Adding the benchmark back into the training process doesn't mean you get an LLM the can weed out irrelevant data, what you get is an LLM that can pass the new metric and you have to design a new metric with different semantic patterns to actually know if it's "eliminating red herrings".
zbyte64
If they know about this then they aren't thinking of the security implications
Misses says my T-shirt tan ruins the look, not the dad bod ❤️
Talk about making a difficult problem (self-driving) more difficult to solve by solving another hard problem.
Being Liberal is not the same as being an American Libertarian. The latter thinks tax is theft and has diverse views on whether children should be regarded as property under the law.
"Gold age of humanity"? Dude we are asking for dignity, not some fully automated luxury gay communism. I mean all that sounds nice, but it starts with valuing people over profits.
Grim, but technically accurate, to say nothing of whether such actions are right or wrong.
I'm stealing this.
The more you can make externalities someone else's problem, the more wealth you can acquire.
Dude is just wholesome.
Guy even gave the plot away when he claims the AI will be incomprehensible, like meeting aliens. So they aren't even pretending it will justify it's answers.
That's not how you sell fish though. You gotta emphasize how at one time we were all basically fish and if you buy my fish for long enough, those fish will eventually evolve hands to climb!