A safety suite that maps natural-language lab rules into typed, executable runtime monitors, and reports reduced unsafe events in simulation from 39.5% to 23.8%.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AI 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
LabGuard: Grounding Natural-Language Laboratory Rules into Runtime Guards for Embodied Laboratory Agents
A safety suite that maps natural-language lab rules into typed, executable runtime monitors, and reports reduced unsafe events in simulation from 39.5% to 23.8%.