The paper is about implementing transformer-like attention using simulated human neurones