相关论文: CRaFT: Circuit-Guided Refusal Feature Selection vi…

Grounded multi-video question answering over real-world news events requires systems to surface query-relevant evidence across heterogeneous video archives while attributing every claim to its supporting source. We introduce CRAFT…

Representation Engineering analyses often characterize refusal using static directions extracted from terminal or pooled representations. We ask whether this view misses how refusal is constructed across layer-token positions. Using causal…

Responsible deployment of language models requires mechanisms for refusing unsafe prompts while preserving model performance. While most approaches modify model weights through additional training, we explore an alternative: steering model…

LLMs are trained to refuse harmful instructions, but do they truly understand harmfulness beyond just refusing? Prior work has shown that LLMs' refusal behaviors can be mediated by a one-dimensional subspace, i.e., a refusal direction. In…

Automated Code Review (ACR) systems integrating Large Language Models (LLMs) are increasingly adopted in software development workflows, ranging from interactive assistants to autonomous agents in CI/CD pipelines. In this paper, we study…

As the deep learning techniques have expanded to real-world recommendation tasks, many deep neural network based Collaborative Filtering (CF) models have been developed to project user-item interactions into latent feature space, based on…

Despite the remarkable performance of Large Language Models (LLMs), they remain vulnerable to jailbreak attacks, which can compromise their safety mechanisms. Existing studies often rely on brute-force optimization or manual design, failing…

Activation steering has emerged as a powerful tool to shape LLM behavior without the need for weight updates. While its inherent brittleness and unreliability are well-documented, its safety implications remain underexplored. In this work,…

The massive scale of modern AI accelerators presents critical challenges to traditional fault assessment methodologies, which face prohibitive computational costs and provide poor coverage of critical failure modes. This paper introduces…

Jailbreaks have been a central focus of research regarding the safety and reliability of large language models (LLMs), yet the mechanisms underlying these attacks remain poorly understood. While previous studies have predominantly relied on…

Refusal refers to the functional behavior enabling safety-aligned language models to reject harmful or unethical prompts. Following the growing scientific interest in mechanistic interpretability, recent work encoded refusal behavior as a…

Instruction Fine-Tuning (IFT) has been widely adopted as an effective post-training strategy to enhance various abilities of Large Language Models (LLMs). However, prior studies have shown that IFT can significantly compromise LLMs' safety,…

Recent work has shown that language models' refusal behavior is primarily encoded in a single direction in their latent space, making it vulnerable to targeted attacks. Although Latent Adversarial Training (LAT) attempts to improve…

Graph prompt tuning has emerged as a promising paradigm to effectively transfer general graph knowledge from pre-trained models to various downstream tasks, particularly in few-shot contexts. However, its susceptibility to backdoor attacks,…

Current alignment evaluation mostly measures whether models encode dangerous concepts and whether they refuse harmful requests. Both miss the layer where alignment often operates: routing from concept detection to behavioral policy. We…

We localize the policy routing mechanism in alignment-trained language models. An intermediate-layer attention gate reads detected content and triggers deeper amplifier heads that boost the signal toward refusal. In smaller models the gate…

Task-oriented LLM-based agents are increasingly used in domains with strict policies, such as refund eligibility or cancellation rules. The challenge lies in ensuring that the agent consistently adheres to these rules and policies,…

The first layer of a deep CNN backbone applies filters to an image to extract the basic features available to later layers. During training, some filters may go inactive, mean ing all weights in the filter approach zero. An inactive fil ter…

Large Language Models (LLMs) suffer from a range of vulnerabilities that allow malicious users to solicit undesirable responses through manipulation of the input text. These so-called jailbreak prompts are designed to trick the LLM into…

Uncovering the mechanisms behind "jailbreaks" in large language models (LLMs) is crucial for enhancing their safety and reliability, yet these mechanisms remain poorly understood. Existing studies predominantly analyze jailbreak prompts by…