Exploring Concept Depth: How Large Language Models Acquire Knowledge at Different Layers?

This paper studies the phenomenon that different concepts are learned in different layers of large language models, i.e. more difficult concepts are fully acquired with deeper layers. We define the difficulty of concepts by the level of abstraction, and here it is crudely categorized by factual, emotional, and inferential. Each category contains a spectrum of tasks, arranged from simple to complex. For example, within the factual dimension, tasks range from lie detection to categorizing mathematical problems. We employ a probing technique to extract representations from different layers of the model and apply these to classification tasks. Our findings reveal that models tend to efficiently classify simpler tasks, indicating that these concepts are learned in shallower layers. Conversely, more complex tasks may only be discernible at deeper layers, if at all. This paper explores the implications of these findings for our understanding of model learning processes and internal representations. Our implementation is available at \url{https://github.com/Luckfort/CD}.

Finding needles in a haystack: A Black-Box Approach to Invisible Watermark Detection

In this paper, we propose WaterMark Detection (WMD), the first invisible watermark detection method under a black-box and annotation-free setting. WMD is capable of detecting arbitrary watermarks within a given reference dataset using a clean non-watermarked dataset as a reference, without relying on specific decoding methods or prior knowledge of the watermarking techniques. We develop WMD using foundations of offset learning, where a clean non-watermarked dataset enables us to isolate the influence of only watermarked samples in the reference dataset. Our comprehensive evaluations demonstrate the effectiveness of WMD, significantly outperforming naive detection methods, which only yield AUC scores around 0.5. In contrast, WMD consistently achieves impressive detection AUC scores, surpassing 0.9 in most single-watermark datasets and exceeding 0.7 in more challenging multi-watermark scenarios across diverse datasets and watermarking methods. As invisible watermarks become increasingly prevalent, while specific decoding techniques remain undisclosed, our approach provides a versatile solution and establishes a path toward increasing accountability, transparency, and trust in our digital visual content.

How to Detect Unauthorized Data Usages in Text-to-image Diffusion Models

Recent text-to-image diffusion models have shown surprising performance in generating high-quality images. However, concerns have arisen regarding the unauthorized usage of data during the training process. One example is when a model trainer collects a set of images created by a particular artist and attempts to train a model capable of generating similar images without obtaining permission from the artist. To address this issue, it becomes crucial to detect unauthorized data usage. In this paper, we propose a method for detecting such unauthorized data usage by planting injected memorization into the text-to-image diffusion models trained on the protected dataset. Specifically, we modify the protected image dataset by adding unique contents on the images such as stealthy image wrapping functions that are imperceptible to human vision but can be captured and memorized by diffusion models. By analyzing whether the model has memorization for the injected content (i.e., whether the generated images are processed by the chosen post-processing function), we can detect models that had illegally utilized the unauthorized data. Our experiments conducted on Stable Diffusion and LoRA model demonstrate the effectiveness of the proposed method in detecting unauthorized data usages.

Alteration-free and Model-agnostic Origin Attribution of Generated Images

Recently, there has been a growing attention in image generation models. However, concerns have emerged regarding potential misuse and intellectual property (IP) infringement associated with these models. Therefore, it is necessary to analyze the origin of images by inferring if a specific image was generated by a particular model, i.e., origin attribution. Existing methods are limited in their applicability to specific types of generative models and require additional steps during training or generation. This restricts their use with pre-trained models that lack these specific operations and may compromise the quality of image generation. To overcome this problem, we first develop an alteration-free and model-agnostic origin attribution method via input reverse-engineering on image generation models, i.e., inverting the input of a particular model for a specific image. Given a particular model, we first analyze the differences in the hardness of reverse-engineering tasks for the generated images of the given model and other images. Based on our analysis, we propose a method that utilizes the reconstruction loss of reverse-engineering to infer the origin. Our proposed method effectively distinguishes between generated images from a specific generative model and other images, including those generated by different models and real images.

NOTABLE: Transferable Backdoor Attacks Against Prompt-based NLP Models

Prompt-based learning is vulnerable to backdoor attacks. Existing backdoor attacks against prompt-based models consider injecting backdoors into the entire embedding layers or word embedding vectors. Such attacks can be easily affected by retraining on downstream tasks and with different prompting strategies, limiting the transferability of backdoor attacks. In this work, we propose transferable backdoor attacks against prompt-based models, called NOTABLE, which is independent of downstream tasks and prompting strategies. Specifically, NOTABLE injects backdoors into the encoders of PLMs by utilizing an adaptive verbalizer to bind triggers to specific words (i.e., anchors). It activates the backdoor by pasting input with triggers to reach adversary-desired anchors, achieving independence from downstream tasks and prompting strategies. We conduct experiments on six NLP tasks, three popular models, and three prompting strategies. Empirical results show that NOTABLE achieves superior attack performance (i.e., attack success rate over 90% on all the datasets), and outperforms two state-of-the-art baselines. Evaluations on three defenses show the robustness of NOTABLE. Our code can be found at https://github.com/RU-System-Software-and-Security/Notable.

UNICORN: A Unified Backdoor Trigger Inversion Framework

The backdoor attack, where the adversary uses inputs stamped with triggers (e.g., a patch) to activate pre-planted malicious behaviors, is a severe threat to Deep Neural Network (DNN) models. Trigger inversion is an effective way of identifying backdoor models and understanding embedded adversarial behaviors. A challenge of trigger inversion is that there are many ways of constructing the trigger. Existing methods cannot generalize to various types of triggers by making certain assumptions or attack-specific constraints. The fundamental reason is that existing work does not consider the trigger's design space in their formulation of the inversion problem. This work formally defines and analyzes the triggers injected in different spaces and the inversion problem. Then, it proposes a unified framework to invert backdoor triggers based on the formalization of triggers and the identified inner behaviors of backdoor models from our analysis. Our prototype UNICORN is general and effective in inverting backdoor triggers in DNNs. The code can be found at https://github.com/RU-System-Software-and-Security/UNICORN.

Backdoor Vulnerabilities in Normally Trained Deep Learning Models

We conduct a systematic study of backdoor vulnerabilities in normally trained Deep Learning models. They are as dangerous as backdoors injected by data poisoning because both can be equally exploited. We leverage 20 different types of injected backdoor attacks in the literature as the guidance and study their correspondences in normally trained models, which we call natural backdoor vulnerabilities. We find that natural backdoors are widely existing, with most injected backdoor attacks having natural correspondences. We categorize these natural backdoors and propose a general detection framework. It finds 315 natural backdoors in the 56 normally trained models downloaded from the Internet, covering all the different categories, while existing scanners designed for injected backdoors can at most detect 65 backdoors. We also study the root causes and defense of natural backdoors.

Rethinking the Reverse-engineering of Trojan Triggers

Deep Neural Networks are vulnerable to Trojan (or backdoor) attacks. Reverse-engineering methods can reconstruct the trigger and thus identify affected models. Existing reverse-engineering methods only consider input space constraints, e.g., trigger size in the input space. Expressly, they assume the triggers are static patterns in the input space and fail to detect models with feature space triggers such as image style transformations. We observe that both input-space and feature-space Trojans are associated with feature space hyperplanes. Based on this observation, we design a novel reverse-engineering method that exploits the feature space constraint to reverse-engineer Trojan triggers. Results on four datasets and seven different attacks demonstrate that our solution effectively defends both input-space and feature-space Trojans. It outperforms state-of-the-art reverse-engineering methods and other types of defenses in both Trojaned model detection and mitigation tasks. On average, the detection accuracy of our method is 93\%. For Trojan mitigation, our method can reduce the ASR (attack success rate) to only 0.26\% with the BA (benign accuracy) remaining nearly unchanged. Our code can be found at https://github.com/RU-System-Software-and-Security/FeatureRE.

BppAttack: Stealthy and Efficient Trojan Attacks against Deep Neural Networks via Image Quantization and Contrastive Adversarial Learning

Deep neural networks are vulnerable to Trojan attacks. Existing attacks use visible patterns (e.g., a patch or image transformations) as triggers, which are vulnerable to human inspection. In this paper, we propose stealthy and efficient Trojan attacks, BppAttack. Based on existing biology literature on human visual systems, we propose to use image quantization and dithering as the Trojan trigger, making imperceptible changes. It is a stealthy and efficient attack without training auxiliary models. Due to the small changes made to images, it is hard to inject such triggers during training. To alleviate this problem, we propose a contrastive learning based approach that leverages adversarial attacks to generate negative sample pairs so that the learned trigger is precise and accurate. The proposed method achieves high attack success rates on four benchmark datasets, including MNIST, CIFAR-10, GTSRB, and CelebA. It also effectively bypasses existing Trojan defenses and human inspection. Our code can be found in https://github.com/RU-System-Software-and-Security/BppAttack.

Neural Network Trojans Analysis and Mitigation from the Input Domain

Deep Neural Networks (DNNs) can learn Trojans (or backdoors) from benign or poisoned data, which raises security concerns of using them. By exploiting such Trojans, the adversary can add a fixed input space perturbation to any given input to mislead the model predicting certain outputs (i.e., target labels). In this paper, we analyze such input space Trojans in DNNs, and propose a theory to explain the relationship of a model's decision regions and Trojans: a complete and accurate Trojan corresponds to a hyperplane decision region in the input domain. We provide a formal proof of this theory, and provide empirical evidence to support the theory and its relaxations. Based on our analysis, we design a novel training method that removes Trojans during training even on poisoned datasets, and evaluate our prototype on five datasets and five different attacks. Results show that our method outperforms existing solutions. Code: \url{https://anonymous.4open.science/r/NOLE-84C3}.