65
this post was submitted on 22 Jan 2025
65 points (98.5% liked)
AI
4592 readers
2 users here now
Artificial intelligence (AI) is intelligence demonstrated by machines, unlike the natural intelligence displayed by humans and animals, which involves consciousness and emotionality. The distinction between the former and the latter categories is often revealed by the acronym chosen.
founded 4 years ago
you are viewing a single comment's thread
view the rest of the comments
view the rest of the comments
I only have a rudimentary understanding of LLMs, so can someone with more knowledge answer me some questions on this topic?
I've heard of data poisoning, which to my understanding means that one can manipulate/bias these models through the training data. Is this a potential problem with this model beyond the obvious censorship that seems to happen in the online version, but apparently can be circumvented? I'm asking because that seems to be fairly obvious, but minor biases might be hard to impossible to detect.
Also is the data it was trained on available as well at all? Or is it just the techniques on how it was trained and the resulting weights? Because without the former i'd imagine it would be impossible to verify any subtle manipulation in the training data or even just its selection.
There is no evidence that poisoning has had any effect on LLMs. It’s likely that it never will, because garbage inputs aren’t likely to get reinforced during training. It’s all just wishful thinking from the haters.
Every AI will always have bias, just as every person has bias, because humanity has never agreed on what is “truth”.
But it is possible, right? As an example from my quick search for example here a paper about Medical large language models
It's probably hard to change major things, like e.g. that Trump is the president of the USA, without it being extremely obvious or degrading performance massively. But smaller random facts? Like for example i have little to no online presence under my real name. So i'd imagine it shouldn't be to hard to add some documents to the training data with made up facts about me. It wouldn't be noticeable until someone actively looks for it and then they'd need to know the truth beforehand to judge them or at least require sources.
That's true, but since we are in a way actively condensing knowledge with LLMs i think there is a difference, if someone has the ability to influence things at this step without it being noticeable.