Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
mistercow
6 months ago
|
parent
|
context
|
favorite
| on:
Refusal in language models is mediated by a single...
I wonder if you could do this with multiple alignment training passes, where you extract the refusal direction each time, and suppress it in future training passes.
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search: