Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
jdw64
21 days ago
|
parent
|
context
|
favorite
| on:
Guess which of these LLM outputs is watermarked
8/10. It was harder to distinguish than I expected. If they had applied something like a humanizer skill, it probably would have been nearly impossible to tell.
kalkin
21 days ago
[–]
It sounds like you have a publishable paper - you've broken Aaronsen's watermarking scheme. Congratulations!
(Or maybe you got lucky.)
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search: