← Back to all news
arXiv Computation and Language By Soyeon Park (KAIST), Seogyeong Jeong (KAIST), Sunwoo Kim (KAIST), Alice Oh (KAIST)

The Role of Fine-grained Harm Signals in LLM Safety

Read the original on arXiv Computation and Language →

The study investigates how category‑specific harmfulness signals, isolated by removing the shared general harmfulness component, influence large language model (LLM) safety. Using activation steering across 11 risk categories in three instruction‑tuned LLMs, the authors find that the presence of harmfulness in these category residuals varies by category and that the pattern of inducing refusal is even more model‑dependent. Additionally, category residuals were shown to enhance the models’ downstream alignment with the shared general harmfulness representation, indicating that fine‑grained signals play a role beyond the general component.

Machine-generated by The Flow from the publisher's headline and feed description — not written or checked by a human. The full article lives at arXiv Computation and Language.

Related stories

arXiv AI
Jul 7

Large Language Models Generate Harmful Responses Using a Distinct Mechanism, Shared Across Harm Types

arXiv:2604. 09544v2 Announce Type: replace-cross Abstract: Large language models (LLMs) undergo alignment training to avoid harmful behaviors, yet the resulting safeguards remain brittle: jailbreaks routinely bypass them, and fine-tuning on narrow domains can induce ``emergent misalignment'' that generalizes broadly.

By Hadas Orgad, Boyi Wei, Kaden Zheng, Martin Wattenberg, Peter Henderson, Seraphina Goldfarb-Tarrant, Yonatan Belinkov
llmsfine-tuningefficiencysafety
More like this →
arXiv Machine Learning
4d ago

How to Tame a Multi-Headed Hydra? Adaptive Multi-Category Safety Steering for Large Language Models

The paper introduces CAM-Steer, a Category‑Adaptive Multi‑category Safety Steering framework that estimates risk for each harm category by comparing hidden states to safe and unsafe prototypes. It then combines safety directions into a single steering vector and applies a rotation whose angle is set by the estimated risks, preserving the hidden‑state norm. Experiments on three LLM backbones and seven harm categories show that CAM‑Steer outperforms baselines in defense success rate, even when multiple harm categories co‑occur, with negligible inference overhead.

By Chenxi Wang, Ruiyang Huang, Li Huang, Yifan Wu
llmssafety
More like this →