Inspect what an internal language-model activation is preparing the model to say.

Soon you can unlock how this was done.

Behind this: the tool used · the method · what actually resulted.