3 ms·

> Yes, but that’s orthogonal to the training set. Of course. I didn't say otherwise. > An LLM that ... Note that I also wrote "AI" rather than "LLM". ;) H

• dragonwriter 3y ago
  > Hopefully there will be better alternatives (to LLMs) at some point that don’t hallucinate. There are AIs that don’t “hallucinate” (a bad metaphor), e.g., most of the technologies developed in the previous rounds of AI – like expert systems. I doubt that we will find more advanced, equally-or-more general AI techniques than LLMs that don’t “hallucinate”, what we’ll probably do is find better ways to build systems around LLMs (or more future AI architectures) that allow them to resolve “hallucination” better before presenting results. It’s not as if natural intelligences don’t both literally hallucinate and, more relevant to what is described as “hallucination” in AIs, confabulate when pressed to answer without access to facts, and we know with them that grounding in physical reality via sensory data and having access to factual references doesn’t mitigate those effects (and we’ve also been able to demonstrate the latter with LLMs.)