Photo by Felipe Albertella on Unsplash
AI
Microsoft researchers derail LLM safety measures with one prompt
With little loss of utility – and it's not just language models either.
