Using secondary rewriting to eliminate stylistic features introduces new human intervention that may affect evaluation fairness, which is one of the core contradictions at the engineering level of double-blind AI evaluation that has yet to be resolved.