five

Draeg82/uk-gdpr-small-business-qa

收藏
Hugging Face2026-05-27 更新2026-05-31 收录
下载链接:
https://hf-mirror.com/datasets/Draeg82/uk-gdpr-small-business-qa
下载链接
链接失效反馈
官方服务:
资源简介:
--- license: mit language: - en tags: - gdpr - uk-gdpr - compliance - legal - small-business - qa - synthetic - deepseek - qwen size_categories: - 1K<n<10K task_categories: - question-answering pretty_name: UK GDPR Small Business Q&A --- # UK GDPR Small Business Q&A 1,000 question-answer pairs covering UK GDPR and Data Protection Act 2018 compliance for small businesses. ## Dataset Description This dataset provides practical, citable answers to real-world GDPR questions faced by UK small business owners, managers, and employees. Each answer references specific UK GDPR articles, ICO guidance documents, and Data Protection Act 2018 provisions where relevant. Topics covered include consent, data breaches, subject access requests (SARs), data protection impact assessments (DPIAs), marketing under PECR, international transfers, children's data, employee monitoring, cloud services, and more. ## Generation Method - **Questions**: Generated locally using Qwen 2.5 14B (Q4_K_M) with a 370-term GDPR concept bank and dual strategy approach (term-anchored + scenario-anchored) to ensure diversity. Continuous deduplication via exact and fuzzy (Jaccard) matching. - **Answers**: Generated via DeepSeek API (`deepseek-chat`) with strict UK GDPR and DPA 2018 prompting. The answer model was instructed to provide specific article references, ICO guidance citations, and practical steps. No real user data was used. Entirely synthetic. ## Data Format Both JSON (newline-delimited) and Parquet formats available. | Field | Type | Description | |---|---|---| | `question` | string | The compliance question | | `answer` | string | Expert answer with article/guidance references | | `terms` | list[string] | GDPR concepts used to generate the question | | `strategy` | string | Generation strategy (A = term-based, B = scenario-based) | | `answer_model` | string | Model used for answer generation | | `generated_at` | ISO 8601 timestamp | When the Q&A pair was generated | ## Limitations - Synthetically generated — not reviewed by a qualified legal professional - Answers reflect UK GDPR and DPA 2018 as of the generation date - May not reflect the latest ICO guidance or regulatory changes - Not a substitute for professional legal advice ## Extended Versions This 1K sample is free. Larger, commercially-licensed datasets and fine-tuned models are available: | Product | Size | Price | |---|---|---| | GDPR Q&A Dataset (Extended) | 5,000 pairs | £19 | | GDPR Q&A Dataset (Full) | 10,000 pairs | £34 | | GDPR Fine-Tuned Model (Qwen 2.5 3B) | Q4_K_M GGUF | £24 | | Bundle — 10K Dataset + Model | — | £89 | [Get the 5K dataset →](https://draeg82.gumroad.com/l/uk-gdpr-5k) · [Browse all products →](https://draeg82.gumroad.com) ## License MIT
提供机构:
Draeg82
二维码
社区交流群
二维码
科研交流群
商业服务