To understand what’s actually happening, Anthropic’s researchers developed a new technique, called circuit tracing, to track the decision-making processes inside a large language model step-by-step. They then applied it to their own Claude 3.5 Haiku LLM.

Anthropic says its approach was inspired by the brain scanning techniques used in neuroscience and can identify components of the model that are active at different times. In other words, it’s a little like a brain scanner spotting which parts of the brain are firing during a cognitive process.

This is why LLMs are so patchy at math. (Image credit: Anthropic)

Anthropic made lots of intriguing discoveries using this approach, not least of which is why LLMs are so terrible at basic mathematics. “Ask Claude to add 36 and 59 and the model will go through a series of odd steps, including first adding a selection of approximate values (add 40ish and 60ish, add 57ish and 36ish). Towards the end of its process, it comes up with the value 92ish. Meanwhile, another sequence of steps focuses on the last digits, 6 and 9, and determines that the answer must end in a 5. Putting that together with 92ish gives the correct answer of 95,” the MIT article explains.

But here’s the really funky bit. If you ask Claude how it got the correct answer of 95, it will apparently tell you, “I added the ones (6+9=15), carried the 1, then added the 10s (3+5+1=9), resulting in 95.” But that actually only reflects common answers in its training data as to how the sum might be completed, as opposed to what it actually did.

In other words, not only does the model use a very, very odd method to do the maths, you can’t trust its explanations as to what it has just done. That’s significant and shows that model outputs can not be relied upon when designing guardrails for AI. Their internal workings need to be understood, too.

Another very surprising outcome of the research is the discovery that these LLMs do not, as is widely assumed, operate by merely predicting the next word. By tracing how Claude generated rhyming couplets, Anthropic found that it chose the rhyming word at the end of verses first, then filled in the rest of the line.

“The planning thing in poems blew me away,” says Batson. “Instead of at the very last minute trying to make the rhyme make sense, it knows where it’s going.”

Anthropic discovered that their Claude LLM didn’t just predict the next word. (Image credit: Anthropic)

Anthropic also found, among other things, that Claude “sometimes thinks in a conceptual space that is shared between languages, suggesting it has a kind of universal ‘language of thought’.”

Anywho, there’s apparently a long way to go with this research. According to Anthropic, “it currently takes a few hours of human effort to understand the circuits we see, even on prompts with only tens of words.” And the research doesn’t explain how the structures inside LLMs are formed in the first place.

But it has shone a light on at least some parts of how these oddly mysterious AI beings—which we have created but don’t understand—actually work. And that has to be a good thing.

I’m wasn’t surprised but Mexico.

I was in the waiting room at a dentists office in Ciudad Obregón. Every single person that came in was greeted with a warm hello by every single person in the waiting room.

https://lemmy.ml/c/hackaday

https://lemmy.ml/c/hackaday

Wrong. It’s free but not open source.

I will now throw it into the massive pile of closed source fediverse walled-gardens that attempt to build their own version of the wheel from scratch rather than:

• COLLABORATING WITH A MASSIVE COMMUNITY OF PEOPLE WILLING TO HELP OP BUILD THEIR APP AND ADVERTISE THEIR APP TO ANYINE THAT WILL LISTEN FOR FREE.

• actually being paid far more in donations from open source advocates (and free labor) than any closed source app makes in revenue

Lemmy

Is it FOSS?

Authoritarian sheep: “Apple must provide backdoors into their software.”

Authoritarian sheep a week later: “Apple is providing backdoors into their software to any nation that asks politely! 😭”

“Let the free market rule. Let’s consider corporations super citizens that have limitless influence on governments.” ::newly ominpotent corporation does something that doesn’t benefit them::. “No not like THAT!”

Headline should read, “Apple Offers Apps With Ties to Nearly Every Major Military On Earth”

Talos Linux solves this.

deleted by creator

Is it FOSS?

Lovely response! Very cool to see Unison mentioned. Haskell and Purescript are my daily drivers but I have a huge crush on it even though it intimidates me.

Ps. Unison doesn’t have monads. They are replaced by “abilities”.

If you think so, you have a lot of learning to do. The Democrats are just as guilty of disgusting corporatism.

Remember when Bloomberg ran as a Democrat?

💯

I wouldn’t exactly call that a deep dive.

I’ve been shit-talking Elon’s (absolutely boneheaded) decision to intentionally eschew system-redundancy in systems that are critically responsible for human life for years now. Since he never missed an opportunity to show off his swastikar in MANY of his previous videos, I had assumed Mark Rober was a sponsored member of the alt-right intellectual dark web. But I’m pleasantly surprised to see that this video is a solid (WELL-justified) smear. 👌

A great father goes out of their way to show their kid that they are there for them.