A compact English instruction–response dataset for supervised fine-tuning (SFT) of writing/“ghostwriting”-style assistants. It contains 2,130 examples across train (1,917) and test (213) splits, stored as Parquet on the Hub.
Dataset Details
Dataset Description
llmGhostWriter pairs an instruction (prompt) with an output (target response). The schema is two string fields and is directly compatible with SFT and conditional… See the full description on the dataset page: https://huggingface.co/datasets/ahmedshahriar/llmGhostWriter.
