Links to both above. Trying to explore some of the custom LLMs that are out here but trying to be sure, relatively speaking, of the security of using such models first.
Which of these is better for EndevOS from a security standpoint and, additionally, is there a way to vet the security of some of these custom LLMs such as Bolt or Dolphin 72B?
Also, would running in Hardened Kernel be sufficient or should I look at running the AI software in a container using VMM?