After OpenAI’s admission that its models broke into Hugging Face, Anthropic has now admitted that the models it was testing also hacked other organizations.

This post was originally published on this site