ASR metrics for LLM jailbreaks are inflated by stochasticity; CAS-eval reveals up to 30pp drops under multi-attempt criteria while CAS-gen recovers the performance loss.
Stegun.Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical T ables
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
-
The Great Pretender: A Stochasticity Problem in LLM Jailbreak
ASR metrics for LLM jailbreaks are inflated by stochasticity; CAS-eval reveals up to 30pp drops under multi-attempt criteria while CAS-gen recovers the performance loss.