3 ms·
> related: I imagine in the future we might several "expert" LLM's and a wrapper can delegate tasks as needed as if it were a "tool". That way we can have segre
by koakuma-chan 2y ago
> related: I imagine in the future we might several "expert" LLM's and a wrapper can delegate tasks as needed as if it were a "tool". That way we can have segregation of expertise - each individual model can excel at one single thing.
In the future? I'm pretty sure people do that already.
- energy123 2y agoIt's crudely done though.
- kratom_sandwich 2y agoMistrals model is a mixture-of-experts model
- simianwords 2y agoNo I disagree. I would want ChatGPT to abstract away expert models - biochemistry model, coding model, physics model and maybe O3 would use these models as tools to come up with an answer. The point being that a separate expert model would be better at its own field than a single model that tries to be good at everything. Intuitively it makes sense, in practice I have seen anecdotes where finetuning a small model on domain data makes the model lose coherence on other topics.
- koakuma-chan 2y ago> have seen anecdotes where finetuning a small model on domain data makes the model lose coherence on other topics This is expected behaviour.
- simianwords 2y agoi know. so why don't we have domain specific models as tools in consumer llm products