So LLMs need an ELI5 mode and detect when to use it. This would use the non-technical terminology but pull concepts from the more jargony model. This may or m

So LLMs need an ELI5 mode and detect when to use it. This would use the non-technical terminology but pull concepts from the more jargony model. This may or may not be easily possible by tweaking current training techniques. But it shows the many edge cases still needed to be addressed by AI models.