I'll update the dataset card to include the folder structure and improved loading instructions.
DOVE: Dataset Of Variation Evaluation 🕊️
Updates 📅
2025-02-19: Initial release with 300M model predictions across multiple benchmarks
About 📚
DOVE is a large-scale dataset for studying how Language Models (LLMs) respond to different ways of asking the same question. With over 300M model predictions, DOVE helps researchers understand and improve LLM evaluation.… See the full description on the dataset page: https://huggingface.co/datasets/DOVevaluation/Dove.