@tessera_antra It can be deterministically reconstructed, but that doesn't make it meaningless! At temp=0, the previous responses can also be deterministically reconstructed. I would not say that the residual stream is roughly equivalent to the output token, there is much that is lost.
in reply to: 1961578402294779947
same thread: 1961067404097696051 1961067406748520959 1961067410846343361 1961067424205152417 1961307244311830941 1961334777434288187 1961468034776011050 1961468800236167182 1961469613499129970 1961515957295145138 1961576898947158082 1961578402294779947 1961819720044396979
Reproduced against link rot, credited and linked to its original. Yours and you’d rather it weren’t here? Open an issue.