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.
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?View quoted post on X
I tested 141 models to find out.
The results are now posted on LessWrong.