Related papers: Guiding Giants: Lightweight Controllers for Weight…
Large Language Models (LLMs) encode vast amounts of parametric knowledge during pre-training. As world knowledge evolves, effective deployment increasingly depends on their ability to faithfully follow externally retrieved context. When…
Safety alignment in large language models (LLMs) is achieved through fine-tuning mechanisms that regulate neuron activations to suppress harmful content. In this work, we propose a novel approach to induce disalignment by identifying and…
Large Language Model (LLM) workloads have distinct prefill and decode phases with different compute and memory requirements which should ideally be accounted for when scheduling input queries across different LLM instances in a cluster.…
Large language models (LLMs) exhibit distinct and consistent personalities that greatly impact trust and engagement. While this means that personality frameworks would be highly valuable tools to characterize and control LLMs' behavior,…
Training large language models (LLMs) is highly memory-intensive, as training must store not only weights and optimizer states but also intermediate activations for backpropagation. While existing memory-efficient methods largely focus on…
Large language models (LLMs), especially generative pre-trained transformers (GPTs), have recently demonstrated outstanding ability in information comprehension and problem-solving. This has motivated many studies in applying LLMs to…
Large Language Models (LLMs) have significantly advanced natural language processing with exceptional task generalization capabilities. Low-Rank Adaption (LoRA) offers a cost-effective fine-tuning solution, freezing the original model…
Large language models (LLMs) are typically aligned to refuse harmful instructions through safety fine-tuning. A recent attack, termed abliteration, identifies and suppresses the single latent direction most responsible for refusal behavior,…
Modern large language models (LLMs) are typically secured by auditing data, prompts, and refusal policies, while treating the forward pass as an implementation detail. We show that intermediate activations in decoder-only LLMs form a…
Masked diffusion language models (MDLMs) generate text via iterative masked-token denoising, enabling mask-parallel decoding and distinct controllability and efficiency tradeoffs from autoregressive LLMs. Yet, efficient representation-level…
As large language models (LLMs) evolve in complexity and capability, the efficacy of less widely deployed alignment techniques are uncertain. Building on previous work on activation steering and contrastive activation addition (CAA), this…
Research demonstrates that the proactivity of in-vehicle conversational assistants (IVCAs) can help to reduce distractions and enhance driving safety, better meeting users' cognitive needs. However, existing IVCAs struggle with user intent…
Latent space steering methods provide a practical approach to controlling large language models by applying steering vectors to intermediate activations, guiding outputs toward desired behaviors while avoiding retraining. Despite their…
Despite the strong performance of large language models (LLMs) across diverse tasks, their susceptibility to adversarial attacks and unsafe content generation remains a significant obstacle to deployment, particularly in high-stakes…
Multimodal LLMs (MLLMs) have reached remarkable levels of proficiency in understanding multimodal inputs. However, understanding and interpreting the behavior of such complex models is a challenging task, not to mention the dynamic shifts…
Large language models (LLMs) are shown to be vulnerable to jailbreaking attacks where adversarial prompts are designed to elicit harmful responses. While existing defenses effectively mitigate single-turn attacks by detecting and filtering…
Adjusting the latency, power, and accuracy of natural language understanding models is a desirable objective of an efficient architecture. This paper proposes an efficient Transformer architecture that adjusts the inference computational…
Despite extensive safety-tuning, large language models (LLMs) remain vulnerable to jailbreak attacks via adversarially crafted instructions, reflecting a persistent trade-off between safety and task performance. In this work, we propose…
The widespread adoption of Large Language Models (LLMs) in critical applications has introduced severe reliability and security risks, as LLMs remain vulnerable to notorious threats such as hallucinations, jailbreak attacks, and backdoor…
Recent findings reveal that much of the knowledge in a Transformer-based Large Language Model (LLM) is encoded in its feed-forward (FFN) layers, where each FNN layer can be interpreted as the summation of sub-updates, each corresponding to…