Two small, page-anchored Thai corpora prepared for retrieval experiments and classroom-style RAG work. The aim is to make the source structure easier to explore while keeping a clear path back to the original document.
This dataset sits behind a small Qwen3.6-35B-A3B Q5 versus OpenThai 2.0 Legal BF16 test. The short write-up, result table and UI captures are on GitHub:… See the full description on the dataset page:
https://huggingface.co/datasets/thodsapon/thai-ncb-digital-fraud-structural-rag.
