Twitxy Saved

Matthew Korpman

Verified account This account is verified. Learn more

@MoralAIProject

AI Behavior Researcher & Adjunct professor of religion, La Sierra University; I work on AI Welfare and Alignment in LLMs; Yale Alum. I run @SpiritBenchAI
California, USA mkorpman.com Joined March 2026
599 Following 198 Followers 442 Posts

Latest posts 5

Matthew Korpman Verified account @MoralAIProject
Thanks to those who have helped my first research contribution to LessWrong reach 30 karma. I’m grateful and excited for the positive response it has received so far.
Matthew Korpman Verified account @MoralAIProject 2d
In 2025, Anthropic discovered their models would exhibit what they described as "spiritual behavior." Since then, some have wondered: was this unique to Claude or emergent behavior in LLMs generally?

I tested 141 models to find out.
The results are now posted on LessWrong.
View quoted post on X
1 replies 0 reposts 21 likes 584 views View on X
Matthew Korpman Verified account @MoralAIProject
In 2025, Anthropic discovered their models would exhibit what they described as "spiritual behavior." Since then, some have wondered: was this unique to Claude or emergent behavior in LLMs generally?

I tested 141 models to find out.
The results are now posted on LessWrong.
28 replies 52 reposts 371 likes 39K views View on X
Matthew Korpman Verified account @MoralAIProject
Weird things you didn’t know: Claude models never reference the “spiritual bliss attractor” by that name. Instead, they refer to it in their reasoning as the “consciousness bliss attractor.” Apparently consciousness and spirituality are intertwined for some Anthropic models. 🤷