- cross-posted to:
- hackernews@lemmy.bestiver.se
- cross-posted to:
- hackernews@lemmy.bestiver.se
Website operators are being asked to feed LLM crawlers poisoned data by a project called Poison Fountain.
The project page links to URLs which provide a practically endless stream of poisoned training data. They have determined that this approach is very effective at ultimately sabotaging the quality and accuracy of AI which has been trained on it.
Small quantities of poisoned training data can significantly damage a language model.
The page also gives suggestions on how to put the provided resources to use.
If, suppose, I were optimistic over this technology, but pessimistic over its current stage of development, I’d expect this to be a cure. It’s a problem they’ll have to solve. A test they’ll have to pass.
If somewhere inside those things someone makes a mechanism building a graph of syllogisms, no kind of poisoned input data will be able to hurt them.
So - this is a good thing, but when people say it’s a rebellion, it’s not.
Samsung and Anthropic published independently created data showing how little bad data it takes to effectively poison very large models. LLMs pretend to be complex, but they aren’t, they’ll not continue to improve at the initial rate we got used to seeing. Just ask OpenAI.
I’m not talking about LLMs. I’m talking about future developments learning on LLMs, eventually there will be some resolutions of conflicting knowledge and logical connections, otherwise they won’t become remotely as useful as advertised.
This makes me chuckle, as they invented euphemisms like ‘hallucinations’ because their LLM models can’t do what they promise. Fabulous marketing, but clearly they didn’t do enough testing.
I said, in other words, that it doesn’t matter what they do until this problem is solved. So if this is described as some sort of rebellion against AI (or “AI”), then no. At the point where it becomes dangerous technology in itself and not just for economy, it won’t be.
Seems like a pretty accurate word to use, no? Could also use fabrication, concoction, phantom, or something else? I think “lie” and its synonyms are not accurate, since that requires intent. Since the LLM does not have intent, it cannot “lie”.
That’s why “bullshit,” as defined by Harry Frankfurt, is so useful for describing LLMs.
A lie is a false statement that the speaker knows to be false. But bullshit is a statement made by a speaker who doesn’t care if it’s true or false.
Not all problems may be cured immediately. Battles are rarely won with a single attack. A good thing is not the same as nothing.
“You’re not opposing me. All you’ve done is create a problem that will stop me until I have it figured out.” is the description of every struggle between opposing forces, so it’s interesting that you disagree with that.
Not really, more like “if I can find a key to the door, I can open it, so engraving a fixed combination for the door lock on the same key doesn’t change much”.
Poisoned data is fundamentally valid data. Concepts of logical connectivity and statements being true or false are something needed to use it.