datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
german-public-sector-c_dbr
Dataset Card for public_sector_c_dbr_QA
Dataset Summary
public_sector_c_dbr_QA is a German-language QA dataset for public-sector and legal-administrative content. The dataset includes question, answer, source context, and LLM-as-a-judge quality metadata.
Dataset Files
Main file: public_sector_c_dbr_QA.jsonl
Base evaluated file: c_dbr_evaluated_top_20_percent.jsonl
Suggested split mapping: train only
Record Counts… See the full description on the dataset page: https://huggingface.co/datasets/Anirbanbhk/german-public-sector-c_dbr.german-public-sector
Dataset Card for public_sector_QA
Dataset Summary
public_sector_QA is a German-language question answering dataset focused on public sector and administrative-domain content. The file contains curated QA samples with source context and LLM-as-a-judge evaluation metadata.
This dataset was created from top-20-percent filtered outputs of three evaluated source files:
anlage-5-handbuch-offene-verwaltungsdaten_1_evaluated_top_20_percent.jsonl… See the full description on the dataset page: https://huggingface.co/datasets/Anirbanbhk/german-public-sector.