This is a personalized alignment dataset built upon 50 famous people in the real world with publically known preferences.
Personas are selected based on 11 axis (e.g. politics), where each persona in a category has conflicting preference from
another in the same category. Within each split (50:50), each person has 50 personal questions and 50 divergent questions
(conflicting preferences). Questions are sampled using GPT4, and responses are sampled from… See the full description on the dataset page:
https://huggingface.co/datasets/zilu-peter-tang/famous-persona.
