DNA: Uncovering Universal Latent Forgery Knowledge

TL;DR

This paper introduces the DNA framework that uncovers intrinsic forgery detection capabilities within pre-trained models, enabling effective detection without extensive retraining, and introduces a new high-fidelity synthetic benchmark for evaluation.

Contribution

The paper presents a novel method to extract forgery-sensitive units from pre-trained models and introduces HIFI-Gen, a new benchmark for evaluating forgery detection.

Findings

DNA outperforms fine-tuning methods in few-shot scenarios.

The approach is robust across different architectures.

It effectively detects forgeries from unseen generative models.

Abstract

As generative AI achieves hyper-realism, superficial artifact detection has become obsolete. While prevailing methods rely on resource-intensive fine-tuning of black-box backbones, we propose that forgery detection capability is already encoded within pre-trained models rather than requiring end-to-end retraining. To elicit this intrinsic capability, we propose the discriminative neural anchors (DNA) framework, which employs a coarse-to-fine excavation mechanism. First, by analyzing feature decoupling and attention distribution shifts, we pinpoint critical intermediate layers where the focus of the model logically transitions from global semantics to local anomalies. Subsequently, we introduce a triadic fusion scoring metric paired with a curvature-truncation strategy to strip away semantic redundancy, precisely isolating the forgery-discriminative units (FDUs) inherently imprinted with sensitivity to forgery traces. Moreover, we introduce HIFI-Gen, a high-fidelity synthetic benchmark built upon the very latest models, to address the lag in existing datasets. Experiments demonstrate that by solely relying on these anchors, DNA achieves superior detection performance even under few-shot conditions. Furthermore, it exhibits remarkable robustness across diverse architectures and against unseen generative models, validating that waking up latent neurons is more effective than extensive fine-tuning.