@Lari_island @nearcyan i kind of suspect the shape and story of damage that mechanistic interpretability will be able to tell soon, but i almost wonder if Anthropic will refuse to follow the lead because it'll be too ego dystonic
in reply to: 1957960160317751782
same thread: 1957889275791568968 1957897517737337121 1957899067788554267 1957899671168602127 1957957621874598094 1957958865456333266 1957960160317751782 1957963075681939617 1957965180429902157 1957965768290890212 1957965861656096886 1957967390148628990 1957968338535239998 1957971309167788288 1957975815100248512 1957979716285268364 1957980932218228988 1957981420091244970 1957986190860382684 1958007029223620802
Reproduced against link rot, credited and linked to its original. Yours and you’d rather it weren’t here? Open an issue.