2 ms·
Unremarkable base model will remain an unremarkable fine-tuned model that memorised a couple thousand of input-output pairings.
by macleginn 3mo ago
Unremarkable base model will remain an unremarkable fine-tuned model that memorised a couple thousand of input-output pairings.
- danw1979 3mo agoYes, neural networks are famously poor at generalising.
- macleginn 3mo agoThey are poor at generalising from a small number of examples; this is why the real generalisation power is achieved in pre-training.
- ACCount37 3mo agoHa ha, as if. Base models have a lot of capabilities - arranged in all the wrong ways for high performance reasoning and problem-solving. The power of fine tuning on "a couple thousand of input-output pairings" is that it can fix some of that. If your pairings are very well chosen, that is.
- Laurel1234 3mo agoIf that were the case, Anthropic wouldn't be throwing a fit over distillation "attacks".
- freejazz 3mo agoWhy? They often don't make sense. They send DMCA takedowns over materials they can't even copyright, for example. They fessed up to creating shadow libraries that they didn't even use in their training corpus, resulting in the largest copyright settlement ever. Your reasoning is flawed.