> that they can internally represent many different complex ideas efficiently and coherently
The Transformer circuits[0] suggest that this representation is not coherent at all.
The Transformer circuits[0] suggest that this representation is not coherent at all.
It's doing compression which does not mean it's coherent.
> The addition circuits are also fairly easy to interpret.
The addition circuits make no sense whatsoever. It's doing great at guessing that's all.