This dataset contains 86 high-quality, open-ended question–answer pairs generated from the U.S. Department of Defense’s Emergency War Surgery (EWS) manual. The content is extracted, chunked, and QA-labeled for use in sovereign AI systems, medical QA fine-tuning, or local RAG pipelines.

Contents

Source: Emergency War Surgery (public domain) Chunks: 450-token segments with overlap QA Pairs: 1–2 generated per chunk, validated and… See the full description on the dataset page: https://huggingface.co/datasets/Paxrad/EWS_v5_USONLY_final.