21 accepted papers
Alignment is vital for safely deploying large language models (LLMs). Existing techniques are either reward-based--train a reward model on preference pairs and optimize with reinforcement learning (RL)--or reward-free--directly fine-tune on ranked outputs. Recent research show that well-tuned reward…
Mixture-of-Experts (MoE) architectures have emerged as a popular paradigm for scaling Large Language Models (LLMs), enabling greater capacity with reduced computational cost by dynamically routing inputs to the most relevant experts based on learned patterns. However, this also introduces a critical…
As Large Language Models (LLMs) are increasingly used, their security risks have drawn increasing attention. Existing research reveals that LLMs are highly susceptible to jailbreak attacks, with effectiveness varying across language contexts. This paper investigates the role of classical Chinese in…
Backdoor attacks pose potential threats to object detection models, highlighting the importance of studying their security. However, existing backdoor attacks mainly rely on trigger-specific intrinsic features, which limits their practicality in real-world scenarios. In this paper, we propose a nove
AI systems may exhibit deceptive behaviors that mislead developers about their capabilities, propensities, or actions. Such deception can take distinct forms across the development lifecycle: training subversion, evaluation gaming, and control evasion. We argue that the AI community should prioritiz…
While large language models (LLMs) demonstrate impressive performance across various tasks, their deployment in real-world scenarios is still constrained by high computational demands. Layer-wise pruning, a commonly employed strategy to mitigate inference costs, can partially address this challenge.…
Although Large Reasoning Models (LRMs) have progressed in solving complex problems, their chain-of-thought (CoT) reasoning often contains harmful content that can persist even when the final responses appear safe. We show that this issue still remains in existing methods which overlook the unique si…
Large Vision Transformers (ViTs) must often be compressed before they can be deployed on resource-constrained edge devices. However, many edge devices require only part of the *all-classes* knowledge of a pre-trained ViT in their corresponding application scenarios. This is overlooked by existing c…
Multimodal Large Language Models (MLLMs) pose unique safety challenges due to their integration of visual and textual data, thereby introducing new dimensions of potential attacks and complex risk combinations. In this paper, we begin with a detailed analysis aimed at disentangling risks through ste…
With the rapid advancement of multimodal large language models (MLLMs), concerns regarding their security have increasingly captured the attention of both academia and industry. Although MLLMs are vulnerable to jailbreak attacks, designing effective jailbreak attacks poses unique challenges, especia…
Multimodal Large Language Models (MLLMs) have achieved impressive performance and have been put into practical use in commercial applications, but they still have potential safety mechanism vulnerabilities. Jailbreak attacks are red teaming methods that aim to bypass safety mechanisms and discover M…
Understanding the vulnerabilities of Large Vision Language Models (LVLMs) to jailbreak attacks is essential for their responsible real-world deployment. Most previous work requires access to model gradients, or is based on human knowledge (prompt engineering) to complete jailbreak, and they hardly c
Ensuring the safety and harmlessness of Large Language Models (LLMs) has become equally critical as their performance in applications. However, existing safety alignment methods typically suffer from safety-performance trade-offs and susceptibility to jailbreak attacks, primarily due to their relian…
Retrieval-augmented generation (RAG) systems enhance large language models (LLMs) with external knowledge but are vulnerable to corpus poisoning and contamination attacks, which can compromise output integrity. Existing defenses often apply aggressive filtering, leading to unnecessary loss of valuab…
Adversarial Training (AT) has been widely proved to be an effective method to improve the adversarial robustness against adversarial examples for Deep Neural Networks (DNNs). As a variant of AT, Adversarial Robustness Distillation (ARD) has demonstrated its superior performance in improving the robu…
The appearance of adversarial examples raises attention from both academia and industry. Along with the attack-defense arms race, adversarial training is the most effective against adversarial examples. However, we find inequality phenomena occur during the $l_{\infty}$-adversarial training, that fe…
Adversarial Training (AT), which is commonly accepted as one of the most effective approaches defending against adversarial examples, can largely harm the standard performance, thus has limited usefulness on industrial-scale production and applications. Surprisingly, this phenomenon is totally oppos…
Recent advances on Vision Transformer (ViT) and its improved variants have shown that self-attention-based networks surpass traditional Convolutional Neural Networks (CNNs) in most vision tasks. However, existing ViTs focus on the standard accuracy and computation cost, lacking the investigation of…
Human can easily recognize visual objects with lost information: even losing most details with only contour reserved, e.g. cartoon. However, in terms of visual perception of Deep Neural Networks (DNNs), the ability for recognizing abstract objects (visual objects with lost information) is still a ch…
Though it is well known that the performance of deep neural networks (DNNs) degrades under certain light conditions, there exists no study on the threats of light beams emitted from some physical source as adversarial attacker on DNNs in a real-world scenario. In this work, we show by simply using a…
Deep neural networks (DNNs) are known to be vulnerable to adversarial examples. Existing works have mostly focused on either digital adversarial examples created via small and imperceptible perturbations, or physical-world adversarial examples created with large and less realistic distortions that a…