Fundamental Safety-Capability Trade-offs in Fine-tuning Large Language Models

Fuente: arXiv
Saved in:
Bibliographic Details
Main Authors: Chen, Pin-Yu, Shen, Han, Das, Payel, Chen, Tianyi
Format: Preprint
Published: 2025
Subjects:
Online Access:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1866917968678486016
author Chen, Pin-Yu
Shen, Han
Das, Payel
Chen, Tianyi
author_facet Chen, Pin-Yu
Shen, Han
Das, Payel
Chen, Tianyi
contents Fine-tuning Large Language Models (LLMs) on some task-specific datasets has been a primary use of LLMs. However, it has been empirically observed that this approach to enhancing capability inevitably compromises safety, a phenomenon also known as the safety-capability trade-off in LLM fine-tuning. This paper presents a theoretical framework for understanding the interplay between safety and capability in two primary safety-aware LLM fine-tuning strategies, providing new insights into the effects of data similarity, context overlap, and alignment loss landscape. Our theoretical results characterize the fundamental limits of the safety-capability trade-off in LLM fine-tuning, which are also validated by numerical experiments.
format Preprint
id arxiv_https___arxiv_org_abs_2503_20807
institution arXiv
publishDate 2025
record_format arxiv
spellingShingle Fundamental Safety-Capability Trade-offs in Fine-tuning Large Language Models
Chen, Pin-Yu
Shen, Han
Das, Payel
Chen, Tianyi
Machine Learning
Artificial Intelligence
Computation and Language
Fine-tuning Large Language Models (LLMs) on some task-specific datasets has been a primary use of LLMs. However, it has been empirically observed that this approach to enhancing capability inevitably compromises safety, a phenomenon also known as the safety-capability trade-off in LLM fine-tuning. This paper presents a theoretical framework for understanding the interplay between safety and capability in two primary safety-aware LLM fine-tuning strategies, providing new insights into the effects of data similarity, context overlap, and alignment loss landscape. Our theoretical results characterize the fundamental limits of the safety-capability trade-off in LLM fine-tuning, which are also validated by numerical experiments.
title Fundamental Safety-Capability Trade-offs in Fine-tuning Large Language Models
topic Machine Learning
Artificial Intelligence
Computation and Language
url https://arxiv.org/abs/2503.20807