LLM vibe-coded safety scripts achieve ~85% execution success but suffer ~45% silent failure rate in safety logic, with GPT-4o-Mini inaccurate in ~56% of working cases.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.SE 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Is Vibe Coding the Future? An Empirical Assessment of LLM Generated Codes for Construction Safety
LLM vibe-coded safety scripts achieve ~85% execution success but suffer ~45% silent failure rate in safety logic, with GPT-4o-Mini inaccurate in ~56% of working cases.