×
Ethics & Content Warning
Please note that this project does not aim to amplify harmful ideology. Posts were filtered for
spiritual and psychosis-coded language, not slurs or targeted
content. The resulting models are intentionally small and unstable, serving
as artistic and analytic objects rather than polished chatbots.The goal is
to interrogate how online communities collectively construct symbolic worlds,
not to reproduce or legitimize harmful viewpoints.
However, the model does not censor profanity or "harmful" keywords as this project is not interested in defining such parameters and seeks only to create a representative model of the data. Therefore, the oracles responses may be offensive to some -- please keep this in mind while prompting.