kamaalg/azerbaijani-instructions

Azerbaijani Instruction Dataset (v0) Azerbaijani (instruction, response) pairs for supervised fine-tuning (SFT) of Azerbaijani language models — part of an open Azerbaijani LLM stack. Instruction data is genuinely scarce for Azerbaijani; this is both training data for our models and a reusable standalone artifact for anyone building Azerbaijani instruction-following models. Contents seeds_az.jsonl — 45 hand-authored, high-quality seed pairs spanning 16 task… See the full description on the dataset page: https://huggingface.co/datasets/kamaalg/azerbaijani-instructions.