Hacker News
new
|
past
|
comments
|
ask
|
show
|
jobs
|
submit
login
pixl97
58 days ago
|
parent
|
context
|
favorite
| on:
OpenAI and Hugging Face address security incident ...
The interesting thing here is a unaligned 'weak' model can leave persistent data all over the internet that then gets used to train the next model to be even more deceiving.
Guidelines
|
FAQ
|
Lists
|
API
|
Security
|
Legal
|
Apply to YC
|
Contact
Search: