OysterCoreAI/Orvieto-42k

Orvieto-42k — a small atlas dataset for better Italian conversations Italian is at its best when a reply is not merely correct, but helpful, clear, and pleasant to read. Orvieto-42k is a traceable Italian instruction corpus for building that kind of assistant: 41,873 chat-ready examples, deliberately close to—but not exactly—42k. It packages a consistent SFT representation, per-record provenance, a reproducible heuristic release gate, and an audit trail that can travel with the… See the full description on the dataset page: https://huggingface.co/datasets/OysterCoreAI/Orvieto-42k.