Any thoughts on how helpful it might be to try to automate the manual inspection and evaluation (for task-relevancy for each feature in the circuit) part from section 4 in the paper, using e.g. a future version of MAIA (to reduce human costs / make the proposal more scalable)?
Any thoughts on how helpful it might be to try to automate the manual inspection and evaluation (for task-relevancy for each feature in the circuit) part from section 4 in the paper, using e.g. a future version of MAIA (to reduce human costs / make the proposal more scalable)?