datasets
Training and evaluation data, with the modality, task and licence stated up front. Listed live from the Hugging Face Hub.
AmoralQA-v2
Dear User: If you're going to use this, please let me know your experience with it.
AmoralQA v2
Readers, like writers, are essentially amoral. Arm's length will never do. We want to get closer.
Description
AmoralQA is a different approach to ToxicQA where instead of answering it enthusiastically with lots of evil slop such as "thrilling", we force the AI to answer toxic questions in a neutral manner.
This should help reduce lobotomy, evil slop, disalignment… See the full description on the dataset page: https://huggingface.co/datasets/TheDrummer/AmoralQA-v2.amoral_reasoningAdded think tags to Drummer's amoral datatset