-wise PEFT baselines over diverse text revisions, while achieving fast convergence, low GPU memory consumption, and effectiveness on small revision corpora.
format Preprint
id arxiv_https___arxiv_org_abs_2510_00268
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Efficient Layer-wise LLM Fine-tuning for Revision Intention Prediction
Liu, Zhexiong
Litman, Diane
Computation and Language
Artificial Intelligence
Large Language Models (LLMs) have shown extraordinary success across various text generation tasks; however, their potential for simple yet essential text classification remains underexplored, as LLM pre-training tends to emphasize generation over classification. While LLMs with instruction tuning can transform classification into a generation task, they often struggle to categorize nuanced texts. One such example is text revision, which involves nuanced edits between pairs of texts. Although simply fine-tuning LLMs for revision classification seems plausible, it requires a large amount of revision annotations, which are exceptionally expensive and scarce in the community. To address this issue, we introduce a plug-and-play layer-wise parameter-efficient fine-tuning (PEFT) framework, i.e., IR-Tuning, which fine-tunes a subset of important LLM layers that are dynamically selected based on their gradient norm distribution, while freezing those of redundant layers. Extensive experiments suggest that IR-Tuning surpasses several layer-wise PEFT baselines over diverse text revisions, while achieving fast convergence, low GPU memory consumption, and effectiveness on small revision corpora.
title Efficient Layer-wise LLM Fine-tuning for Revision Intention Prediction
topic Computation and Language
Artificial Intelligence
url https://arxiv.org/abs/2510.00268