Related papers: Activation Steering Meets Preference Optimization:…

Current safety alignment techniques for large language models (LLMs) face two key challenges: (1) under-generalization, which leaves models vulnerable to novel jailbreak attacks, and (2) over-alignment, which leads to the excessive refusal…

Modern language models often need to optimize a primary accuracy objective while also accommodating secondary behavioral preferences, such as verbosity, agreeableness, or the level of technical expertise in its response. In practice, a base…

As Spoken Language Models (SLMs) integrate speech and text modalities, they inherit the safety vulnerabilities of their LLM backbone and an expanded attack surface. SLMs have been previously shown to be susceptible to jailbreaking, where…

Large Vision-Language Models (LVLMs) demonstrate exceptional performance across multimodal tasks, yet remain vulnerable to jailbreak attacks that bypass built-in safety mechanisms to elicit restricted content generation. Existing black-box…

We present HaoMo Vision-Language Model (HMVLM), an end-to-end driving framework that implements the slow branch of a cognitively inspired fast-slow architecture. A fast controller outputs low-level steering, throttle, and brake commands,…

Generative models are often deployed to make decisions on behalf of users, such as vision-language models (VLMs) identifying which person in a room is a doctor to help visually impaired individuals. Yet, VLM decisions are influenced by the…

Multimodal Large Language Models (MLLMs) have demonstrated impressive capabilities in cross-modal understanding, but remain vulnerable to adversarial attacks through visual inputs despite robust textual safety mechanisms. These…

Recent work has demonstrated the potential of contrastive steering for jailbreaking Large Language Models (LLMs). However, existing methods rely on limited and inherently biased contrastive prompts and require laborious manual tuning of…

Autonomous driving systems often infer pedestrian yielding behavior from geometric and kinematic cues alone, limiting their ability to reason about visual scene context and age-dependent behavioral variability. This limitation can produce…

Jailbreak prompts can trigger harmful completions on aligned LLMs, In accordance, safety steering has been proposed: test-time activation interventions that steer jailbreak activations to trigger refusal while preserving benign utility.…

Large vision-language models (VLMs) often exhibit weakened safety alignment with the integration of the visual modality. Even when text prompts contain explicit harmful intent, adding an image can substantially increase jailbreak success…

Steering Large Language Models (LLMs) through activation interventions has emerged as a lightweight alternative to fine-tuning for alignment and personalization. Recent work on Bi-directional Preference Optimization (BiPO) shows that dense…

Vision language models (VLMs) extend the reasoning capabilities of large language models (LLMs) to cross-modal settings, yet remain highly vulnerable to multimodal jailbreak attacks. Existing defenses predominantly rely on safety…

Large language models (LLMs) exhibit reasoning biases, often conflating content plausibility with formal logical validity. This can lead to wrong inferences in critical domains, where plausible arguments are incorrectly deemed logically…

Clinical ML workflows are often fragmented and inefficient: triage, task selection, and model deployment are handled by a patchwork of task-specific networks. These pipelines are rarely aligned with data-science practice, reducing…

The rapid advancement of pre-trained language models (PLMs) has demonstrated promising results for various code-related tasks. However, their effectiveness in detecting real-world vulnerabilities remains a critical challenge. While existing…

Vision-Language Models (VLMs) face significant safety vulnerabilities from malicious prompt attacks due to weakened alignment during visual integration. Existing defenses suffer from efficiency and robustness. To address these challenges,…

The visual modality of vision-language models (VLMs) is an underexplored attack surface for bypassing safety alignment. We introduce four jailbreak attacks exploiting the vision component: (1) encoding harmful instructions as visual symbol…

Although safely enhanced Large Language Models (LLMs) have achieved remarkable success in tackling various complex tasks in a zero-shot manner, they remain susceptible to jailbreak attacks, particularly the unknown jailbreak attack. To…

Safety-aligned LLMs suffer from two failure modes: jailbreak (answering harmful inputs) and over-refusal (declining benign queries). Existing vector steering methods adjust the magnitude of answer vectors, but this creates a fundamental…