Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
seydor
on July 24, 2023
|
parent
|
context
|
favorite
| on:
Attention Is Off By One
Can i please ask why lim{x->-inf} softmax(x) = 1/k ?
hawkice
on July 24, 2023
[–]
It's splitting the probability across all x_i equally, even when they're all massively negatively weighted.
This change would have all the probability going into the null option, in that case, basically.
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search:
