Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
vidarh
on March 21, 2023
|
parent
|
context
|
favorite
| on:
GPT-4 and professional benchmarks: the wrong answe...
I
said that,and you're missing the point. We don't need to understand the models to be able to evaluate
the output
manually.
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search: