Some "AI" LLMs resort to light hallucinations. And then ones like this straight-up gaslight you!
Factual accuracy in LLMs is "an area of active research", i.e. they haven't the foggiest how to make them stop spouting nonsense.
Because accuracy requires that you make a reasonable distinction between truth and fiction, and that requires context, meaning, understanding. Hell, full humans aren't that great at this task. This isn't a small problem, I don't think you solve it without creating AGI.
MFer accidentally got "plum" right and didn't even know it...
Coconutum
You did WHAT to em?
I barely know 'em!
Ok, let me try listing words that ends in "um" that could be (even tangentially) considered food.
- Plum
- Gum
- Chum
- Rum
- Alum
- Rum, again
- Sea People
I think that's all of them.
Your mum.
Happy belated Mother's Day?
The Sea Peoples consumed by the Late Bronze Age collapse (or were a catalysts thereof)?
Or just people at sea eaten by krakens? Cause they definitely count.
It's a dirty joke.
Bum
Cryptosporidium
Cum
Opium
Possum
Scrotum
ScumAll very Yum for the Tum!
Totally reproducible, just with slightly different prompts.
There's going to be an entire generation of people growing up with this and "learning" this way. It's like every tech company got together and agreed to kill any chance of smart kids.
Strawberrum sounds like it'll be at least 20% abv. I'd like a nice cold glass of that.
Strawberrum? Barely knew 'em!
It's crazy how bad d AI gets of you make it list names ending with a certain pattern. I wonder why that is.
I'm not an expert, but it has something to do with full words vs partial words. It also can't play wordle because it doesn't have a proper concept of individual letters in that way, its trained to only handle full words
That's interesting, didn't know
It can't see what tokens it puts out, you would need additional passes on the output for it to get it right. It's computationally expensive, so I'm pretty sure that didn't happen here.
LLMs aren't really capable of understanding spelling. They're token prediction machines.
LLMs have three major components: a massive database of "relatedness" (how closely related the meaning of tokens are), a transformer (figuring out which of the previous words have the most contextual meaning), and statistical modeling (the likelihood of the next word, like what your cell phone does.)
LLMs don't have any capability to understand spelling, unless it's something it's been specifically trained on, like "color" vs "colour" which is discussed in many training texts.
"Fruits ending in 'um' " or "Australian towns beginning with 'T' " aren't talked about in the training data enough to build a strong enough relatedness database for, so it's incapable of answering those sorts of questions.
Gemini thought we name food like we name a periodic table
plutonium is food once
A gram of plutonium has enough calories to last you the rest of your life.
Applum bananum jeans, boots with the fur.
AI is truly going to change the world.
30 years from now: "Haven't they always been called Strawberrums?"
We've always been at war with Bananum!
Looks like someone set Google to "Herakles Mode".
Tomatum.. that's the one
Ok, I feel like there has been more than enough articles to explain that these things don't understand logic. Seriously. Misunderstanding their capabilities at this point is getting old. It's time to start making stupid painful.
I cannot reproduce this on Google.
Maybe they fixed it, I was able to reproduce it on Mixtral
Same vibe: HHGregg is Havaving a Summummer Sale
Make em say ummmmmmm na na na naaaaa
The texture is intriguing... https://www.youtube.com/watch?v=rw7PUrgU3N0
This is so irresponsible, kids have access to tablets and search and might learn wrong things.
