Preemptive Detection and Steering of LLM Misalignment via Latent Reachability
Large language models (LLMs) are now ubiquitous in everyday tools, raising urgent safety concerns about their tendency to generate harmful content. The dominant safety approach -- reinforcement learning from human feedback (RLHF) -- effectively shapes model behavior during training but offers no safeguards at inference time, where unsafe continuations may still arise. We propose BRT-Align, a reachability-based framework that brings control-theoretic safety tools to LLM inference. BRT-Align models autoregressive generation as a dynamical system in latent space and learn a safety value function via backward reachability, estimating the worst-case evolution of a trajectory. This enables two complementary mechanisms: (1) a runtime monitor that forecasts unsafe completions several tokens in advance, and (2) a least-restrictive steering filter that minimally perturbs latent states to redirect generation away from unsafe regions. Experiments across multiple LLMs and toxicity benchmarks demonstrate that BRT-Align provides more accurate and earlier detection of unsafe continuations than baselines. Moreover, for LLM safety alignment, BRT-Align substantially reduces unsafe generations while preserving sentence diversity and coherence. Qualitative results further highlight emergent alignment properties: BRT-Align consistently produces responses that are less violent, less profane, less offensive, and less politically biased. Together, these findings demonstrate that reachability analysis provides a principled and practical foundation for inference-time LLM safety.
Code (0)
등록된 구현이 없습니다.
Tasks
Reinforcement LearningSimilar Papers 제목 키워드 기반
Concept Reachability in Diffusion Models: Beyond Dataset Constraints
Despite significant advances in quality and complexity of the generations in text-to-image models, prompting does not always lead to the desired outputs. Controlling model behaviour by directly steering intermediate mode…
Preemptive Detection and Correction of Misaligned Actions in LLM Agents
Deploying LLM-based agents in real-life applications often faces a critical challenge: the misalignment between agents' behavior and user intent. Such misalignment may lead agents to unintentionally execute critical acti…
Action DetectionDecision MakingUncertainty-aware Latent Safety Filters for Avoiding Out-of-Distribution Failures
Recent advances in generative world models have enabled classical safe control methods, such as Hamilton-Jacobi (HJ) reachability, to generalize to complex robotic systems operating directly from high-dimensional sensor …
Conformal PredictionActivation Steering Induces Emergent Misalignment: A More Comprehensive Evaluation
Activation steering has emerged as a popular inference-time technique for modulating the behavior of large language models (LLMs). By constructing a steering vector from examples of a target behavior and injecting it int…
Vision-Language Introspection: Mitigating Overconfident Hallucinations in MLLMs via Interpretable Bi-Causal Steering
Object hallucination critically undermines the reliability of Multimodal Large Language Models, often stemming from a fundamental failure in cognitive introspection, where models blindly trust linguistic priors over spec…