Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
LeoPanthera
4 months ago
|
parent
|
context
|
favorite
| on:
Gemma 3 Technical Report [pdf]
This only works until it doesn't. Start with a model that simply hasn't been trained on anything your shareholders find objectionable, and there will be nothing to reveal with abliteration.
xpl
4 months ago
|
next
[–]
Maybe there exists a dataset consisting
entirely
of objectionable content, so people can finetune neutered models on it?
anticensor
4 months ago
|
parent
|
next
[–]
PH maybe?
Sharlin
4 months ago
|
root
|
parent
|
next
[–]
More like literotica.
xpl
4 months ago
|
root
|
parent
|
prev
|
next
[–]
I mean not only sex, but also swearing, drugs, violence, etc. Basically everything R-rated (but not illegal) which usually gets censored.
anticensor
4 months ago
|
root
|
parent
|
next
[–]
PH is not porn-only. A significant portion of non-porn content also exists there.
anticensor
4 months ago
|
prev
[–]
Such models would actually run against their long term interests of being able to automate away the work currently done by humans.
Consider applying for YC's Fall 2025 batch! Applications are open till Aug 4
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search:
