Domain-Adaptive Pretraining

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.

Recent text embedding models are often adapted to specialized domains via contrastive pre-finetuning (PFT) on a naive collection of scattered, heterogeneous tasks. However, this approach often introduces task-induced bias alongside domain knowledge, leading to uncontrolled representation shifts that distort the pretrained embedding geometry and cause substantial performance degradation. To address this issue, we propose REZE, a representation regularization framework that explicitly controls representation shift during embedding pre-finetuning. REZE operates on the relations of anchor-positive pairs and decomposes them in an eigenspace. It then measures task-wise dispersion along each eigencomponent to identify task-variant directions and applies adaptive soft-shrinkage to suppress task-induced noise while preserving task-invariant semantic structure, without inference-time overhead. Experiments across multiple embedding backbones and specialized benchmarks show that REZE outperforms standard pre-finetuning and isotropy-oriented post-hoc regularization in most settings, remaining stable where existing PFT variants collapse. Embedding space analyses further confirm that REZE induces controlled shifts aligned with the original embedding manifold, underscoring representation shift control as a key principle for robust embedding pre-finetuning under heterogeneous supervision.

In Natural Language Processing (NLP) classification tasks where a lack of labeled data is an issue, continued pretraining (CPT) of transformer models on unlabeled data is an established approach. In this paper, we have two aims. (1) We describe our observations from continued pretraining of the Finnish BERT transformer model (FinBERT) on a Finnish histopathological dataset (below, the Histopathology data). (2) Since the Histopathology data has no classification labels, we gather public Finnish datasets as proxy data to analyze whether the signals observed in (1) are associated with downstream classification gains. We observe that CPT train-time loss curves differ strongly by domain, and that, in an exploratory analysis, certain CPT-derived features correlate with proxy classification improvement. In particular, this report contributes to the limited literature on NLP for Finnish healthcare data.

We study whether in-domain pretraining of Bidirectional Encoder Representations from Transformer (BERT) model improves subdomain-level detection of exfiltration at low false positive rates. While previous work mostly examines fine-tuned generic Transformers, it does not aim to isolate the effect of pretraining on the downstream task of classification. To address this gap, we develop a controlled pipeline where we freeze operating points on validation and transfer them to the test set, thus enabling clean ablations across different label and pretraining budgets. Our results show significant improvements in the left tail of the Receiver Operating Characteristic (ROC) curve, especially against randomly initialized baseline. Additionally, within pretrained model variants, increasing the number of pretraining steps helps the most when more labeled data are available for fine-tuning.

Large language models (LLMs) have demonstrated remarkable capabilities across diverse domains, yet their adaptation to specialized fields remains challenging, particularly for non-English languages. This study investigates domain-adaptive pre-training (DAPT) as a strategy for specializing small to mid-sized LLMs in the French biomedical domain through continued pre-training. We address two key research questions: the viability of specialized continued pre-training for domain adaptation and the relationship between domain-specific performance gains and general capability degradation. Our contributions include the release of a fully open-licensed French biomedical corpus suitable for commercial and open-source applications, the training and release of specialized French biomedical LLMs, and novel insights for DAPT implementation. Our methodology encompasses the collection and refinement of high-quality French biomedical texts, the exploration of causal language modeling approaches using DAPT, and conducting extensive comparative evaluations. Our results cast doubt on the efficacy of DAPT, in contrast to previous works, but we highlight its viability in smaller-scale, resource-constrained scenarios under the right conditions. Findings in this paper further suggest that model merging post-DAPT is essential to mitigate generalization trade-offs, and in some cases even improves performance on specialized tasks at which the DAPT was directed.

Natural language processing (NLP) advances have powered a generation of LegalTech systems, but Turkish law remains under-served by domain-specific data and models. English has legal encoders such as LEGAL-BERT; no comparable high-volume Turkish counterpart exists. We introduce HukukBERT, a Turkish legal language model trained on a 19 GB cleaned corpus using a hybrid domain-adaptive pre-training (DAPT) recipe that mixes Whole-Word Masking, Token Span Masking, Word Span Masking, and targeted Keyword Masking. We compared our 48K WordPiece tokenizer and DAPT pipeline against general-purpose and existing domain-specific Turkish models. On the Legal Cloze Test - a masked legal term prediction benchmark over Turkish court decisions - HukukBERT reaches 84.40% Top-1 accuracy and beats every baseline we tested. The Legal Cloze Test is synthetically constructed, so its passages are absent from the pre-training corpus by construction, eliminating train-test contamination. On the downstream task of structural segmentation of official Turkish court decisions, it reaches a 92.8% document pass rate. We release HukukBERT to support Turkish legal NLP work in named entity recognition, judgment prediction, and document classification.

Next-generation radio astronomy surveys are delivering millions of resolved sources, yet scalable morphology analysis remains difficult across heterogeneous telescopes and imaging pipelines. We present STRADAViT, a self-supervised continued-pretraining framework for learning transferable radio-astronomy encoders from Vision Transformer (ViT) backbones. It combines mixed-survey data curation, radio astronomy-aware training-view generation, and a ViT-MAE-initialized encoder family with optional register tokens. It supports reconstruction-only, contrastive-only, and two-stage branches. Our pretraining dataset comprises 512x512 radio astronomy cutouts drawn from four complementary sources (MeerKAT, ASKAP, LOFAR/LoTSS, and SKA SDC1 simulated data). We evaluate transfer with linear probing (LP) and fine-tuning (FT) on three morphology benchmarks spanning binary and multi-class settings (MiraBest, LoTSS DR2, and Radio Galaxy Zoo). An exploratory three-fold ablation grid guides selection of a register-based two-stage checkpoint using a fixed cross-dataset criterion. Across subsequent 15-seed paired downstream evaluations on fixed partitions, this checkpoint improves linear-probe Macro-F1 over its ViT-MAE initialization on all three benchmarks and improves fine-tuning on MiraBest and RGZ DR1, while LoTSS DR2 fine-tuning declines; all six differences remain statistically supported after Holm correction. A parallel DINOv2 experiment yields mixed adaptation effects: the procedure transfers, but the benefit is not uniform. STRADAViT thus improves frozen ViT representations while retaining clear dataset-dependent limitations and remaining below task-specialized methods on standard MiraBest classification.

This work presents a collection of multi-modal domain adapted large language models built upon the instruction tuned variants of open weight models (Gemma 3, Qwen 3, Gemma 4) using a relatively small dataset of around 50 million tokens. The dataset consists of open-access additive manufacturing journal articles with data extracted for the domain adaptive pretraining and visual instruction tuning processes. Various stages of the developed model are evaluated with the Additive-Manufacturing-Benchmark which consists of additive manufacturing domain specific tasks compiled published resources. Domain adapted and instruction tuned models exhibit proficiency in both language and vision based tasks, achieving accuracies upwards of 90% in general additive manufacturing knowledge. This domain adaptive pretraining and instruction tuning strategy outline an accessible specialization method for large language models to a domain such as additive manufacturing.

Language models achieve impressive performance on a variety of knowledge, language, and reasoning tasks due to the scale and diversity of pretraining data available. The standard training recipe is a two-stage paradigm: pretraining first on the full corpus of data followed by specialization on a subset of high quality, specialized data from the full corpus. In the multi-domain setting, this involves continued pretraining of multiple models on each specialized domain, referred to as split model training. We propose a method for pretraining multiple models independently over a general pretraining corpus, and determining the optimal compute allocation between pretraining and continued pretraining using scaling laws. Our approach accurately predicts the loss of a model of size N with D pretraining and D' specialization tokens, and extrapolates to larger model sizes and number of tokens. Applied to language model training, our approach improves performance consistently across common sense knowledge and reasoning benchmarks across different model sizes and compute budgets.

Real-world model deployments demand strong performance on narrow domains where data is often scarce. Typically, practitioners finetune models to specialize them, but this risks overfitting to the domain and forgetting general knowledge. We study a simple strategy, specialized pretraining (SPT), where a small domain dataset, typically reserved for finetuning, is repeated starting from pretraining as a fraction of the total tokens. Across three specialized domains (ChemPile, MusicPile, and ProofPile), SPT improves domain performance and preserves general capabilities after finetuning compared to standard pretraining. In our experiments, SPT reduces the pretraining tokens needed to reach a given domain performance by up to 1.75x. These gains grow when the target domain is underrepresented in the pretraining corpus: on domains far from web text, a 1B SPT model outperforms a 3B standard pretrained model. Beyond these empirical gains, we derive overfitting scaling laws to guide practitioners in selecting the optimal domain-data repetition for a given pretraining compute budget. Our observations reveal the finetuner's fallacy: while finetuning may appear to be the cheapest path to domain adaptation, introducing specialized domain data during pretraining stretches its utility. SPT yields better specialized domain performance (via reduced overfitting across repeated exposures) and better general domain performance (via reduced forgetting during finetuning), ultimately achieving stronger results with fewer parameters and less total compute when amortized over inference. To get the most out of domain data, incorporate it as early in training as possible.

We investigate the impact of tokeniser pretraining on the accuracy and efficiency of physics emulation. Modern high-resolution simulations produce vast volumes of data spanning diverse physical regimes and scales. Training foundation models to learn the dynamics underlying such data enables the modelling of complex multiphysics phenomena, especially in data-limited settings. The emerging class of physics foundation models typically aims to learn two tasks jointly: (i) extracting compact representations of high-resolution spatiotemporal data, and (ii) capturing governing physical dynamics. However, learning both tasks from scratch simultaneously can impede the effectiveness of either process. We show that pretraining the tokeniser with an autoencoding objective prior to training the dynamics model enhances computational efficiency for physics emulation. Notably, the magnitude of this benefit depends on domain alignment: pretraining on the same physical system as the emulation task yields the largest improvements, while pretraining on other systems provides moderate gains. In-domain pretraining reduces VRMSE by 64% after 10,500 training steps compared to training from scratch. To our knowledge, this is the first systematic investigation of tokeniser pretraining for physics foundation models. We further introduce flexible spatiotemporal compression operations that extend causal convolutions to support runtime-adjustable compression ratios, enabling efficient adaptation to diverse downstream tasks. Our findings provide practical guidance for training efficient physics emulators and highlight the importance of strategic pretraining data selection.

Automatic summarization of radiology reports is an essential application to reduce the burden on physicians. Previous studies have widely used the "pre-training, fine-tuning" strategy to adapt large language models (LLMs) for summarization. This study proposed a subdomain adaptation through a mid-training method to improve summarization. We explored three adaptation strategies: (1) general-domain pre-training, (2) clinical-domain pre-training, and (3) clinical-domain pre-training followed by subdomain mid-training. We developed models using large-scale clinical text from the University of Florida (UF) Health and conducted mid-training and fine-tuning experiments using widely used benchmark datasets including OpenI and MIMIC-CXR. The experimental results show that the mid-trained model, GatorTronT5-Radio, achieved the best performance, outperforming models without mid-training in both text-based measures (ROUGE-L) and factuality measures (RadGraph-F1). Our mid-training methods also demonstrate better few-shot learning and could alleviate the "cold start" problem reported in previous studies as a learning barrier. Our findings support the use of "pre-training, mid-training, fine-tuning," instead of the widely used direct fine-tuning strategy.

Differential privacy limits how much each training image can influence a medical imaging model but lowers its diagnostic accuracy. Pretrained weights reduce this loss. It is unknown whether the pretraining objective or domain improves accuracy more. Here we train chest radiograph classifiers under differential privacy from five starting points on 591,884 radiographs from five datasets. We show that supervised pretraining on chest radiographs ranks first in 24 of 25 combinations of dataset and privacy setting. Measured by the area under the receiver operating characteristic curve, its lead over supervised pretraining on natural images is 2.6% without privacy and 15.3% at the strictest budget. From that starting point, changing the domain improves this area 2.2 to 3.7 times more than changing the objective. Private pretraining on chest radiographs lowers this area by about 5% under privacy but outperforms the three starting points without patient data at the strictest budget. Under privacy, supervised chest radiograph pretraining leads in the worst-performing demographic subgroup.

The increasing scale of AI workloads demands High-Performance Computing (HPC) infrastructure and training methodologies that are both scalable and sustainable. While Large Language Models (LLMs) demonstrate exceptional natural language capabilities, general-purpose models often lack the specialized domain knowledge necessary for effective cybersecurity analysis. We investigate Domain-Adaptive Continuous Pretraining (DAP) as a scalable, resource-efficient methodology for enhancing cybersecurity understanding in pretrained LLMs, implemented through a distributed Fully Sharded Data Parallel (FSDP) pipeline across multi-node GPU clusters. We systematically adapted three decoder-based architectures -- Llama-3.1-8B, DeepSeek-R1-Distill-Qwen-14B, and Llama-3.3-70B-Instruct -- using a curated 126-million-word cybersecurity corpus from standards, academic literature, and technical documentation. Evaluation across three cybersecurity benchmarks -- CTI-MCQ, CyberMetric, and SecEval -- demonstrates consistent improvements post-adaptation. Notably, our Llama-3.3-70B-Ins-DAP model achieves state-of-the-art performance with accuracies of 0.718, 0.933, and 0.864, respectively, surpassing parameter-efficient baselines and specialized models including Llama-Primus-Base (trained on 2.77 billion tokens) and Foundation-Sec-8B (trained on 5 billion tokens), despite utilizing only 118.8 million tokens -- representing a 23-to-42-fold reduction in training data. Targeted continuous pretraining via scalable HPC infrastructure enables effective cybersecurity domain adaptation with a substantially reduced computational and energy footprint, supporting specialized AI assistants in threat analysis, vulnerability assessment, and security documentation, while advancing sustainable and responsible AI development.