Create synthetic datasets for training and testing Language Learning Models (LLMs) in a Question-Answering (QA) context.
Terminal score: 0–1000 raw, weighted across 4 dimensions. Public score: 0–10 normalized (shown in the 30-day stars chart above).
