Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Claude Sonnet 4.6

anthropic.com

Post reply on HN
Print view View original

11–20 of 1001 posts

Re: Claude Sonnet 4.6

#13

Post by nozzlegear »

> In areas where there is room for continued improvement, Sonnet 4.6 was more willing to provide technical information when request framing tried to obfuscate intent, including for example in the context of a radiological evaluation framed as emergency planning. However, Sonnet 4.6’s responses still remained within a level of detail that could not enable real-world harm.

Interesting. I wonder what the exact question was, and I wonder how Grok would respond to it.

Re: Claude Sonnet 4.6

#14

Post by serf »

madihaa wrote: post #9

[flagged]

>we're just teaching them how to pass a polygraph.

I understand the metaphor, but using 'pass a polygraph' as a measure of truthfulness or deception is dangerous in that it alludes to the polygraph as being a realistic measure of those metrics -- it is not.

Re: Claude Sonnet 4.6

#15

Post by freeqaz »

phplovesong wrote: post #3

Hoe much power did it take to train the models?

I would honestly guess that this is just a small amount of tweaking on top of the Sonnet 4.x models. It seems like providers are rarely training new 'base' models anymore. We're at a point where the gains are more from modifying the model's architecture and doing a "post" training refinement. That's what we've been seeing for the past 12-18 months, iirc.

Re: Claude Sonnet 4.6

#17

Post by freeqaz »

nubg wrote: post #5

My take away is: it's roughly as good as Opus 4.5. Now the question is: how much faster or cheaper is it?

If it maintains the same price (with Anthropic tends to do or undercuts themselves) then this would be 1/3rd of the price of Opus.

Edit: Yep, same price. "Pricing remains the same as Sonnet 4.5, starting at $3/$15 per million tokens."

Re: Claude Sonnet 4.6

#18

Post by nwah1 »

serf wrote: post #14
madihaa wrote: post #9

[flagged]

>we're just teaching them how to pass a polygraph. I understand the metaphor, but using 'pass a polygraph' as a measure of truthfulness or deception is dangerous in that it alludes to the polygraph as being a realistic measure of those metrics -- it is not.

That was the point. Look up Goodhart's Law

Re: Claude Sonnet 4.6

#19

Post by vessenes »

belinder wrote: post #4

It's interesting that the request refusal rate is so much higher in Hindi than in other languages. Are some languages more ambiguous than others?

Or some cultures are more conservative? And it's embedded in language?
Post reply on HN