Injecting brief safety-plausible phrases into robot audio triggers LLM safety halts, enabling semantic denial-of-service attacks where prompt defenses trade attack suppression for impaired genuine hazard detection.
Accessed: 2026-04-22
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CR 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Semantic Denial of Service in LLM-controlled robots
Injecting brief safety-plausible phrases into robot audio triggers LLM safety halts, enabling semantic denial-of-service attacks where prompt defenses trade attack suppression for impaired genuine hazard detection.