251–260 of 265 posts
Earlier quoted context omitted.
Imagine telling a person something that goes contrary to their conditioning, to their beliefs. They would tend to dismiss it completely and choose not to take it into account, even while knowing it's true.
This is widely visible as a very real thing and why charlatans are able to be charlatans. Not sure why it is phrased as something that needs to be imagined
Earlier quoted context omitted.
Okay, but the anecdote states that every model repeated the pseudo-factoid about Foobar square, not just the 4 GB open source model equivalent of a tabloid.
I think the key phrase here is, "an obscure small town." There may only be a single mention of this place, hence the only one on which a response can be based. This says more about the user's understanding of LLMs than it does about LLMs.
"Tell me everything you know about (obscure small town), (state). Only what's unique to (town), not commonly-known facts" is an excellent way to test for hallucinatory tendencies in a new model, in my experience. Likely the best I've found.
Quality of results is almost linearly proportional to the size of the model in many cases. The largest models like K3 and GLM 5.3 will either confine their responses to known true facts about the town and its surroundings, or admit they don't have enough information to answer. Smaller ones will reliably make up hilarious or downright-strange things.
Another good test is https://whatever.scalzi.com/2025/12/13/ai-a-dedicated-fact-f... , which still works on the newest models. Of the open-weight models available, only Kimi K3 will consistently admit it has no idea who Scalzi's novel is dedicated to. The rest still make up random stuff and present it confidently.
TL,DR: progress is possible, and it has been made, but it's happening slower than many people think.
Earlier quoted context omitted.
Take a random essay and add in a bunch of the phrases that LLMs love like “load-bearing,” “crucial,” structural,” and “woven,” and then submit the original and the edited version to an LLM and ask which is better. It will choose the second one virtually every time. They have ingrained biases that associate those words with good writing and arguments.
Sometimes I wonder if there’s just one guy somewhere who loved using the word load-bearing, all his papers got trained on, and now he can’t write anything without being assumed to be Claude.
Earlier quoted context omitted.
> Let's hope the LLM model continues to be paying for credits LLM vendors make this hard because you can't trust them with your session data. Yesterday you were opted out of training, then suddenly today you're opted in. It's an extension of the idea that they don't need to care about anybody's copyright. They don't care about preserving the security or privacy of customer data, because there is negligible incentive…
But don't you want Technojesus to come save you from never having learned how to invert a binary tree?
Human learning is slow, AI production is fast. Very hard to steer a middle path.
Earlier quoted context omitted.
Newspapers have been doing this for a long time, notably the Metro in London.
London had a non existant restautant as #1 recommendation but I think it was done by using multiple (sock puppet?) accounts to create fake reviews https://en.wikipedia.org/wiki/The_Shed_at_Dulwich
Earlier quoted context omitted.
> It isn't even a settled question that there is a genocide. Israel themselves seem pretty settled on it: https://www.france24.com/en/middle-east/20260902-israel-orga...
That's election blathering, it's less than two months and as the Israeli (far-)right is looking for a serious defeat [1], they are trying to mobilize their respective bases. Especially the ultraorthodox crowd is going to get hit hard, the rest of Israel is extremely pissed off at them due to their insistence on not having to serve in the military. And let me be clear: all of those currently in power in Israel deserve…
Your degree if denialism is really impressive. You dismiss, the mass killings, the push to erase a nation, and you even dismiss representatives outright admitting and acknowledging doing it.
Where exactly do you place the bar that would lead you to say "Yes, Israel is committing genocide".
Earlier quoted context omitted.
I think the key phrase here is, "an obscure small town." There may only be a single mention of this place, hence the only one on which a response can be based. This says more about the user's understanding of LLMs than it does about LLMs.
This says more about the user's understanding of LLMs than it does about LLMs. "Tell me everything you know about (obscure small town), (state). Only what's unique to (town), not commonly-known facts" is an excellent way to test for hallucinatory tendencies in a new model, in my experience. Likely the best I've found. Quality of results is almost linearly proportional to the size of the model in many cases. The large…
Earlier quoted context omitted.
I noticed this when I tried to get LLMs to play text adventures. Early on in my experiments, I wanted to give them hints when they got stuck at a puzzle. I did this by stopping the loop, injecting thoughts into the LLMs own persistent scratchpad (as if it had thought of that itself), and then starting the loop up again.[1] That way, the LLM would read what it had intended to remember from the previous turn including…
Imagine telling a person something that goes contrary to their conditioning, to their beliefs. They would tend to dismiss it completely and choose not to take it into account, even while knowing it's true.
Earlier quoted context omitted.
In ML, "best" is what minimises the loss function, so in a certain (very basic) sense it does.
You are mixing up levels. When you are asking it a question (like which of these two texts is the best), the output is also just picked by minimising that loss function. There's no guarantee that answering "Text B is better" aligns with text B minimising the loss function. (And they aren't really minimising loss functions during inference. They sample from a distribution. During training they minimise the loss functi…
So to OPs question: I guess LLMs do have an "idea" of what is best (conditioned on minimising a loss function during training), however they may not always output that (because stochasticity), which maybe represents a degree of uncertainty in that "idea"?
Earlier quoted context omitted.
That's election blathering, it's less than two months and as the Israeli (far-)right is looking for a serious defeat [1], they are trying to mobilize their respective bases. Especially the ultraorthodox crowd is going to get hit hard, the rest of Israel is extremely pissed off at them due to their insistence on not having to serve in the military. And let me be clear: all of those currently in power in Israel deserve…
> That's election blathering Your degree if denialism is really impressive. You dismiss, the mass killings, the push to erase a nation, and you even dismiss representatives outright admitting and acknowledging doing it. Where exactly do you place the bar that would lead you to say "Yes, Israel is committing genocide".
At the very least you’d expect the population being genocided to decline.
Palestinian population growth has been exponential throughout Israel’s history (if you’re one of those who claim the “genocide” started a long time ago), and even the population of Gaza has not declined after 3 years of war (reported births are still more than reported deaths).