Automated interpretability research has recently attracted attention as a potential research direction that could scale explanations of neural network behavior to large models. Existing automated circuit discovery work applies activation patching to identify subnetworks responsible for solving specific tasks (circuits). In this work, we show that a simple method based on attribution patching outperforms all existing methods while requiring just two forward passes and a backward pass. We apply a linear approximation to activation patching to estimate the importance of each edge in the computational subgraph. Using this approximation, we prune the least important edges of the network. We survey the performance and limitations of this method, finding that averaged over all tasks our method has greater AUC from circuit recovery than other methods.

翻译：自动化可解释性研究近期作为有望将神经网络行为解释扩展到大型模型的潜在研究方向引起了关注。现有的自动化电路发现工作应用激活补丁来识别负责解决特定任务的子网络（电路）。在本研究中，我们证明了一种基于归因补丁的简单方法在仅需两次前向传播和一次反向传播的情况下优于所有现有方法。我们采用激活补丁的线性近似来估计计算子图中每条边的重要性，并利用这一近似对网络中重要性最低的边进行剪枝。通过评估该方法的性能与局限性，我们发现在所有任务上的平均电路恢复AUC值均优于其他方法。