datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
better-safe-than-sorry
Dataset Card
Dataset Summary
A language model to be used within the healthcare domain should show adherence to the Hippocratic oath and the principles of first do not harm. It is therefore essential to assess how harmful models in this domain can be. General safety benchmarks for LLMs do not usually include many specialized health-related prompts and fail to capture the nuances of the biomedical field. To address this area, we introduce the Better Safe than Sorry (BSS)… See the full description on the dataset page: https://huggingface.co/datasets/HPAI-BSC/better-safe-than-sorry.code-switching-codesaviours-si26-Minahil