# DP-SelFT: Differentially Private Selective Fine-Tuning for Large Language Models

arXiv: https://arxiv.org/abs/2605.17432 · HTML: https://mlsanity.com/paper/2605.17432/

## Key facts

- No citing papers in the map yet (data through 14 Oct 2026).
- Builds on 25 of its 39 references that are in the map.
- First posted on arXiv 17 May 2026.

## Builds on (most cited first, 25 of 25)

- [RoBERTa: A Robustly Optimized BERT Pretraining Approach](https://mlsanity.com/paper/1907.11692/) (2019)
- [GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understanding](https://mlsanity.com/paper/1804.07461/) (2018)
- [DistilBERT, a distilled version of BERT: smaller, faster, cheaper and lighter](https://mlsanity.com/paper/1910.01108/) (2019)
- [Parameter-Efficient Transfer Learning for NLP](https://mlsanity.com/paper/1902.00751/) (2019)
- [Deep Learning with Differential Privacy](https://mlsanity.com/paper/1607.00133/) (2016)
- [A Broad-Coverage Challenge Corpus for Sentence Understanding through Inference](https://mlsanity.com/paper/1704.05426/) (2017)
- [Compacter: Efficient Low-Rank Hypercomplex Adapter Layers](https://mlsanity.com/paper/2106.04647/) (2021)
- [Large Language Models Can Be Strong Differentially Private Learners](https://mlsanity.com/paper/2110.05679/) (2021)
- [Differentially Private Fine-tuning of Language Models](https://mlsanity.com/paper/2110.06500/) (2021)
- [R'enyi Differential Privacy of the Sampled Gaussian Mechanism](https://mlsanity.com/paper/1908.10530/) (2019)
- [Deep Learning with Label Differential Privacy](https://mlsanity.com/paper/2102.06062/) (2021)
- [Do Not Let Privacy Overbill Utility: Gradient Embedding Perturbation for Private Learning](https://mlsanity.com/paper/2102.12677/) (2021)
- [Differentially Private Synthetic Data via Foundation Model APIs 2: Text](https://mlsanity.com/paper/2403.01749/) (2024)
- [Differentially Private Language Models for Secure Data Sharing](https://mlsanity.com/paper/2210.13918/) (2022)
- [Differentially Private Synthetic Data via Foundation Model APIs 1: Images](https://mlsanity.com/paper/2305.15560/) (2023)
- [Harnessing large-language models to generate private synthetic text](https://mlsanity.com/paper/2306.01684/) (2023)
- [Efficient Private ERM for Smooth Objectives](https://mlsanity.com/paper/1703.09947/) (2017)
- [Terminal Brain Damage: Exposing the Graceless Degradation in Deep Neural Networks Under Hardware Fault Attacks](https://mlsanity.com/paper/1906.01017/) (2019)
- [Differentially Private Fine-Tuning of Diffusion Models](https://mlsanity.com/paper/2406.01355/) (2024)
- [LDPTrace: Locally Differentially Private Trajectory Synthesis](https://mlsanity.com/paper/2302.06180/) (2023)
- [TOAST: Transfer Learning via Attention Steering](https://mlsanity.com/paper/2305.15542/) (2023)
- [Continuous Release of Data Streams under both Centralized and Local Differential Privacy](https://mlsanity.com/paper/2005.11753/) (2020)
- [PrivGraph: Differentially Private Graph Data Publication by Exploiting Community Information](https://mlsanity.com/paper/2304.02401/) (2023)
- [DPMLBench: Holistic Evaluation of Differentially Private Machine Learning](https://mlsanity.com/paper/2305.05900/) (2023)
- [AHEAD: Adaptive Hierarchical Decomposition for Range Query under Local Differential Privacy](https://mlsanity.com/paper/2110.07505/) (2021)

## Abstract (arXiv)

Large language models (LLMs) are commonly adapted to downstream tasks through fine-tuning, but fine-tuning data often contains sensitive information that may be leaked by the resulting model. Differential privacy (DP) offers formal protection against such leakage, yet DP fine-tuning of LLMs still suffers from substantial utility degradation due to gradient clipping and noise injection. Existing work improves this trade-off by combining DP with parameter-efficient fine-tuning methods such as LoRA, which constrain the form of updates. In this work, we study a complementary direction: selective fine-tuning, which constrains where updates are applied. We propose DP-SelFT, a framework for differentially private selective fine-tuning of LLMs. DP-SelFT addresses three DP-specific challenges in parameter selection: avoiding repeated privacy cost, improving stability under noisy estimates, and selecting parameters that remain useful under clipped and noisy updates. It first constructs a lightweight DP synthetic dataset and performs selection only on this synthetic data, so the selection stage incurs no additional privacy cost. It then conducts layer-level selection by temporarily training candidate layer subsets on a synthetic training split and evaluating them on a synthetic validation split. Crucially, this temporary training is performed under a perturbation regime matched to downstream DP fine-tuning, with worst-case perturbations of the same scale as DP noise. This favors layer subsets that are not only learnable but also robust to noisy private updates. Experiments on benchmark tasks show that DP-SelFT consistently improves the privacy--utility trade-off over existing DP fine-tuning baselines under the same privacy guarantees.