| license: cc-by-4.0 | |
| language: | |
| - ja | |
| task_categories: | |
| - text-generation | |
| tags: | |
| - space | |
| - science-explainer | |
| - japanese | |
| - original-content | |
| size_categories: | |
| - n<1K | |
| # Space AI Explainer JA | |
| 100 original Japanese Q&A rows (×2 instruction templates = 200 examples) explaining space and AI basics: rockets, orbital mechanics, satellites, deep space, machine learning, and their intersections. Answers are short maintainer-written explanations of established basics with concept-level source notes. No scraped text, no invented citations. | |
| ## Load with Python | |
| ```bash | |
| pip install datasets | |
| ``` | |
| ```python | |
| from datasets import load_dataset | |
| data = load_dataset("LoNebula/space-ai-explainer-ja") | |
| print(data) | |
| print(data["train"][0]["input"], "→", data["train"][0]["output"]) | |
| ``` | |
| ## Split | |
| Deterministic 90/10 split by question (seed `20261014`). Both prompt variants of a question stay in the same split. | |
| ## Limitations | |
| Short introductory explanations, not expert references. Representative numbers may be rounded. Verify technical claims independently before relying on them. | |
| ## License | |
| CC BY 4.0. | |