Related papers: Prompt-Activation Duality: Improving Activation St…

This work introduces a new Transformer model called Cached Transformer, which uses Gated Recurrent Cached (GRC) attention to extend the self-attention mechanism with a differentiable memory cache of tokens. GRC attention enables attending…

Trajectory prediction models in autonomous driving are vulnerable to perturbations from non-causal agents whose actions should not affect the ego-agent's behavior. Such perturbations can lead to incorrect predictions of other agents'…

Driver's cognitive ability at a given moment is the most elusive variable in assessing driver's safety. In contrast to other physical conditions, such as short-sight, or manual disability cognitive ability is transient. Safety regulations…

Activation steering methods are widely used to control large language model (LLM) behavior and are often interpreted as revealing meaningful internal representations. This interpretation assumes that steering directions are identifiable and…

Self-attention has greatly contributed to the success of the widely used Transformer architecture by enabling learning from data with long-range dependencies. In an effort to improve performance, a gated attention model that leverages a…

Applying steering vectors to large language models (LLMs) is an efficient and effective model alignment technique, but we lack an interpretable explanation for how it works-- specifically, what internal mechanisms steering vectors affect…

Recent work on activation and latent steering has demonstrated that modifying internal representations can effectively guide large language models (LLMs) toward improved reasoning and efficiency without additional training. However, most…

Controllable generation requires language models to realize output characteristics such as reading level, politeness, and toxicity. Existing steering methods are often indirect, require access to internal activations, or depend on auxiliary…

Controlling specific behaviors in large language models while preserving their general capabilities is a central challenge for safe and reliable artificial intelligence deployment. Current steering methods, such as vector addition and…

Steering vectors offer a training-free mechanism for controlling reasoning behaviors in large language models, but constructing effective vectors requires identifying genuine behavioral signals in the model's hidden states. For behaviors…

Autonomous driving in multi-agent dynamic traffic scenarios is challenging: the behaviors of road users are uncertain and are hard to model explicitly, and the ego-vehicle should apply complicated negotiation skills with them, such as…

Understanding how large audio models represent music, and using that understanding to steer generation, is both challenging and underexplored. Inspired by mechanistic interpretability in language models, where direction vectors in…

Previous studies have demonstrated the effectiveness of point-based neural models on the point cloud analysis task. However, there remains a crucial issue on producing the efficient input embedding for raw point coordinates. Moreover,…

We present an approach called Dialogue Action Tokens (DAT) that adapts language model agents to plan goal-directed dialogues. The core idea is to treat each utterance as an action, thereby converting dialogues into games where existing…

Large language models (LLMs) can memorize and reproduce training sequences verbatim -- a tendency that undermines both generalization and privacy. Existing mitigation methods apply interventions uniformly, degrading performance on the…

The recent works on a deep learning (DL)-based joint design of preamble set for the transmitters and data-aided active user detection (AUD) in the receiver has demonstrated a significant performance improvement for grant-free sparse code…

Attention mechanisms underpin the success of large language models (LLMs), yet their substantial computational and memory overhead poses challenges for optimizing efficiency and performance. A critical bottleneck arises as KV cache and…

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…

Inference-time guided sampling steers state-of-the-art diffusion and flow models without fine-tuning by interpreting the generation process as a controllable trajectory. This provides a simple and flexible way to inject external constraints…

Vision-Language-Action (VLA) models leverage powerful perceptual priors from web-scale Vision-Language Model (VLM) pre-training, yet they remain surprisingly brittle in practice, frequently failing at simple robotic tasks. To mitigate this,…