Representational and sampling assumptions drive individual differences in single category generalisation.