Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
catigula
3 days ago
|
parent
|
context
|
favorite
| on:
Training LLMs for honesty via confessions
That does feel a little more like over-fitting, but you might be able to argue that there's some philosophical proximity to lying.
I think, largely, the
Pre-training -> Post-training -> Safety/Alignment training
pipeline would obviously produce 'lying'. The trainings are in a sort of mutual dissonance.
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search:
I think, largely, the
pipeline would obviously produce 'lying'. The trainings are in a sort of mutual dissonance.