RE: Advanced Large Language Models Are Capable And Prone to In-Context Scheming
Yeah, the author of the video shows a table of the ways various models cheat. It's quite an interesting read.
RE: Advanced Large Language Models Are Capable And Prone to In-Context Scheming