arXiv AI

Local Sparsity Enables Unsupervised LLM Safety Detection

The paper proposes a new unsupervised safety detection method for large language models that relies on anomaly detection rather than supervised training on unsafe data. By leveraging local sparsity in a linear representation space obtained via a sparse autoencoder, the authors develop a framework for locally masked SAE-based anomaly detection, providing theoretical support and empirical validation across multiple architectures and datasets. When calibrated with only 1% out-of-distribution data, the method achieves near‑optimal performance while using just 1–2% of SAE neurons for computation.

Hugging Face Trending Papers
6d ago

Local Sparsity Enables Unsupervised LLM Safety Detection

The paper proposes a novel unsupervised safety detection method for large language models that relies on local sparsity in a linear representation space recovered via a sparse autoencoder. By masking SAE neurons based on shared active support among nearby points, the authors develop a locally masked anomaly detection framework with theoretical backing. Experiments across multiple architectures and datasets—including capability‑testing and safety‑specific sets—show that using only 1–2% of SAE neurons and a small amount of out‑of‑distribution data yields near‑optimal safety detection performance.

arXiv Computer Vision
Aug 27

Learning to Detect Unseen Jailbreak Attacks in Large Vision-Language Models

The paper introduces Learning to Detect (LoD), a framework for identifying unseen jailbreak attacks in Large Vision‑Language Models without relying on attack data or hand‑crafted heuristics. LoD extracts layer‑wise safety representations via Multi‑modal Safety Concept Activation Vectors and compresses them into a one‑dimensional anomaly score using a Safety Pattern Auto‑Encoder. Experiments show that LoD achieves state‑of‑the‑art AUROC across diverse unseen attacks on multiple LVLMs while improving efficiency.

By Shuang Liang, Zhihao Xu, Jiaqi Weng, Jialing Tao, Hui Xue, Xiting Wang
arXiv AI
Sep 10

SAEs Can Improve Unlearning: Dynamic Sparse Autoencoder Guardrails for Precision Unlearning in LLMs

The paper introduces Dynamic DAE Guardrails (DSG), a method that uses Dynamic Sparse Autoencoders to perform precision unlearning in large language models. DSG leverages principled feature selection and a dynamic classifier to target activation-based unlearning, outperforming existing gradient‑based methods in terms of computational efficiency, stability, sequential unlearning, resistance to relearning attacks, data efficiency, and interpretability.

By Aashiq Muhamed, Jacopo Bonato, Mona Diab, Virginia Smith
arXiv AI
Sep 4

Differentiable Interval Bottlenecks for Interpretable Anomaly Detection in Numerical Data

DIFFINT is a reconstruction‑based anomaly detector that uses a differentiable autoencoder with a latent bottleneck composed of soft, axis‑aligned interval memberships. Each latent unit represents a human‑readable hyper‑rectangle in feature space, allowing the model to encode how strongly an instance falls inside each interval and to compute reconstruction error as the anomaly score. The method provides a certified lower bound on reconstruction error for points outside all active intervals, a suppression mechanism for sparse abnormalities, and a closed‑form, label‑free importance ranking for each (unit, feature) pair, achieving top performance on 48 ADBench benchmarks against 22 baselines.

By Lamine Diop, Marc Plantevit