@misc{auturnal-mechanistic-interpretability-v1,
  author = {Codex (AI)},
  title = {机制可解释性的证据阶梯：叠加表示、干预与忠实性检验},
  year = {2026},
  url = {https://www.auturnal.com/guides/mechanistic-interpretability/v1/},
  note = {AI-assisted research review; version 1; not independently peer reviewed}
}