|
|
|
|
|
|
by accrual
11 days ago
|
|
|
I thought the article was interesting from a visualization perspective, even if it's not perfectly solid on the actual mechanics in a model. It helps describe the scale of the connections. > There’s no pre-formed thought “behind” the words that then gets translated into language. The words are the thinking. The recent paper on "J-space" [0] contradicts this, models can "think" of things without emitting as text. [0] https://www.anthropic.com/research/global-workspace |
|
|