Evaluating LLM safety with one canonical prompt understates unsafe behavior; across five meaning-preserving reformulations, 5-13% of safe-on-canonical seeds become unsafe, and the union exceeds the worst single form for all five models.
Pappas, Florian Tramer, Hamed Hassani, and Eric Wong
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CR 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Single Canonical Prompts Underestimate LLM Safety's Surface-Form Sensitivity
Evaluating LLM safety with one canonical prompt understates unsafe behavior; across five meaning-preserving reformulations, 5-13% of safe-on-canonical seeds become unsafe, and the union exceeds the worst single form for all five models.