Related papers: Dynamically Scaled Activation Steering

Automated Driving Systems (ADS) development relies on utilizing real-world vehicle data. The volume of data generated by modern vehicles presents transmission, storage, and computational challenges. Focusing on Dynamic Behavior (DB) offers…

Alignment in LLMs is more brittle than commonly assumed: misalignment can be triggered by adversarial prompts, benign fine-tuning, emergent misalignment, and goal misgeneralization. Recent evidence suggests that some misalignment behaviors…

Activation steering controls language model behavior by adding directions to internal representations at inference time, but standard residual-stream steering can fail in stateful dialogue. We identify KV-cache contamination as a key…

Language models often default to a narrow set of high-probability outputs, leaving their generation paths homogeneous and prone to mode collapse. Sampling-based strategies inject randomness but still struggle to guarantee diversity across…

Recent progress in Multimodal Large Language Models (MLLMs) has unlocked powerful cross-modal reasoning abilities, but also raised new safety concerns, particularly when faced with adversarial multimodal inputs. To improve the safety of…

Safety alignment is indispensable for Large Language Models (LLMs) to defend threats from malicious instructions. However, recent researches reveal safety-aligned LLMs prone to reject benign queries due to the exaggerated safety issue,…

Sparse autoencoders (SAEs) have recently emerged as a powerful tool for language model steering. Prior work has explored top-k SAE latents for steering, but we observe that many dimensions among the top-k latents capture non-semantic…

Intelligent agents need a physical understanding of the world to predict the impact of their actions in the future. While learning-based models of the environment dynamics have contributed to significant improvements in sample efficiency…

Sparse Autoencoders (SAEs) have been proposed as an unsupervised approach to learn a decomposition of a model's latent space. This enables useful applications such as steering - influencing the output of a model towards a desired concept -…

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…

Automated Driving Systems (ADS) have made great achievements in recent years thanks to the efforts from both academia and industry. A typical ADS is composed of multiple modules, including sensing, perception, planning, and control, which…

Autonomous cars can reduce road traffic accidents and provide a safer mode of transport. However, key technical challenges, such as safe navigation in complex urban environments, need to be addressed before deploying these vehicles on the…

Activation steering, or representation engineering, offers a lightweight approach to align large language models (LLMs) by manipulating their internal activations at inference time. However, current methods suffer from two key limitations:…

We present Causal Amortized Active Structure Learning (CAASL), an active intervention design policy that can select interventions that are adaptive, real-time and that does not require access to the likelihood. This policy, an amortized…

Despite extensive efforts in safety alignment, large language models (LLMs) remain vulnerable to jailbreak attacks. Activation steering offers a training-free defense method but relies on fixed steering coefficients, resulting in suboptimal…

Adapting LLM agents to domain-specific tool calling remains notably brittle under evolving interfaces. Prompt and schema engineering is easy to deploy but often fragile under distribution shift and strict parsers, while continual…

Vision-Language-Action (VLA) models are a promising path to realizing generalist embodied agents that can quickly adapt to new tasks, modalities, and environments. However, methods for interpreting and steering VLAs fall far short of…

Model-based design of experiments (MBDOE) is essential for efficient parameter estimation in nonlinear dynamical systems. However, conventional adaptive MBDOE requires costly posterior inference and design optimization between each…

In order to drive safely and efficiently under merging scenarios, autonomous vehicles should be aware of their surroundings and make decisions by interacting with other road participants. Moreover, different strategies should be made when…

Aligning Large Language Models (LLMs) with specific personas typically relies on expensive and monolithic Supervised Fine-Tuning (SFT) or RLHF. While effective, these methods require training distinct models for every target personality…