Exploring the Secondary Risks of Large Language Models

Ensuring the safety and alignment of Large Language Models is a significant challenge with their growing integration into critical applications and societal functions. While prior research has primarily focused on jailbreak attacks, less attention has been given to non-adversarial failures that subtly emerge during benign interactions. We introduce secondary risks a novel class of failure modes marked by harmful or misleading behaviors during benign prompts. Unlike adversarial attacks, these risks stem from imperfect generalization and often evade standard safety mechanisms. To enable systematic evaluation, we introduce two risk primitives verbose response and speculative advice that capture the core failure patterns. Building on these definitions, we propose SecLens, a black-box, multi-objective search framework that efficiently elicits secondary risk behaviors by optimizing task relevance, risk activation, and linguistic plausibility. To support reproducible evaluation, we release SecRiskBench, a benchmark dataset of 650 prompts covering eight diverse real-world risk categories. Experimental results from extensive evaluations on 16 popular models demonstrate that secondary risks are widespread, transferable across models, and modality independent, emphasizing the urgent need for enhanced safety mechanisms to address benign yet harmful LLM behaviors in real-world deployments.

FGS-Audio: Fixed-Decoder Framework for Audio Steganography with Adversarial Perturbation Generation

The rapid development of Artificial Intelligence Generated Content (AIGC) has made high-fidelity generated audio widely available across the Internet, offering an abundant and versatile source of cover signals for covert communication. Driven by advances in deep learning, current audio steganography frameworks are mainly based on encoding-decoding network architectures. While these methods greatly improve the security of audio steganography, they typically employ elaborate training workflows and rely on extensive pre-trained models. To address the aforementioned issues, this paper pioneers a Fixed-Decoder Framework for Audio Steganography with Adversarial Perturbation Generation (FGS-Audio). The adversarial perturbations that carry secret information are embedded into cover audio to generate stego audio. The receiver only needs to share the structure and weights of the fixed decoding network to accurately extract the secret information from the stego audio, thus eliminating the reliance on large pre-trained models. In FGS-Audio, we propose an audio Adversarial Perturbation Generation (APG) strategy and design a lightweight fixed decoder. The fixed decoder guarantees reliable extraction of the hidden message, while the adversarial perturbations are optimized to keep the stego audio perceptually and statistically close to the cover audio, thereby improving resistance to steganalysis. The experimental results show that the method exhibits excellent anti-steganalysis performance under different relative payloads, outperforming existing SOTA approaches. In terms of stego audio quality, FGS-Audio achieves an average PSNR improvement of over 10 dB compared to SOTA method.

Protecting Copyright of Medical Pre-trained Language Models: Training-Free Backdoor Watermarking

Pre-training language models followed by fine-tuning on specific tasks is standard in NLP, but traditional models often underperform when applied to the medical domain, leading to the development of specialized medical pre-trained language models (Med-PLMs). These models are valuable assets but are vulnerable to misuse and theft, requiring copyright protection. However, no existing watermarking methods are tailored for Med-PLMs, and adapting general PLMs watermarking techniques to the medical domain faces challenges such as task incompatibility, loss of fidelity, and inefficiency. To address these issues, we propose the first training-free backdoor watermarking method for Med-PLMs. Our method uses rare special symbols as trigger words, which do not impact downstream task performance, embedding watermarks by replacing their original embeddings with those of specific medical terms in the Med-PLMs' word embeddings layer. After fine-tuning the watermarked Med-PLMs on various medical downstream tasks, the final models (FMs) respond to the trigger words in the same way they would to the corresponding medical terms. This property can be utilized to extract the watermark. Experiments demonstrate that our method achieves high fidelity while effectively extracting watermarks across various medical downstream tasks. Additionally, our method demonstrates robustness against various attacks and significantly enhances the efficiency of watermark embedding, reducing the embedding time from 10 hours to 10 seconds.

A Survey of Fragile Model Watermarking

Model fragile watermarking, inspired by both the field of adversarial attacks on neural networks and traditional multimedia fragile watermarking, has gradually emerged as a potent tool for detecting tampering, and has witnessed rapid development in recent years. Unlike robust watermarks, which are widely used for identifying model copyrights, fragile watermarks for models are designed to identify whether models have been subjected to unexpected alterations such as backdoors, poisoning, compression, among others. These alterations can pose unknown risks to model users, such as misidentifying stop signs as speed limit signs in classic autonomous driving scenarios. This paper provides an overview of the relevant work in the field of model fragile watermarking since its inception, categorizing them and revealing the developmental trajectory of the field, thus offering a comprehensive survey for future endeavors in model fragile watermarking.

Fragile Model Watermarking: A Comprehensive Survey of Evolution, Characteristics, and Classification

Model fragile watermarking, inspired by both the field of adversarial attacks on neural networks and traditional multimedia fragile watermarking, has gradually emerged as a potent tool for detecting tampering, and has witnessed rapid development in recent years. Unlike robust watermarks, which are widely used for identifying model copyrights, fragile watermarks for models are designed to identify whether models have been subjected to unexpected alterations such as backdoors, poisoning, compression, among others. These alterations can pose unknown risks to model users, such as misidentifying stop signs as speed limit signs in classic autonomous driving scenarios. This paper provides an overview of the relevant work in the field of model fragile watermarking since its inception, categorizing them and revealing the developmental trajectory of the field, thus offering a comprehensive survey for future endeavors in model fragile watermarking.

AutoBreach: Universal and Adaptive Jailbreaking with Efficient Wordplay-Guided Optimization

Despite the widespread application of large language models (LLMs) across various tasks, recent studies indicate that they are susceptible to jailbreak attacks, which can render their defense mechanisms ineffective. However, previous jailbreak research has frequently been constrained by limited universality, suboptimal efficiency, and a reliance on manual crafting. In response, we rethink the approach to jailbreaking LLMs and formally define three essential properties from the attacker' s perspective, which contributes to guiding the design of jailbreak methods. We further introduce AutoBreach, a novel method for jailbreaking LLMs that requires only black-box access. Inspired by the versatility of wordplay, AutoBreach employs a wordplay-guided mapping rule sampling strategy to generate a variety of universal mapping rules for creating adversarial prompts. This generation process leverages LLMs' automatic summarization and reasoning capabilities, thus alleviating the manual burden. To boost jailbreak success rates, we further suggest sentence compression and chain-of-thought-based mapping rules to correct errors and wordplay misinterpretations in target LLMs. Additionally, we propose a two-stage mapping rule optimization strategy that initially optimizes mapping rules before querying target LLMs to enhance the efficiency of AutoBreach. AutoBreach can efficiently identify security vulnerabilities across various LLMs, including three proprietary models: Claude-3, GPT-3.5, GPT-4 Turbo, and two LLMs' web platforms: Bingchat, GPT-4 Web, achieving an average success rate of over 80% with fewer than 10 queries

FaceCat: Enhancing Face Recognition Security with a Unified Generative Model Framework

Face anti-spoofing (FAS) and adversarial detection (FAD) have been regarded as critical technologies to ensure the safety of face recognition systems. As a consequence of their limited practicality and generalization, some existing methods aim to devise a framework capable of concurrently detecting both threats to address the challenge. Nevertheless, these methods still encounter challenges of insufficient generalization and suboptimal robustness, potentially owing to the inherent drawback of discriminative models. Motivated by the rich structural and detailed features of face generative models, we propose FaceCat which utilizes the face generative model as a pre-trained model to improve the performance of FAS and FAD. Specifically, FaceCat elaborately designs a hierarchical fusion mechanism to capture rich face semantic features of the generative model. These features then serve as a robust foundation for a lightweight head, designed to execute FAS and FAD tasks simultaneously. As relying solely on single-modality data often leads to suboptimal performance, we further propose a novel text-guided multi-modal alignment strategy that utilizes text prompts to enrich feature representation, thereby enhancing performance. For fair evaluations, we build a comprehensive protocol with a wide range of 28 attack types to benchmark the performance. Extensive experiments validate the effectiveness of FaceCat generalizes significantly better and obtains excellent robustness against input transformations.

Fragile Model Watermark for integrity protection: leveraging boundary volatility and sensitive sample-pairing

Neural networks have increasingly influenced people's lives. Ensuring the faithful deployment of neural networks as designed by their model owners is crucial, as they may be susceptible to various malicious or unintentional modifications, such as backdooring and poisoning attacks. Fragile model watermarks aim to prevent unexpected tampering that could lead DNN models to make incorrect decisions. They ensure the detection of any tampering with the model as sensitively as possible.However, prior watermarking methods suffered from inefficient sample generation and insufficient sensitivity, limiting their practical applicability. Our approach employs a sample-pairing technique, placing the model boundaries between pairs of samples, while simultaneously maximizing logits. This ensures that the model's decision results of sensitive samples change as much as possible and the Top-1 labels easily alter regardless of the direction it moves.

Adaptive White-Box Watermarking with Self-Mutual Check Parameters in Deep Neural Networks

Artificial Intelligence (AI) has found wide application, but also poses risks due to unintentional or malicious tampering during deployment. Regular checks are therefore necessary to detect and prevent such risks. Fragile watermarking is a technique used to identify tampering in AI models. However, previous methods have faced challenges including risks of omission, additional information transmission, and inability to locate tampering precisely. In this paper, we propose a method for detecting tampered parameters and bits, which can be used to detect, locate, and restore parameters that have been tampered with. We also propose an adaptive embedding method that maximizes information capacity while maintaining model accuracy. Our approach was tested on multiple neural networks subjected to attacks that modified weight parameters, and our results demonstrate that our method achieved great recovery performance when the modification rate was below 20%. Furthermore, for models where watermarking significantly affected accuracy, we utilized an adaptive bit technique to recover more than 15% of the accuracy loss of the model.

AdvFAS: A robust face anti-spoofing framework against adversarial examples

Ensuring the reliability of face recognition systems against presentation attacks necessitates the deployment of face anti-spoofing techniques. Despite considerable advancements in this domain, the ability of even the most state-of-the-art methods to defend against adversarial examples remains elusive. While several adversarial defense strategies have been proposed, they typically suffer from constrained practicability due to inevitable trade-offs between universality, effectiveness, and efficiency. To overcome these challenges, we thoroughly delve into the coupled relationship between adversarial detection and face anti-spoofing. Based on this, we propose a robust face anti-spoofing framework, namely AdvFAS, that leverages two coupled scores to accurately distinguish between correctly detected and wrongly detected face images. Extensive experiments demonstrate the effectiveness of our framework in a variety of settings, including different attacks, datasets, and backbones, meanwhile enjoying high accuracy on clean examples. Moreover, we successfully apply the proposed method to detect real-world adversarial examples.