News for hackers. Stuff that matters.
Re:The Dual LLM pattern for building AI assistants that can ... by datascientist on Monday May 15, 2023 @01:53PM (#35951957)
Alignment is hard: "If the LLM has finite probability of exhibiting negative behavior, there exists a prompt for which the LLM will exhibit negative behavior with probability 1." Source: Fundamental Limitations of Alignment in LLMs https://arxiv.org/abs/2304.11082

There are no comments on this comment yet.