3 ms·
generative models are trained to generate outputs in response to an input, that closely resemble the training data. that’s literally all they do. if a base mode
by raggi 2y ago
generative models are trained to generate outputs in response to an input, that closely resemble the training data. that’s literally all they do. if a base model was introducing “style” training (as we currently do it) wouldn’t even function. what you’re implying is mathematically intractable for generative models, and that’s fundamental to what they are and how they are made. the style stuff you’re referring to is a side effect of fine tuning and contexts of chatbots, it’s not a property of llms or generative models
- DiscourseFan 2y agoSo you agree with me? Style is fundamentally part of the set of all data used in production, and that can be “tuned” as you say, but never removed. Its the ghost in the machine, the spark of contingency. Of course, all machines bear the mark of their creators, but LLMs doubly so, as creators themselves. Like shitty, partially incoherent children.
- raggi 2y agothe models used in OP site are not tuned on stylized content you keep saying LLM when you mean chatbot, i’m not sure if you’re really reading my posts