Efficiency vs. Alignment: Investigating Safety and Fairness Risks in Parameter-Efficient Fine-Tuning of LLMs Article Swipe

MAKE IMPACT

Mina Taraghi , Yann Pequignot , Amin Nikanjam , Mohamed Amine Merzouk , Foutse Khomh · YOU? · · 2025 · Open Access · · DOI: https://doi.org/10.48550/arxiv.2511.00382

Organizations are increasingly adopting and adapting Large Language Models (LLMs) hosted on public repositories such as HuggingFace. Although these adaptations often improve performance on specialized downstream tasks, recent evidence indicates that they can also degrade a model's safety or fairness. Since different fine-tuning techniques may exert distinct effects on these critical dimensions, this study undertakes a systematic assessment of their trade-offs. Four widely used Parameter-Efficient Fine-Tuning methods, LoRA, IA3, Prompt-Tuning, and P-Tuning, are applied to four instruction-tuned model families (Meta-Llama-3-8B, Qwen2.5-7B, Mistral-7B, and Gemma-7B). In total, 235 fine-tuned variants are evaluated across eleven safety hazard categories and nine demographic fairness dimensions. The results show that adapter-based approaches (LoRA, IA3) tend to improve safety scores and are the least disruptive to fairness, retaining higher accuracy and lower bias scores. In contrast, prompt-based methods (Prompt-Tuning and P-Tuning) generally reduce safety and cause larger fairness regressions, with decreased accuracy and increased bias. Alignment shifts are strongly moderated by base model type: LLaMA remains stable, Qwen records modest gains, Gemma experiences the steepest safety decline, and Mistral, which is released without an internal moderation layer, displays the greatest variance. Improvements in safety do not necessarily translate into improvements in fairness, and no single configuration optimizes all fairness metrics simultaneously, indicating an inherent trade-off between these objectives. These findings suggest a practical guideline for safety-critical deployments: begin with a well-aligned base model, favour adapter-based PEFT, and conduct category-specific audits of both safety and fairness.

Type

preprint

Landing Page

http://arxiv.org/abs/2511.00382

PDF

https://arxiv.org/pdf/2511.00382

OA Status

green

OpenAlex ID

https://openalex.org/W4415937521

Raw OpenAlex JSON

OpenAlex ID

https://openalex.org/W4415937521

Canonical identifier for this work in OpenAlex

DOI

https://doi.org/10.48550/arxiv.2511.00382

Digital Object Identifier

Title

Efficiency vs. Alignment: Investigating Safety and Fairness Risks in Parameter-Efficient Fine-Tuning of LLMs

Work title

Type

preprint

OpenAlex work type

Publication year

2025

Year of publication

Publication date

2025-11-01

Full publication date if available

Authors

Mina Taraghi, Yann Pequignot, Amin Nikanjam, Mohamed Amine Merzouk, Foutse Khomh

List of authors in order

Landing page

https://arxiv.org/abs/2511.00382

Publisher landing page

PDF URL

https://arxiv.org/pdf/2511.00382

Direct link to full text PDF

Open access

Yes

Whether a free full text is available

OA status

green

Open access status per OpenAlex

OA URL

https://arxiv.org/pdf/2511.00382

Direct OA link when available

Cited by

0

Total citation count in OpenAlex

Full payload