Can Continual Pre-training Bridge the Performance Gap between General-purpose and Specialized Language Models in the Medical Domain?

• URL: http://arxiv.org/abs/2604.19394v1
• Date: Tue, 21 Apr 2026 12:19:48 GMT
• Title: Can Continual Pre-training Bridge the Performance Gap between General-purpose and Specialized Language Models in the Medical Domain?
• Abstract summary: We build a high-quality German medical corpus (FineMed-de) from FineWeb2.<n>This corpus is used to continually pre-train and merge three well-known LLMs.<n>A comprehensive evaluation confirms that specialization dramatically enhances $7B$ model performance on German medical benchmarks.
• Score: 1.8065388020741275
• License: http://creativecommons.org/licenses/by/4.0/
• Abstract: This paper narrows the performance gap between small, specialized models and significantly larger general-purpose models through domain adaptation via continual pre-training and merging. We address the scarcity of specialized non-English data by constructing a high-quality German medical corpus (FineMed-de) from FineWeb2. This corpus is used to continually pre-train and merge three well-known LLMs (ranging from $7B$ to $24B$ parameters), creating the DeFineMed model family. A comprehensive evaluation confirms that specialization dramatically enhances $7B$ model performance on German medical benchmarks. Furthermore, the pairwise win-rate analysis of the Qwen2.5-based models demonstrates an approximately $3.5$-fold increase in the win-rate against the much larger Mistral-Small-24B-Instruct through domain adaptation. This evidence positions specialized $7B$ models as a competitive, resource-efficient solution for complex medical instruction-following tasks. While model merging successfully restores instruction-following abilities, a subsequent failure mode analysis reveals inherent trade-offs, including the introduction of language mixing and increased verbosity, highlighting the need for more targeted fine-tuning in future work. This research provides a robust, compliant methodology for developing specialized LLMs, serving as the foundation for practical use in German-speaking healthcare contexts.

Related papers

• Is Biomedical Specialization Still Worth It? Insights from Domain-Adaptive Language Modelling with a New French Health Corpus [6.725200230250009]
  This study investigates domain-adaptive pre-training (DAPT) as a strategy for specializing small to mid-sized LLMs in the French biomedical domain.<n>Our contributions include the release of a fully open-licensed French biomedical corpus suitable for commercial and open-source applications.
  arXiv  Detail & Related papers  (2026-04-08T09:59:50Z)
• A Federated and Parameter-Efficient Framework for Large Language Model Training in Medicine [59.78991974851707]
  Large language models (LLMs) have demonstrated strong performance on medical benchmarks, including question answering and diagnosis.<n>Most medical LLMs are trained on data from a single institution, which faces limitations in generalizability and safety in heterogeneous systems.<n>We introduce the model-agnostic and parameter-efficient federated learning framework for adapting LLMs to medical applications.
  arXiv  Detail & Related papers  (2026-01-29T18:48:21Z)
• MedAlign: A Synergistic Framework of Multimodal Preference Optimization and Federated Meta-Cognitive Reasoning [52.064286116035134]
  We develop MedAlign, a framework to ensure visually accurate LVLM responses for Medical Visual Question Answering (Med-VQA)<n>We first propose a multimodal Direct Preference Optimization (mDPO) objective to align preference learning with visual context.<n>We then design a Retrieval-Aware Mixture-of-Experts (RA-MoE) architecture that utilizes image and text similarity to route queries to a specialized and context-augmented LVLM.
  arXiv  Detail & Related papers  (2025-10-24T02:11:05Z)
• mSCoRe: a $M$ultilingual and Scalable Benchmark for $S$kill-based $Co$mmonsense $Re$asoning [74.97363626515236]
  We propose a textbfMultilingual and Scalable Benchmark for textbfSkill-based textbfCommonsense textbfReasoning (textbfmSCoRe)<n>Our benchmark incorporates three key components that are designed to systematically evaluate LLM's reasoning capabilities.<n>Our results reveal the limitations of such reasoning-reinforced models when confronted with nuanced multilingual general and cultural commonsense.
  arXiv  Detail & Related papers  (2025-08-13T18:59:02Z)
• Med-R$^3$: Enhancing Medical Retrieval-Augmented Reasoning of LLMs via Progressive Reinforcement Learning [25.67288561609553]
  We introduce **Med-R$3$**, a **Med**ical **R**etrieval-augmented **R**easoning framework driven by progressive **R**einforcement learning.<n>In this framework, we first develop the model's ability to perform logical reasoning over medical problems.<n>We then adaptively optimize the retrieval capability to better align with the characteristics of knowledge corpus and external information utilization.
  arXiv  Detail & Related papers  (2025-07-31T13:31:01Z)
• Uncertainty-Driven Expert Control: Enhancing the Reliability of Medical Vision-Language Models [52.2001050216955]
  Existing methods aim to enhance the performance of Medical Vision Language Model (MedVLM) by adjusting model structure, fine-tuning with high-quality data, or through preference fine-tuning.<n>We propose an expert-in-the-loop framework named Expert-Controlled-Free Guidance (Expert-CFG) to align MedVLM with clinical expertise without additional training.
  arXiv  Detail & Related papers  (2025-07-12T09:03:30Z)
• The Unreasonable Effectiveness of Model Merging for Cross-Lingual Transfer in LLMs [45.08958917457921]
  Large language models (LLMs) still struggle across tasks outside of high-resource languages.<n>In this work, we investigate cross-lingual transfer to lower-resource languages where task-specific post-training data is scarce.
  arXiv  Detail & Related papers  (2025-05-23T20:28:31Z)
• Adapting LLMs for the Medical Domain in Portuguese: A Study on Fine-Tuning and Model Evaluation [1.922611370494431]
  This study evaluates the performance of large language models (LLMs) as medical agents in Portuguese. The InternLM2 model, with initial training on medical data, presented the best overall performance. DrBode models, derived from ChatBode, exhibited a phenomenon of catastrophic forgetting of acquired medical knowledge.
  arXiv  Detail & Related papers  (2024-09-30T19:10:03Z)
• LoRKD: Low-Rank Knowledge Decomposition for Medical Foundation Models [59.961172635689664]
  "Knowledge Decomposition" aims to improve the performance on specific medical tasks. We propose a novel framework named Low-Rank Knowledge Decomposition (LoRKD) LoRKD explicitly separates gradients from different tasks by incorporating low-rank expert modules and efficient knowledge separation convolution.
  arXiv  Detail & Related papers  (2024-09-29T03:56:21Z)
• LLMs-in-the-loop Part-1: Expert Small AI Models for Bio-Medical Text Translation [0.0]
  This study introduces a novel "LLMs-in-the-loop" approach to develop supervised neural machine translation models optimized for medical texts. Custom parallel corpora in six languages were compiled from scientific articles, synthetically generated clinical documents, and medical texts. Our MarianMT-based models outperform Google Translate, DeepL, and GPT-4-Turbo.
  arXiv  Detail & Related papers  (2024-07-16T19:32:23Z)
• Towards Building Multilingual Language Model for Medicine [54.1382395897071]
  We construct a multilingual medical corpus, containing approximately 25.5B tokens encompassing 6 main languages. We propose a multilingual medical multi-choice question-answering benchmark with rationale, termed as MMedBench. Our final model, MMed-Llama 3, with only 8B parameters, achieves superior performance compared to all other open-source models on both MMedBench and English benchmarks.
  arXiv  Detail & Related papers  (2024-02-21T17:47:20Z)
• DAEDRA: A language model for predicting outcomes in passive pharmacovigilance reporting [0.0]
  DAEDRA is a large language model designed to detect regulatory-relevant outcomes in adverse event reports. This paper details the conception, design, training and evaluation of DAEDRA.
  arXiv  Detail & Related papers  (2024-02-10T16:48:45Z)
• SANSformers: Self-Supervised Forecasting in Electronic Health Records with Attention-Free Models [48.07469930813923]
  This work aims to forecast the demand for healthcare services, by predicting the number of patient visits to healthcare facilities. We introduce SANSformer, an attention-free sequential model designed with specific inductive biases to cater for the unique characteristics of EHR data. Our results illuminate the promising potential of tailored attention-free models and self-supervised pretraining in refining healthcare utilization predictions across various patient demographics.
  arXiv  Detail & Related papers  (2021-08-31T08:23:56Z)

This list is automatically generated from the titles and abstracts of the papers in this site.

This site does not guarantee the quality of this site (including all information) and is not responsible for any consequences.