datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
closedqa-eval-11-v1danish-wiki-closedqa-v1
Danish Wiki closed-QA v1
Factual question-and-answer pairs generated from Danish Wikipedia via
google/gemma-3-12b-it. Each row is one question + short factual answer
grounded in a specific Wikipedia article.
Article selection
Articles were scored for general-knowledge salience and filtered before
generation. Signals combined:
page_len — article byte length (from page.sql.gz)
langlinks — number of other language editions of the article
(from langlinks.sql.gz) —… See the full description on the dataset page: https://huggingface.co/datasets/jensjepsen/danish-wiki-closedqa-v1.danish-wiki-closedqa-stem-v1closedqa-response-11-v1claude3sonnet-closedqa-eval-by-claude3sonnetgemma2b-closedqa-eval-by-claude3sonnetclosedqa-claude3sonnet-responsellama3-8b-closedqa-eval-by-gemini15flashclaude3sonnet-closedqa-eval-by-gemini1_5flashgpt4o-closedqa-eval-by-gemini1_5flashclaude3sonnet-closedqa-eval-by-gpt-4ogemma2b-closedqa-eval-by-gemini15flashgemma7b-closedqa-eval-by-claude3sonnetgpt4o-closedqa-eval-by-claude3sonnetgemini1_5flash-closedqa-eval-by-gpt-4ogemma7b-closedqa-eval-by-gemini15flashgemini1_5flash-closedqa-eval-by-claude3sonnetgemini1_5flash-closedqa-eval-by-gemini1_5flashclosedqa-gemini1_5flash-responseclosed-QA-vi
Dataset Card for "closed-QA-vi"
More Information needed
closedqa-evalflan2021-closed-book-qadolly_closed_qa
Dataset Card for "dolly_closed_qa"
More Information needed
closedqa-eval-11-v2llama3-1b-closedqa-eval-by-gemini15flashvietnamese_legal_closed_qa_validation
Evaluation dataset for Vietnamse Legal RAG Chatbot
The dataset is manually filterred and refined by humans. Each sample in the dataset contains a question, K contexts and an answer generated by OpenAI GPT-4. The answer is refined by humans.
gpt4o-closedqa-eval-by-gpt-4oclosedqa-eval-11-v2.1llama3-1b-closedqa-eval-by-gpt4ollama3-8b-closedqa-eval-by-claude3sonnet
