It was in the announcement, too. I’m 99% sure they edited it after they changed their mind, because I knew about it from reading that, and never opened the model card.
On the earliest web archive snapshot I can find [0], I do not see any mention of the safeguard/sabotage under discussion [1].
And to be clear, this isn't the safeguard where the model is explicitly downgraded to Opus, but rather where the Fable/Mythos model's "effectiveness" is transparently "limited" via "prompt modification, steering vectors, or parameter-efficient fine-tuning (PEFT)".
They didn’t get caught, they explicitly said they would do that in the announcement. I think it was both bad and a weird idea, but it certainly wasn’t sneaky.
Tons of people have died for a variety of causes throughout human history. It does not mean that the causes were not silly or deluded. (e.g religious wars, Communism, Fascism etc.)
Carlin didn’t say it was silly, he said it didn’t matter. Religion, communism, fascism were all world-defining ideas, and they very much mattered to those who fought for them and against them. What am I saying, “mattered”. Go check social media, they matter right now.
The United States predates you, and will almost certainly outlive you.
You, an individual with a microscopic lifespan and negligible influence in the grand sweep of history, do not in fact share any guilt for the things your ancestors did, nor do you deserve credit for the greatness they wrought. The United States, however, deserves blame where guilty and praise where worthy for what it did during its lifespan, which expands hundreds of years into the past, and with luck will continue into the future.
There's a reason stochastic was used in the original phrase instead of "probabilistic."
While most inference executions are intentionally non-deterministic, even a purely deterministic one would still be stochastic in that the model itself was built in a process such that the statistical frequency, sequencing, etc of the training text and followup processes all heavily influence the result.
Because of that, the output is the sort of thing that is not expected to generate 100% perfect output 100% of the time, but to have a good probability of being like-in-kind-to-the-training-data (and useful/relevant as a result).
(As compared to a non-stochastic model, like arithmetic on integers, where 2+2 is always gonna be 4 and you don't have a chance of coming up with some novel pair of inputs to addition that will cause your arithmetic to miss the mark.)
Agreed. My point was to question the use of “just“ to obscure an incredibly complicated process, which has been shown repeatedly to rely on generalizations that are indistinguishable from world models.
Now, it is true that the world they’re modeling is the world of tokens. But insofar as those tokens, be they text or images or videos, are themselves modeling the real world, LLMs do have a model of the real world.
reply