Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
petercooper
6 months ago
|
parent
|
context
|
favorite
| on:
Claude Opus 4.6
One possible trick could be to search and replace them all with nonsense alternatives then see if it extracts those.
andai
6 months ago
|
next
[–]
That might actually boost performance since attention pays attention to stuff that stands out. If I make a typo, the models often hyperfixate on it.
jazzyjackson
6 months ago
|
prev
[–]
A fine instruction following task but if harry potter is in the weights of the neural net, it's going to mix some of the real ones with the alternates.
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search: