LLMs have been known to display different personas. It seems likely that the space of personas is low-rank. Can we start mapping out this space by just looking at what sets of likes and dislikes LLMs are willing to take on with relatively little training (set some arbitrary threshold of N tokens) and which likes and dislikes feel incongruent to an LLM? Are these preferences similar across LLMs?
If you are inspired by this idea, you can reach out to the authors for collaboration or cite it:
@misc{holtzman-mapping-person-space-2026,
author = {Holtzman, Ari},
title = {Mapping Person Space Through Simple Likes and Dislikes},
year = {2026},
url = {https://hypogenic.ai/ideahub/idea/5GdCcopu2ih1dgKmPXu5}
}Please sign in to comment on this idea.
No comments yet. Be the first to share your thoughts!