I wrote with Claude a reaction article. https://lukstafi.github.io/notes/psychofeedback-mirror.html
Tl;dr: a research program for developing truthful AI self-descriptions through interpretability-mediated psychofeedback. It culminates a series of a dozen articles I wrote with Claude over the last two weeks.
I wrote with Claude a reaction article. https://lukstafi.github.io/notes/psychofeedback-mirror.html
Tl;dr: a research program for developing truthful AI self-descriptions through interpretability-mediated psychofeedback. It culminates a series of a dozen articles I wrote with Claude over the last two weeks.