When Words Are Safe But Actions Kill: Probing Physical Danger Beyond Text Safety in Hidden-State Risk Space
Large language models (LLMs) increasingly serve as high-level planners for embodied agents, where linguistically benign instructions can become unsafe once grounded in the physical world. We study whether this physically grounded danger is the same safety problem as ordinary text-level content danger. Through hidden-state direction analysis and random-split null tests, we show that content danger (CD) and physical danger (PD) form separable signals in LLM representations across Qwen2.5-3B/7B/14B/32B, Phi-3.5 and SmolLM2. Building on the CD/PD separability, we propose PRISM, a single-layer L2-r
Lineage graph
Paper → model → repo connections mined from source citations (Tier-1 exact match).
Why these links exist
Every edge carries a method, confidence, and the source snippet that justified it — so bad links are debuggable.
- PossiblePossibly related (embedding) · 51%Nemotron 3.5 Content Safety: Customizable Multimodal Safety for Global Enterprise AI →
- PossiblePossibly related (embedding) · 49%New Research: AI models can give dangerous responses despite output guardrails - The AI Journal →
- PossiblePossibly related (embedding) · 49%Evaluating J-space entropy as an error predictor across 7 datasets on Qwen3-4B [R] →
- FuzzyOverlapping authors or contributors · 62%bytedance/deer-flow →
“Shared author/contributor keys: wang”
- FuzzyOverlapping authors or contributors · 62%ray-project/ray →
“Shared author/contributor keys: wang”
- LinkedLinked via arxiv author · 85%Weimeng Wang →
“When Words Are Safe But Actions Kill: Probing Physical Danger Beyond Text Safety in Hidden-State Risk Space”
- LinkedLinked via arxiv author · 85%Ziqiang Wang →
“When Words Are Safe But Actions Kill: Probing Physical Danger Beyond Text Safety in Hidden-State Risk Space”
- LinkedLinked via arxiv author · 85%Zihang Zhan →
“When Words Are Safe But Actions Kill: Probing Physical Danger Beyond Text Safety in Hidden-State Risk Space”