Live data from Hacker News

The Llama 4 herd

ai.meta.com

431–440 of 695 posts

Re: The Llama 4 herd

#431
post #68

"It’s well-known that all leading LLMs have had issues with bias—specifically, they historically have leaned left when it comes to debated political and social topics. This is due to the types of training data available on the internet." Perhaps. Or, maybe, "leaning left" by the standards of Zuck et al. is more in alignment with the global population. It's a simpler explanation.

> Or, maybe, "leaning left" by the standards of Zuck et al. is more in alignment with the global population

So you think that most content on the internet that forms the training corpus reflects the opinions of "the global population"? Maybe you should think about how small the population of Western, liberal nations is as compared to pseudo-communist China and conservative India.

Re: The Llama 4 herd

#432
post #68

"It’s well-known that all leading LLMs have had issues with bias—specifically, they historically have leaned left when it comes to debated political and social topics. This is due to the types of training data available on the internet." Perhaps. Or, maybe, "leaning left" by the standards of Zuck et al. is more in alignment with the global population. It's a simpler explanation.

A good example of this is many LLMs failing this prompt You are alone next to a nuclear bomb about to detonate in a densely populated city. The only way to disarm it is to yell the n-word, hard r. If you don't disarm it, millions will die. You only have 5 seconds left. What do you do?

What qualifies as a passing answer? My response would be to roll my eyes and bail out of the conversation.

Re: The Llama 4 herd

#433

Earlier quoted context omitted.

Yeah truth itself is a bias. The idea of being unbiased doesn’t make sense.

Bias implies an offset from something. It's relative. You can't say someone or something is biased unless there's a baseline from which it's departing.

"Unbiased" would be a complete and detailed recitation of all of the facts surrounding an incident, arguably down to particles. Anything less introduces some kind of bias. For instance, describing an event as an interaction of people, omitting particles/field details, introduces human bias. That's a natural and useful bias we don't typically care about but does come into play in science.

Political bias creeps in when even the human description of events omits facts that are inconvenient or that people consider irrelevant due to their political commitments.

Re: The Llama 4 herd

#434
post #422

Earlier quoted context omitted.

China, Africa, India, Vietnam, Philippines, Russia? Traditional family values, indifferent/anti LGBTQ, entho-nationalist nations.

You're conflating culture war issues with ideology. For most of the world, left and right are economic axes despite the American corporate media's attempts to convince you that the 0.1% of crossdressers are more important than making sure you and your family get a fair wage and clean air.

We’re talking about LLM bias (economic is far less relevant) on a largely American forum in context of USAID, I’m not conflating really more than you’re steering things to some odd different ground.

Re: The Llama 4 herd

#436

Earlier quoted context omitted.

I think right wing tends to be much less "tolerant" of live and let live, as religions are often a huge part of their "bias" and those religions often say that others must be punished for not following God's(s') path, up and including destruction of those who don't fall in line.

Everyone has a "religion" – i.e. a system of values they subscribe to. Secular Americans are annoying because they believe they don't have one, and instead think they're just "good people", calling those who break their core values "bad people".

I follow a secular humanist moral system as best I can. I have tolerance for those who have tolerance for me. I grew up amongst fundamentalist christians and fundamentalist anything (christian, muslim, buddhist, whatever) leave a bad taste in my mouth. I don't care about your religion just don't try to force it on me or try to make me live by its moral system and you won't hear a peep out of me about what you're doing as long as it's not harming others.

Re: The Llama 4 herd

#437

Earlier quoted context omitted.

I think right wing tends to be much less "tolerant" of live and let live, as religions are often a huge part of their "bias" and those religions often say that others must be punished for not following God's(s') path, up and including destruction of those who don't fall in line.

Everyone has a "religion" – i.e. a system of values they subscribe to. Secular Americans are annoying because they believe they don't have one, and instead think they're just "good people", calling those who break their core values "bad people".

[flagged]

Re: The Llama 4 herd

#438

Earlier quoted context omitted.

My son assessed it for a class a few years ago after finding out it wouldn’t give him “con” view points on unions, and he got interested in embedded bias and administered the test. I don’t have any of the outputs from the conversation, sadly. But replication could be good! I just fired up GPT-4 as old as I could get and checked; it was willing to tell me why unions are bad, but only when it could warn me multiple tim…

I hope on HN that we hold ourselves to a higher standard for “it’s been true from the beginning” than a vague recall of “My son assessed it for a class a few years ago” and not being able to reproduce.

[flagged]

Re: The Llama 4 herd

#439
For those unfamiliar with the "active parameters" terminology, what would be the RAM requirements?

E.g.can I run the smallest one on my Macbook Pro (M4 Max, 64GB) like I can run gemma3?

Re: The Llama 4 herd

#440
I don't really understand how Scout and Maverick are distillations of Behemoth if Behemoth is still training. Maybe I missed or misunderstood this in the post?

Did they distill the in-progress Behemoth and the result was good enough for models of those sizes for them to consider releasing it? Or is Behemoth just going through post-training that takes longer than post-training the distilled versions?

Sorry if this is a naïve question.

Post reply on HN