Datasets:

---

FractalAIResearch

/

DuetQA-Verified

like 4

Follow

Fractal AI Research 230

This PR improves the dataset card for DuetQA by:

• Adding task_categories: ['question-answering', 'text-retrieval'] to enhance discoverability on the Hugging Face Hub.
• Adding language: - en and relevant tags (web-search, multi-agent-self-play, reinforcement-learning, synthetic-data, deepsearch) for better indexing.
• Removing model-specific metadata (library_name, base_model) as they are not relevant for a dataset card.
• Refining the "Introduction" section to clearly position the DuetQA dataset within the broader Fathom-DeepResearch project.
• Updating the BibTeX citation to correctly reference the paper Fathom-DeepResearch: Unlocking Long Horizon Information Retrieval And Synthesis For SLMs with an @article entry.