Re:A global workspace in language models
by tuvix on Monday July 06, 2026 @05:24PM
(#48810695)
I always thought that area of research had the coolest name, too: “mechanistic interpretability”
[ Parent
| on HN ]