“I don’t have to drive a symbol of racism, greed and ignorance! Life is suddenly much better!”
—Tor Bette Middler expresses his joy when selling his tesla, Insider information.
The great story
Large language models can do amazing things. But nobody knows exactly why.
Two years ago, Yuri Burda and Harri Edwards, Operai researchers, were trying to find out what would be needed to obtain a large language model to make a basic arithmetic. At first, things did not go too well. The models memorized the sums they saw but could not solve new ones.
By accident, Burda and Edwards left some of their experiments running for days instead of hours. The models were shown the sums of example again and again, and finally learned to add two numbers; He had just taken much longer than someone thought he should.
In certain cases, the models could apparently fail to learn a task and, suddenly, he simply gets it, as if a light bulb was lit, a behavior that the researchers called Grokking. Grokking is only one of several strange phenomena who have researchers from the head. Larger models and large language models in particular seem to behave so that textbook mathematics say they should not.
This highlights a remarkable fact about deep learning, the fundamental technology behind Ia’s boom today: despite all its fugitive success, nobody knows exactly how, or why it works. Read the complete story.
—It will be the sky of Douglas
We can still have beautiful things
A place for comfort, fun and distraction to brighten your day. (Do you have any idea? Leave me a line either Be up to me.
+ What happened when WHAM! He took Western pop music to China 40 years ago.
+ Who knew that? Sharks make noises after all? 🦈
+ Microsoft is celebrating its 50th anniversary, and with it, 50 very strange inventions.
+ What devils is a prototaxiteanyway?
