Indic PersonaHub Domain QA Synthetic Corpus is a large-scale synthetic instruction dataset designed for the Indian context. It was generated using a synthetic population reflecting India's linguistic, cultural, occupational, and cognitive diversity, with personas tagged by domain expertise. The dataset was last updated on May 7, 2026.
Use Cases
- Training models for India-specific reasoning based on domain-anchored persona questions.
- Evaluating long-form instruction following based on detailed persona-conditioned answers.
- Multilingual instruction tuning across 22 Indian languages based on aligned translations.
- Testing cultural compliance of language models based on culturally grounded questions and answers.
Strengths
- Each sample is anchored to a unique synthetic persona tagged with a domain of expertise.
- Answers are detailed responses of approximately 800–900 words, designed for high-entropy text.
- Quality is enforced using two automated LLM-based evaluators for cultural compliance and task relevance.
- The underlying persona population was constructed using a dual strategy to capture long-tail diversity.
Limitations
- Description metadata is limited; actual data quality requires manual inspection after download.
- Column-level documentation is absent; field semantics must be inferred after download.
- Row count is unknown, which may limit suitability assessment.
Provenance
- Source
- Pre-Training-Corpus
- Collection Method
- Generated using Indic PersonaHub, a synthetic population, with a persona-conditioned two-step pipeline for question generation and answering.
- Freshness
- Last updated 2026-05-07 06:07:12; freshness should be verified.
- Geography
- India