Hacker News new | ask | show | jobs
by nielsole 491 days ago
With a bit of fiddling you should be able to get the LLM to translate/summarize the thinking process. Not a 1:1 thing, but still
1 comments
WithinReason 491 days ago
how would you do it?
link
nielsole 491 days ago
my naive way would be to try to do seq2seq with the hidden state as input. Not sure how to replace the supervised samples though.
link
WithinReason 491 days ago
OK but what would you use as ground truth?
link