Adversarial attacks pose security concerns to deep learning applications, but their characteristics are under-explored. Yet largely imperceptible, a strong trace could have been left by PGD-like attacks in an adversarial example. Recall that PGD-like attacks trigger the “local linearity” of a network, which implies different extents of linearity for benign or adversarial examples. Inspired by this, we construct an Adversarial Response Characteristics (ARC) feature to reflect the model’s gradient consistency around the input to indicate the extent of linearity. Under certain conditions, it qualitatively shows a gradually varying pattern from benign example to adversarial example, as the latter leads to Sequel Attack Effect (SAE). To quantitatively evaluate the effectiveness of ARC, we conduct experiments on CIFAR-10 and ImageNet in a challenging setting. The results suggest that SAE, reflected through the ARC feature, is an effective and unique trace of PGD-like attacks. Our method is designed to generalize with a scarce amount of data, which remains feasible even when access to the full training dataset is impossible. Code: https://github.com/cdluminate/advtrace .