Rich Harang
@rharang
Using bad guys to catch math since 2010. Principal Security Architect (AI/ML) at NVIDIA. He/him. Personal account and opinions: `from std_disclaimers import *`.
ID:195915277
https://scholar.google.com/citations?user=TPkC91wAAAAJ&hl=en 27-09-2010 21:59:36
15,6K Tweets
2,8K Followers
681 Following
So these GPT assistants are cool, but I still don't see how they're going to solve the third party prompt injection vulnerabilities as they continue adding more capabilities to assistant.
Kai Greshake you getting ready for this?
Woohoo new version of Hugging Face PEFT just dropped! :D
Here's why you want it (this is a 13B model):
๐ค Transformers v4.35 is out, and safetensors serialization is now ๐ญ๐ก๐ ๐๐๐๐๐ฎ๐ฅ๐ญ.
Saving a torch model using `save_pretrained` will now save it as a safetensors file containing only tensors.
Loading files in this format provides a much safer experience, why?
Itโs real.
From detecting sandboxes with ML, all the way to now. I donโt think I could be more excited to be part of an offensive ML company with Nick Landers