aarajbhattarai/nepali-agri-gov-instruct
Nepali Source-Grounded Instruction Dataset Synthetic Nepali instruction-tuning data generated with NVIDIA NeMo Data Designer from authoritative Nepali documents (agriculture manuals, legal texts). Answers are grounded strictly in the source; unanswerable questions get an explicit refusal. Records use chat messages format plus metadata and per-record quality_scores (grounding / correctness / naturalness, 1-5, LLM-as-judge). One data/train-<shard>.jsonl per source document; shards… See the full description on the dataset page: https://huggingface.co/datasets/aarajbhattarai/nepali-agri-gov-instruct.
Nepali Source-Grounded Instruction Dataset
Synthetic Nepali instruction-tuning data generated with NVIDIA NeMo Data Designer from authoritative Nepali documents (agriculture manuals, legal texts). Answers are grounded strictly in the source; unanswerable questions get an explicit refusal. Records use chat messages format plus metadata and per-record quality_scores (grounding / correctness / naturalness, 1-5, LLM-as-judge). One data/train-<shard>.jsonl per source document; shards are overwritten idempotently on re-runs.
This is the TRAINING repo: records judged and passing the quality gate (grounding>=4, correctness>=3, naturalness>=3). Companion repos hold quality-gate rejects (rejected-*) and records whose judge call failed (unjudged-*).