Hacker News
new
past
comments
ask
show
jobs
points
by
bitwize
2 hours ago
|
comments
by
ctrust
2 hours ago
|
[-]
The accurate inferences that can be drawn from subtle linguistic attributes should freak you out more than they do.
reply
by
alchemism
1 hours ago
|
parent
|
[-]
Switching one good synonym can send the model off an entirely different direction in response, or so I’ve observed.
reply