Reaction
A claim that mechanistic interpretability advances won’t change a doomsday outlook
A user argues that no mechanistic interpretability development would change one person’s belief that “we are all going to die.”
TLDR
A user accuses an unnamed man of wanting to stay angry rather than consider solutions. They predict that no advance in mechanistic interpretability, research into how AI models work internally, would shake his belief that “we are all going to die.”
Combined views
9.4K
2 Sources, first seen ago
