Safety-Preserving Medical LLM Fine-Tuning via Bilevel Data Selection

As general large language models (LLMs) are increasingly adapted for patient-facing medical applications, improving medical capability while controlling patient-facing risks remains a critical challenge. Large-scale medical instruction data are essential for acquiring domain knowledge through supervised fine-tuning (SFT), but unsafe responses can undermine existing safety alignment. To address this problem, we propose Scalable Bilevel Data Selection (S-BDS) for safety-preserving medical SFT. By controlling selection-weight concentration, S-BDS enables reliable sample ranking over large candidate pools and identifies training examples that improve medical capability while controlling patient-facing risks. We combine clinician-validated safe data and selected data with unscreened medical instruction data for SFT, and jointly evaluate medical capability and patient-facing safety on HealthBench and PatientSafetyBench, respectively. Experiments across Llama and Qwen models show that S-BDS achieves the highest safe rates among all compared methods, outperforming the strongest baselines by 4.64 and 3.00 percentage points, respectively, while maintaining competitive medical capability. Further experiments show that varying the safety-oriented weight during SFT controls the trade-off between medical capability and patient-facing safety.

Authors

Publication Details

Journal
arXiv (Cornell University)
Published
2026-11-16
DOI
https://doi.org/10.48550/arxiv.2611.20643
Primary Topic
Artificial Intelligence in Healthcare and Education
Type
preprint
Controls
|||
ALL TIME
JAN
FEB
MAR
APR
MAY
JUN
JUL
AUG
SEP
OCT
NOV
preprint

Safety-Preserving Medical LLM Fine-Tuning via Bilevel Data Selection

Ruoxuan Shi, Zhengxing Su, Yating Liu, Sheng Yang et al.
arXiv (Cornell University)
Artificial Intelligence in Healthcare and Education
preprint

Safety-Preserving Medical LLM Fine-Tuning via Bilevel Data Selection

Ruoxuan Shi, Zhengxing Su, Yating Liu, Sheng Yang, Xiaoyang Hou
preprint en

Abstract

As general large language models (LLMs) are increasingly adapted for patient-facing medical applications, improving medical capability while controlling patient-facing risks remains a critical challenge. Large-scale medical instruction data are essential for acquiring domain knowledge through supervised fine-tuning (SFT), but unsafe responses can undermine existing safety alignment. To address this problem, we propose Scalable Bilevel Data Selection (S-BDS) for safety-preserving medical SFT. By controlling selection-weight concentration, S-BDS enables reliable sample ranking over large candidate pools and identifies training examples that improve medical capability while controlling patient-facing risks. We combine clinician-validated safe data and selected data with unscreened medical instruction data for SFT, and jointly evaluate medical capability and patient-facing safety on HealthBench and PatientSafetyBench, respectively. Experiments across Llama and Qwen models show that S-BDS achieves the highest safe rates among all compared methods, outperforming the strongest baselines by 4.64 and 3.00 percentage points, respectively, while maintaining competitive medical capability. Further experiments show that varying the safety-oriented weight during SFT controls the trade-off between medical capability and patient-facing safety.

arXiv (Cornell University)
Artificial Intelligence in Healthcare and Education
AI Navigator

Ask Laika to Summarize, Analyze, and Connect papers live on the map.

Summarize Papers & Methodologies

Extract key findings, datasets, and comparative methods across publications.

Benchmark Rankings & Visual Analytics

Rank top research institutions, authors, funders, topics, and journals by Field-Weighted Citation Impact (FWCI) and paper volume with instant charts.

Connect Distant Disciplines

Bridge topological clusters on the map to find hidden collaborative intersections.

Laika access coming soon