10/07/2024 10:41:22 AM PDT
· 9 of 31 glorgau
to ConservativeMind
I use a local road with a 600 ft climb for my sprint training. Mid 60s and I’m far more likely to get hit and killed by a car than die from a heart attack ;-)
What it comes down to is the RAG placed in front of most all public facing LLMs. It’s also done for in-house LLMs.
RAG (Retrieval-Augmented Generation) is an AI framework for retrieving facts from an external knowledge base to ground large language models (LLMs). You pretty much have to place a RAG in order to hide the “hallucinations” that LLMs generate to fill in gaps in their responses. Basically, they filter all the problematic requests and give preprogrammed responses such as “it is inappropriate to make a joke about Mohammed but not about Jesus”