Rossetta is a dataset created to train models that transform raw human prompts into structured, optimized, and enriched instructions. It is designed for research in prompt engineering, instruction rewriting, and natural-language-to-instruction compilation.
Dataset Details
Dataset Description
Rossetta provides paired examples that map a raw, unrefined human prompt (prompt_original) to a structured, context-rich… See the full description on the dataset page: https://huggingface.co/datasets/msoler18/rossetta.
