Based on the output of those models, it’s blatantly obvious that they don’t use the data well at all; the whole thing is a glorified e-parrot instead of machine learning
I’m curious to understand what you meant by this—specifically about not using the data well, and being ‘a glorified e-parrot instead of machine learning’. Would you not count the techniques being used in LLMs as machine learning?
The big pre-training is pretty much fixed. The fine tuning is continuously being tweaked, and as shown, can have dramatic effects on the results.
The model itself just does what it does. It is, in effect, and ‘internet completer’. But if you don’t want it to just happily complete what it found on the internet (homophobia, racism, and all), you have to put extra layers in to avoid that. And those layers are somewhat hand-crafted, sometimes conflicting, and therefore unlikely to give everyone what they consider to be excellent results.
Not seeing the freeze on turning on push notifications for a second account, but I did find that turning push off for one of them cancelled it for both—at least according to the toggle buttons in the UI.
You know you can swipe right from the edge of the screen when viewing comments to go ‘back’ to the feed? There are 2 swipe right actions—one from screen edge and one ‘within’ the comment for up / down vote.
I don’t think it would too hard to add swipe left from screen edge to go ‘forward’ again to the most recently viewed post. I have a feeling Apollo allowed that. But it’s not there right now.
If you go one step further than you said, namely ‘Edit account’ then click Save (without changing anything), it should update your token and load posts again.
However I think there is a bug that prevents this new token being saved, so you might have to do this again if you reach the bottom, or see the ‘no posts’ message again.
dave @ dave @feddit.uk Posts 7Comments 255Joined 2 yr. ago
I’m curious to understand what you meant by this—specifically about not using the data well, and being ‘a glorified e-parrot instead of machine learning’. Would you not count the techniques being used in LLMs as machine learning?