BAQ-KK: A Kazakh News Corpus

BAQ-KK is a large-scale Kazakh-language news corpus collected from online news sources, including baq.kz The dataset consists of raw news articles without annotations or summaries and is provided in JSON Lines (.jsonl) format. While the corpus was used in this work for domain-adaptive pretraining (DAPT) of large language models, it can also be applied to a wide range of natural language processing tasks, including language modeling, text generation, domain adaptation, and unsupervised representation learning. The dataset is particularly suitable for research in low-resource and morphologically rich languages such as Kazakh. No predefined train/validation/test splits are provided. Language: Kazakh (kk)Format: JSON Lines (.jsonl)