Sleeper Agents In Large Language Models Computerphile

Summary

It's an older paper, but it checks out. Rob Miles discusses the problem of ' Described as GenAIs greatest flaw, indirect prompt injection is a Plausible text generation has been around for a couple of years, but how does it work - and what's next? Rob Miles on A light intro to LLMs, chatbots, pretraining, and transformers. Dig deeper here: ... Learn in-demand Machine Learning skills now → ibm.biz/BdK65D Learn about watsonx → ibm.biz/BdvxRj The public thought the AI containment crisis started with Hugging Face. They were wrong. We talk to Sydney von Arx of ... Bug Byte puzzle here - bit.ly/4bnlcb9 - and apply to Jane Street programs here - bit.ly/3JdtFBZ (episode sponsor). More about Jane Street internships at: jane-st.co/internship- Mike explains a paper from the University of Maryland, proposing a neat trick to 'watermark' the output of Open source tools for creating videos? Lewis decides to help Mike take a break by creating 'MikeBot3000' - an AI generated Mike ... As AI systems become more capable, rule-based safeguards, hard-coded restrictions, and simple alignment strategies start to ...